January 16, 2025 • 53 mins
What if harnessing the power of edge computing could transform content delivery and redefine your streaming experience? Join us as we explore this exciting frontier with Mark Fisher from Quilt, a leading innovator in the space. Mark shares his extensive journey through significant internet trends, highlighting Quilt's groundbreaking partnerships with industry giants like Comcast and their pivotal role in setting open caching standards. His insights reveal how these collaborations are not only enhancing performance but also shaping the very future of streaming technologies.
Our conversation takes a deep dive into the dynamic world of edge computing partnerships. We uncover how strategic alliances with operators like Comcast and technology titans like Cisco are facilitating the deployment of edge nodes, creating a seamless integration into cloud-controlled platforms. This discussion unveils the potential of real-time applications, such as gaming and rendering services, at the network edge, and offers a glimpse into the exciting industry developments that are unfolding. Mark helps us navigate through the complexities of implementing technologies like WebRTC, ensuring no single solution is seen as a one-size-fits-all.
The future of streaming is rife with challenges and opportunities, especially as the migration from broadcast to IP networks accelerates. We discuss the critical importance of network capacity for high-quality streaming, particularly during major global events, and the race for ultra-low latency delivery to sync with broadcast signals. As we look at emerging trends, Mark offers an intriguing perspective on the evolving landscape, including Netflix's shift towards live content and advertising. Plus, we get a fascinating peek into Quilt's expansion in the Middle East through a partnership with Saudi Telecom, underscoring the global reach and influence of edge computing advancements.
Stay tuned for more in-depth insights on video technology, trends, and practical applications. Subscribe to Voices of Video: Inside the Tech for exclusive, hands-on knowledge from the experts. For more resources, visit Voices of Video.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:07):
voices of video.
Voices of video.
The voices of video voices ofvideo okay, well, good morning
everyone.
We are back for another veryexciting edition of Voices of
Video, and I have with me todayMark Fisher from Quilt.
(00:29):
So first of all, mark, welcometo Voices of Video.
Speaker 2 (00:34):
Thanks, mark, good to be here with you.
Speaker 1 (00:36):
Yeah, yeah, so we were talking before we started,
you know, before we opened thesession, and it's been 10 years
since we crossed paths Now Ithink maybe it was a few years
before that.
Even We'll get into exactlywhat we're referencing, because
we're going to talk about thiskind of a cool thing that you
(00:57):
know, Quilt has been doingbeyond building your product.
So, anyway, a little teaserwe're dangling out there, right?
But, yes, welcome to the show.
And I just want to say foreverybody listening live, if you
have questions, this is afireside chat type format and we
(01:21):
want to get to your questionsand we're very happy to take
really any question and if wecan't answer it, we'll just say
so, right, yeah, so so feel freeto just, you know, type in some
questions in the chat and ourproducer, anita, will pick those
(01:41):
up and then you know we'llwe'll do our best to get to them
.
So with that you know, mark,why don't you say a few words
about who you are and what youdo at Quilt?
And then you know, I wouldn'tassume that everybody knows
exactly who Quilt is and whatyou do.
So let's start there.
Speaker 2 (02:04):
Yeah, great, and thanks for the invitation, mark,
to join you today.
So I've been working in variousstartups now for a number of
years, maybe tracing or trackingwith the big trends across the
internet Early 90s, thecommercialization of the
internet, internet serviceproviders, then to the mobile
(02:27):
Internet with a company, danger,that was acquired by Microsoft,
then to Internet of Things, acompany, jasper, which was
acquired by Cisco, and now atQuilt.
I'm in every case and now I'mresponsible for marketing
business development.
And now I'm responsible formarketing business development.
It's been, you know, a greatexperience across the board, and
(02:47):
what makes it great are thepeople that you are with, that
you work with, whether they'rein the company where you work or
across the industry.
That, for me, has always beenvery rewarding.
Yes, so Quilt and I agree withyou, not everyone would know us.
We obviously had big news thisweek, an announcement of a
partnership with Comcast, whichwe're very proud of and was
(03:09):
years Absolutely Spend some timetalking about that for sure.
So Quilt is a software and cloudservices company.
We're unique because we reallyare focused on edge computing as
a general category, but what wedo to go to market is unique.
(03:33):
In particular, we have a verystrategic relationship with
Cisco, who's an investor, andwith them we go out and partner
with big service providers andsometimes smaller service
providers, and we do that to putour software on compute
(03:55):
hardware deep inside theirnetworks, close to consumers or
businesses.
And once that's done, thatinfrastructure is in place, that
edge cloud, service provider,edge cloud then we turn, we
unite it, we federate it, if youwill, with our cloud services.
And then we turn to today bigcontent providers, some of the
(04:17):
biggest stream platforms in theworld, and we say, hey, we've
got a better way to deliver yourcontent live streams, vod or,
in some cases, just softwaredownloads.
And we can demonstrate thedifference by our performance,
where we routinely outperformevery other commercial CDN
because we have this advantageof being closer.
(04:38):
It's sort of that simple.
At the beginning of what we'redoing, we're taking that use
case as the foundation.
That's what gets us in the doorwith these big service
providers, like a Comcast whowe've announced this week, and
then we're going to turn ourattention to other edge
computing use cases.
We can talk about that as timegoes on, yeah, yeah, amazing.
Speaker 1 (05:03):
Remind me, by the way , when was the company founded?
What year was quilt?
Speaker 2 (05:08):
yeah, so we've got some history.
Uh, we were founded in 2010 umthat's right the company over 12
years now.
we've, you know, it's been quitea journey and I should say,
before talking about comcast,that, um, you know, one of the
things we recognized early on,because we believe deeply in
(05:28):
this notion of deep partnerships, collaborative partnerships
with service providers that wecan then turn and use as an
infrastructure and network to doall sorts of things today
content delivery.
We were one of the foundingmembers of what was at the time
the Streaming Video Alliance.
(05:49):
It's now the Streaming VideoTechnology Alliance, and we were
very keen to be the catalyst,one of the catalyst companies
for that, because we believethat the industry needed to get
together in a collaborative wayto build, among other things,
the specifications, thestandards for things like open
(06:11):
caching, which is the foundation.
With open caching and our workthere, then the likelihood that
they'll you know as time goes onembrace this, endorse this,
deploy it and support it is thatmuch better and that's proven
(06:33):
to be true.
Speaker 1 (06:35):
Yeah, let's talk about open caching.
Why don't you give you know theone-minute primer on what open
caching is, and I also would bevery interested in having you
tell the evolution of you know.
Explain the evolution of opencaching from when the company
was founded to now, because Ithink it's fair to say and I'm
(06:59):
not going to take words out ofyour mouth but in 2010, networks
looked maybe maybe verydifferent.
Is networks looked maybe verydifferent goes too far, but
looked different.
The way content was deliveredwas different.
In other words, 2010,.
There was a different problemthan there is today in 2024, so
(07:22):
nearly 15 years later.
So explain open caching andthen you know how the technology
and its application has evolved.
Speaker 2 (07:31):
Yeah, happy to do that.
So, and just for background, onQuilt and I'm sure some of your
viewers will know this andunderstand it perhaps better
than I but we began with thisnotion of a product around
transparent caching.
Without going into that indetail, that was a clever way to
help service providers moreefficiently deliver content like
(07:53):
big software downloads andstreaming video in their
networks.
However, that depended onunencrypted traffic and it was
clear to us pretty quickly thatyou couldn't simply depend on
that.
You know streams in the clearthat could be then candidates
for transparent caching.
(08:14):
So with that in mind, and toyour point about open caching,
we had the vision for serviceprovider direct participation in
the value chain of contentdelivery.
Obviously, there were CDNs atthe time Akamai for example who
were doing this in a differentway, and we recognized, partly
(08:36):
because we knew serviceproviders and spent a lot of
time with them, that I'll callit the traditional CDN
architecture and methodologyoffered very little
collaboration with serviceproviders, and it's the way
things have developed.
It's the fault of no one, butthere was a better way.
(08:58):
So I'll just give you theextreme example We've had
recently NFL games in the US,thursday Night Football, for
example.
That's publicly known.
Amazon delivers that throughstreaming.
Peacock delivered famously someNFL games earlier in the year
and then just recently In thetraditional method.
A big CDN, a commercial CDN,would deliver literally millions
(09:25):
of redundant unicast streamsinto the exchange points,
peering points, of a majorservice provider like Comcast.
And those would trace their paththrough the core, out through
distribution and then into theaccess network, into viewers
homes.
(09:45):
That creates a reallyinefficient traffic flow and
even contact.
During the earlier in januarythis year, when the nfl game the
playoff game was was streamedvia peacock.
They said it was the biggestday on the internet for them
ever.
That one event.
Speaker 1 (10:06):
Amazing.
Something like 16 million rightConcurrent.
Speaker 2 (10:11):
Correct, it was the largest streamed event in the US
so far and I'm coming back atopen cash and so sorry for the
journey here, but we recognizethere's a better way and and in
that in in looking at this, wesaid you know, if we could get a
service provider to be aparticipant in value stream,
deploy the caching technologydeep in their network in a very
(10:35):
distributed way and close tousers, we could send and it's
literally a simple you know oneor two Unicast streams into the
network.
They reach those caches and thenfrom their fan out to the
user's homes.
As many as I want to watch.
It could be 16 million, itcould be 100 million for the
(10:55):
super and the offload benefit tothe service provider is massive
, Massive.
99% of the traffic vanishes offthe core.
And they don't have to build abigger network through brute
force.
They can simply depend on opencaching.
We took what was a networktransport problem or use case.
(11:17):
We transformed it into an edgecomputing use case.
Edge computing wins inefficiency and economy every
time, but we recognize that todo this, we have to get the
industry together, sit at atable together, start to
translate and build thespecifications that would allow
(11:38):
this to happen in such a waythat everyone at the table and I
mean by that service providers,content providers and even CDNs
everyone's a participant inthat, such that, when it's done,
everyone can look at it and say, yeah, that's exactly what I
wanted, let's do this, andthat's what happened.
Yeah, it was really.
Speaker 1 (12:00):
I know it was a ton of work.
You know, to build an industryforum is.
I mean, if someone has neverdone it, they just you just have
to do it to understand how hardit is and then not only to
build it but to keep it runningand then to have it achieve its
mission.
You know, I think and you knowwe don't have to talk about all
(12:22):
the failed attempts and thegroups that started and then
kind of flamed out, butcongratulations, because it was
a huge initiative for a vendorto take on and I was fortunate
and honored to be in some ofthose very first In fact I think
I was at the very first meetingand some of the other
(12:43):
subsequent meetings.
First, in fact, I think I wasat the very first meeting and
some of the other subsequentmeetings.
So it was really cool to seehow the SVA and then now the
SVTA, you know, has evolved.
Speaker 2 (12:52):
Yeah, and just to add to that quickly, you know we
had, at the time of theannouncement of the formation,
we had some outstanding foundingcompanies right Among those.
Speaker 1 (13:04):
That was amazing.
Speaker 2 (13:06):
Among those was Comcast right, verizon level
three.
Now Rumen Fox was on board.
Speaker 1 (13:14):
Yeah.
Speaker 2 (13:14):
Baseball.
You may remember Joe, andZarrillo was the-.
Yeah, yeah.
Speaker 1 (13:20):
Of course, of course.
Speaker 2 (13:22):
It's not the result of any one company or one
person's vision.
It's the collective power of agroup saying let's make this
better.
And I should say the SVTA, nowrun by Jason Thiebaud.
It's got a huge remit.
The domain of opportunity anduse cases and problems they're
(13:48):
looking at is pretty vast, andthey've done good work across a
wide variety of issues that theindustry faces, and one of the
unique things we felt about theSBTA is that it wasn't a narrow
audience of people trying tosolve a narrow problem there's
room for that but this was thefull ecosystem saying let's look
(14:12):
at each of these things.
It could be live streaming, itcould be quality measurement, it
could be advertising a reallywide range of things and so I'm
really happy to see howeffective it is yeah, yeah, well
, let's so.
Speaker 1 (14:26):
So tell me this for those who aren't as familiar or
haven't deployed Quilt or, youknow, haven't talked to you
directly, tell us is Quilt asoftware company and you know
you're deploying your, yoursolution, you're licensing your
solution that then an operatoris deploying?
(14:47):
Are you actually deployinghardware?
What is your model?
And you know?
I think that would be good,yeah absolutely.
Speaker 2 (14:55):
Thank you and yeah, so in simple terms, and I'll
describe the method by which wedeploy, maybe that'll help
illustrate.
We're a software and cloudservices company.
So let's first take the side ofthe equation where we're
working with an operator.
We create a partnership likeComcast.
We won't buy the hardware.
(15:18):
We work with Cisco and in somecases they provide their
hardware UCS into the solution,or the service provider can
bring its own hardware.
Either way is fine.
We then, and they deploy thatright, so they provide the space
, the power, the networking, youknow the IP that goes into that
(15:40):
.
They put it as deep in theirnetwork as possible.
This is something we work withthem in advance on the design.
We then put our edge softwarethere OS, and then the edge
applications that we offer.
We web connect that, so weconnect that to our cloud, so we
(16:01):
cloud enable each of thosenodes and just as an example say
, for example, with Verizon,there would be hundreds of those
nodes across their network tosupport.
And then we in every case, forevery operator and there are,
you know, 180 plus operatorsworldwide who support this
today- worldwide who support?
(16:22):
this today.
We unite those or often it'sused, the term's used.
We federate those onto a singleunified platform where the
control plane is in the cloudthat we run and then we can
onboard a content provider.
It's publicly known thatPeacock, for example, is on the
platform and with that singleAPI interface, which is a SVTA
(16:50):
open caching standard, we canbring that content provider onto
the platform and then expose tothem, make available to them,
the universe of federatedservice providers that are using
our open caching technology.
Interesting, yeah, so that'sthe short version.
Hopefully that.
Speaker 1 (17:10):
Okay, one of the things, one of the questions I
have, because it's very much apersonal interest and for NetEnt
, you know we're gettinginvolved more and more as
hardware is making its way.
As hardware is making its way,um, dedicated hardware is making
its way into the video networkand so I'm speaking of, you know
(17:31):
, gpu or vpu to do real-timemuch higher capacity video
encoding, transcoding, that's,um, you know, not bound by a CPU
so much, much more efficient.
Does the open caching standardenable, for example, now,
(17:53):
encoders to be at the edge ofthe network, so that you're
actually even doing the encodingin real time at the edge of the
network?
Can you comment on that?
Is that built into the standard?
Is anybody deploying that?
Are they thinking aboutdeploying it or is it not you?
Speaker 2 (18:10):
know possible.
Well, it's certainly possible.
Let me approach it from acouple of perspectives.
So we don't have anyone who'sdoing that today.
We've had that conversationbecause it's of interest to some
.
For the reasons you're saying,it could be a more efficient way
to do this.
Um, open caching doesn't?
(18:31):
The specifications at themoment don't explicitly address
such a use case, but there'sthere's no reason why it
couldn't be.
Speaker 1 (18:38):
You know, part of the suite of use cases that are in
other words, rather than, ratherthan pulling a stored uh,
either a stored asset or a astream being the input, it could
, literally the input could bean encoder.
Right, it's just sitting thereand it's producing.
You know it's outputting theappropriate files in real time,
(19:00):
you know the appropriate encodesfor whatever device.
Speaker 2 (19:03):
You know it's is uh being supported, I guess, or
devices uh, yeah, yeah, exactlyright, and, as I said that there
have been a number of peoplewho've talked about it with
interest, um, yeah but the otherone just throwing you know new
ideas.
Uh uh, gaming companies havetalked about doing rendering at
the edge.
I think it's kind of a similarnotion which is, yeah, you can
(19:25):
do it there more efficiently andfaster, so these things are
coming, I think and there's noreason why open cache and can't
support you, you, you, no doubt.
Speaker 1 (19:37):
uh, netflix knows who you are, but you know you
should be talking to netflixbecause obviously it's very
public.
They are building a cloudgaming network and it would just
seem to me with their OpenConnect platform, which of
course they write a lot about.
So this is all public is kindof the largest edge distributed
(20:00):
network, probably in thebusiness, and maybe not, but at
least with my limitedunderstanding.
So, yeah, very interesting welllet me comment on that just
generally.
Speaker 2 (20:13):
This is not about netflix specifically, but just
generally.
One of the things that we'veheard from service providers
telcos, broadband providers,cable providers, mobile
providers is this idea that, um,a lot of people, a lot of and
you can imagine, I'm sure youknow them all but a lot of
companies some hyperscalers,others with streaming services
(20:36):
want to get into those networksand want to be deployed deeply
with their, I'll say devices,their edge delivery devices.
But it seems to be a consensusamong service providers that
they're saying, yeah, I can't dothat for 10, 20, 30 people.
It would be a nightmare tomanage.
That's a lot of hardware.
(20:56):
That's exactly right, and so atleast several that we've spoken
to, who are major serviceproviders, are saying we'd like
to have this capability on aunified platform from one
partner that we can then offerto all, and and that's at least
(21:18):
um a sense that we're gettingabout open caching, because it,
as we've already said, it's it'sa industry specification, yeah,
adopted as a standard by theitf, uh, and it's open, and so
so this is something that wethink works in our favor so far.
Speaker 1 (21:37):
Yeah, it's great.
Well, this is the perfectlead-in, because you referenced
the Comcast announcement thatcame out.
That was Tuesday, I believe, sowe're recording this on
Thursday, so a couple days agoago now.
Congratulations, uh, uh.
We even have some listeners whohave, uh, who have posed some
questions.
So, um, I I want to talk aboutthis, and one of the questions
(22:02):
that I have and and I saw in thechat, is you know, maybe, maybe
you can explain in layman'sterms, you know, how, how this
is going to improve thestreaming experience for Comcast
customers?
And, like, my question is sortof a follow-up is is this a
Comcast customer of the Cape?
(22:22):
Because you know one of thethings we always, when we say
Comcast, are we talking thecable network, are we talking
the you know the IP over QAM, orare we talking the open
Internet?
So maybe you can even clarifythat.
But just in layman's terms, howis this going to improve the
experience?
Comcast didn't do this justbecause they needed a press
(22:43):
release, so there's a reasonthat they're adopting.
Speaker 2 (22:50):
Yeah, and that's exactly right.
Um, and you can imagine now,they were one of the founding uh
members of the streaming videotech.
Speaker 1 (22:58):
I remember that yeah, so we we've got some.
You've had a long deep.
We spent time with them, youknow but very specifically.
Speaker 2 (23:06):
So this is comcast, the you know cable provider okay
, okay.
Interesting.
It's the network, yes, and sothey're a national ISP three
plus million broadband users.
They're very, very big in theUS, one of the biggest cable
providers in the world, and, asI mentioned earlier that we will
(23:31):
have compute hardware placeddeep in their network in a
highly distributed architecture.
It will be the most distributedCDN if you will ever.
And on that hardware that theyprovide and they run and own and
(23:52):
operate, we will put oursoftware, we will cloud connect
it and then we will provide thecontrol plane that unites that
into the rest of this federatednetwork I've been talking about.
So why did they do this?
I think and there's a sorry forthe just like straight up
marketing promotion here, butthere's a great video on our
site in which our CEO, alainMayor, interviews network
(24:15):
officer Elad and he talks aboutyou know the dynamics, the
forces that are driving them tothis kind of outcome, a
partnership with us, and theyhave other announcements about
virtualization.
But specifically, they'retrying to activate their edge
(24:35):
because they know it's valuableand it's of interest to many in
the industry, not just streamingproviders In such a way that
it's unified, that it can bemade available to anyone who
wants to make use of it, to makeuse of it, and it would be then
used for a whole range of usecases, not just streaming, but
potentially gaming.
(24:56):
Other consumer use cases.
There are enterprise use casesright that, for example,
networking, security,collaboration, iot, connected
cards.
There's a really broad range ofedge computing use cases that I
think.
Well, in their case, I think wesee with them coming and their
(25:21):
goal is to create the foundationon which all that stuff can be
built and run efficiently.
Speaker 1 (25:27):
And so this is hopefully that's helpful.
This is what's driving, I think, their effort here.
Can you quantify?
You said it's the largestdistributed CDN or network.
What does that mean?
Is that based on number ofmachines?
Is it number of locations?
Is it capacity?
What exactly does that mean?
The largest?
Speaker 2 (25:47):
Yeah thank you, it's good to be clear about this.
So, in our view, when we talkabout distributed, we're
referring to some people call iton net distribution.
We often refer to it as deeplyembedded caches of who has the
(26:09):
most deeply embedded caches.
I'm putting aside capacity forthe moment, because we have more
than enough capacity to deliveragainst the demand we and the
customer demand we have today,but it's the proximity and the
distribution that gives usadvantage, right?
So we're inside the network,we're downstream of peering
points, exchange points that canbe congested, but, more
(26:31):
importantly, we're deep in thenetwork, over on the access side
, past the core.
We're downstream of peeringpoints, exchange points that can
be congested, but, moreimportantly, we're deep in the
network.
You know, over on the accessside, past the core, past
aggregation, sitting on theaccess side.
So there's nothing closer,there's nothing with more
proximity than we have.
And if you were to aggregateall of these in this federated
CDN that we have, all of thesedeeply embedded caches, nobody
(26:54):
has.
This is thousands of cachesthat we have.
Yeah.
Speaker 1 (26:59):
Yeah.
Speaker 2 (27:01):
I'm trying to.
Speaker 1 (27:03):
Okay, yeah, yeah, definitely that's helpful.
You know, I recall that Comcastrecently launched an enhanced
4K service, and maybe that wasaround the Olympics, I'm not
remembering now exactly, but I'mjust curious is there any
(27:24):
relation there or can you tipyour hat towards?
Is this now going to enablesome, shall we say, say, higher
quality experiences?
Because, let's face it, whenyou talk about, you know whether
it's quote enhanced 4k or, youknow, xr, vr, even gaming for
that matter the big challenge,the big hurdle, is really
(27:48):
network capacity.
At the end of the day, um, soyou know, people say, you know
how come.
You know service x, service y,service z, you know they don't
have more 4k or they don't.
You know, you know they're notstreaming with better quality
and it's like, well it?
It's pretty simple it comesdown to capacity and cost you
know.
(28:09):
So yeah, I curious.
Can you comment on?
Is there any you know?
Were you an enabler to that oryou know that announcement?
Speaker 2 (28:18):
Well, 4k is one of the use cases that's enabled, if
you will, by what we're doing.
Absolutely, you know.
I don't think it was the driver, for the reasons I've already
said.
They have a vision for edgecomputing that transcends 4K or
any one thing.
Speaker 1 (28:37):
But yeah, 4k would be an example.
Speaker 2 (28:39):
But back to your earlier point, and I completely
agree capacity is the scarceresource here.
As we look at what's happeningwith the use of internet and
streaming, just look narrowly atlive streaming.
Uh, you know, we.
We say 16 million live streams.
That's a us record.
You know the the super bowl isa hundred million users.
(29:02):
Do we want to somedayexclusively stream that event?
Maybe, and so it's.
You're exactly right, the, thecapacity to do it.
It becomes the constraint and,as I said, that's what we solve
for when we work with serviceproviders, like we do, because
we're offloading, you know, 99%of the core traffic associated
(29:23):
with a big event.
That's the way you get to 100million.
I mean big events, right, thecricket finals, that's a billion
viewers, and if they all wantto stream it, that's a problem.
Speaker 1 (29:39):
Yeah, definitely.
Well, I am looking here in thechat and there are a number of
questions that came in and thenare coming in, so why don't we
switch here?
It continues.
Many of them are in thedirection we're already talking,
so yeah, so okay, what are themain gaps?
(30:02):
To bridge the gap of the bigmigration of broadcast to IP?
Ip Boy, this is an interestingquestion, but you're working
primarily with service providers, so I think you're in a good
position.
Just comment generally.
Speaker 2 (30:20):
Yeah, generally and there's several parts to this at
least that we see so many ofthese big service providers,
cable companies, for example,they're already going to all IP
services, right.
So, for example, I mean I'm aComcast customer at home.
When I watch Xfinity, I'mstreaming it, right, it's not
(30:42):
the broadcast signal that I'mgetting, I'm streaming from
their platform.
Others are doing the same thing, and when we're deployed in one
of these networks, and whenwe're when we're deployed in one
of these networks, we cansupport that streaming of their
managed content.
We do that today with Verizon,fios TV, for example.
So we help enable that.
I'm kind of doing thisresponding in part to your, to
(31:04):
your question.
The larger question of themassive transformation from
broadcast to IP is certainlyunderway.
We see those trends, but itwill take time and, as I said,
(31:26):
capacity is one of theconstraints that needs to be
solved for to get us to theother side.
You know it'll take years, forsure, right, for everyone to
stop using broadcast televisionand start streaming, um, but but
it's, it's an unstoppable forcein the industry, it's
inevitable, and so I think it'sjust.
You know, so I think I, I Iwould.
Speaker 1 (31:46):
I would like you to comment, though, because I think
the question, uh, is okay.
All all of what you said istrue and you know the trends and
, like you say a lot really, allof these networks, they're
already IP networks, so you knowone of my.
So my initial response to aquestion like this is like well,
what gap are you talking about?
(32:08):
Because, fundamentally, otherthan I don't know, you know some
really remote regions of theworld, or you know where there
still are significant legacybuild-outs, like I don't know in
the Americas who isn't runningan IP network, you know.
So now that's my response.
I'm sure others, who are muchcloser to the details, could say
(32:30):
, oh, but Mark, you know.
I'm sure others who are muchcloser to the details could say,
(32:59):
oh, but Mark, controllednetwork, my closed network, and
then I have my inner theinternet right, because that is
still true and even insideComcast, right, you know there,
there there's sort of twonetworks.
I mean not sort of.
There are two networks you knowthat you interact with.
So I'm wondering do you have anyperspective of you know like at
what point, or will they nevermerge for whatever technical or
(33:23):
political reasons, or you knowwhatever.
Speaker 2 (33:25):
Well, again, I think it's going to be a long, long
time before somebody shuts offthe broadcast network at one of
these big companies.
Speaker 1 (33:35):
For example.
Speaker 2 (33:35):
The other gap, though just to continue on the list
would be solving for ultra-lowlatency delivery.
So we want the signal, that'sright.
If someone's streaming toarrive at the same time the
broadcast signal does, so thatyou know, the classic problem of
social media spoiler doesn'ttake place.
I think we're getting there andwe're working with one of the
(33:59):
largest streaming platforms inthe world on a ultra low latency
protocol, awesome UDP basedthat is delivering the signal
within you know, a few seconds,five seconds.
So we're starting to see thatpart of it being, because it
used to be right, maybe 60seconds, which was an
(34:21):
unacceptable delay, and peoplewouldn't consider that if they
had you know, money riding onthe game, but we're getting
there on that front.
But yeah hopefully the problem,you know, the transformation,
the migration.
It'll take time.
Speaker 1 (34:37):
Yeah, yeah, for sure, okay.
Well, a question that came inwas about WebRTC and
specifically the person isasking you know, well, I'll just
read the question, because it'spretty well formed what do you
think about the future ofstreaming via WebRTC and then
its impact on CDNinfrastructures?
(34:59):
And they go on to say andtranscoding.
So yeah, webrtc, is it playinga factor?
Will it never play a factor?
Is everything going to go toWebRTC?
What's your view there?
Speaker 2 (35:17):
Yeah, I mean, I think this is promising.
I'm afraid I'm going to answerit in the more general case,
because I've used the example ofone of our customers who's a
global streaming platform andthey're using a UDP-based stack
for live streaming.
That gets us to a few secondsof delivery, that is, latency
(35:38):
between the broadcast and thestream.
If it's going to be WebRTC,great, we can support all of
this.
So we're looking for you knowwhat our customers want, as
opposed to for us, you know,promoting a particular path or a
very specific stack.
(35:59):
Beyond that, I, you know, Iwill see, we'll see where this,
where this goes.
Speaker 1 (36:05):
Yeah, exactly, exactly it's.
It's a really exciting space.
It's a really exciting spaceand I think you know.
Just my personal commentary isI think one of the challenges
excuse me that we have in theindustry and when we're having
these sorts of conversations isthat there is no sort of hand
(36:25):
wave.
You know and I'm picking onWebRTC here but there almost
never is a scenario where it'slike, well, webrtc is going to
solve that, or XYZ is going tosolve that, or oh, this is going
to you know, sort of that.
And oftentimes I find I'm inconversations where questions
are framed or where it's kind oflike.
(36:47):
It's like, oh, but isn't youknow, fill in the blank,
technology going to solve that.
And it's like, well, let's zoomout.
And first of all, you have tounderstand the incredibly
complex architectures of howcontent is delivered from you
(37:09):
know, from origin, from source,from you know, whatever the
topology, whatever the topologyis, to the viewer.
In one scenario, it actuallymight be true that WebRTC
largely is going to be the waythe content is delivered, but
for that very same stream, therecould be another scenario based
(37:30):
on a different geographicproximity and the networks that
are involved and the devicesthat are involved, that WebRTC
will never work, you know, andthere can be a third, and
there's a fourth and a fifth anda sixth.
So I make this observation morefor listeners who are listening
, you know, to conversationslike this or just having other
(37:50):
conversations throughout theindustry, and it's very natural
and it's perfectly reasonable tobe really leaning in like, hey,
what technology should we belooking at, building on, maybe
even doubling down on?
But regretfully, for all theyears that I've worked in the
space, I have never seen, and Idon't see in the future, a
(38:14):
scenario where a technology iswant to debate.
You know, feel free to commenton LinkedIn, you know, let's
(38:38):
talk about it.
So, all right.
Speaker 2 (38:41):
Let's move on, Completely agree.
Speaker 1 (38:42):
Mark, yeah, yeah, yeah, Okay, let's move on.
So all right.
So here's an interestingquestion.
All right, so here's aninteresting question how can
hundreds of gigabytes of data beused for real-time analytics?
What can the best turnaroundexpect?
(39:02):
How can hundreds of data beused for real-time analytics?
Speaker 2 (39:24):
Do you know what this question means use edge
computing as a resource tohandle processing of data in
proximity to the device that'sproducing it, because that's
more efficient and moreresponsive than centralized
cloud computing?
Yeah, the answer is verypossibly yes, and at least there
(39:48):
are definitely use cases aroundedge computing where the
assumption is we want to processanalytic data from a device as
close to that device in thecloud edge cloud as close to the
device as possible.
Uh, so so?
we're not doing that todaythat's not a current um use case
we're supporting, but I thinkit's okay there, in theory it's
it's one that's possible yeah Idon't know if that's the
(40:10):
question okay, all right, well,um, well, feel free.
Speaker 1 (40:13):
I know Mark is very accessible on LinkedIn.
I'm accessible on LinkedIn, soyou know.
If we don't hit your questionor if you have a question that
you weren't able to raise, yeah,hit us up.
Ok, are you developing?
This person wants to know ifyou're developing the Gulf
Cooperation Council countries,so I think it's like Saudi
(40:35):
Arabia, bahrain.
So yeah, are you you know?
Do you have any deploymentsthere or anything coming up, or
can you hint something coming up?
Speaker 2 (40:47):
Yeah, I can say what's public, which is we have.
We announced some time ago apartnership with STC, saudi
Telecom.
Speaker 1 (40:53):
Okay.
Speaker 2 (40:54):
And we are deployed there and they have announced
again publicly that they aredelivering content throughout
the region.
I can't name exactly thecountries within that region,
but they're referring broadly tothe Middle East.
So yes, we're there with thatpartnership and it're referring
broadly to the Middle East.
So so, yes, we're there withthat partnership and it's active
(41:15):
.
Speaker 1 (41:15):
Awesome, yeah, yeah, that's great, that's great.
Okay, um, well, let's, uh,let's, you know, let's wrap up,
um, where I have a two-partquestion for you and I won't do
the usual, you know, crystalball.
Where are we going to be infive years?
(41:36):
I, I, personally, I hate thosequestions, yeah, but but I do
have a two-part question andit's and, and I'll start I'll.
I'll start with the first partand that is where.
How are networks?
You know, I referenced it whenyou started the company, when
(41:57):
Quilt was founded in 2010.
It was.
You know, network architectureslook different.
You know, largely.
You know 95% of, or what youknow.
I'm just guessing.
I don't know the real number,but it's probably in that range
of content being streamed.
It was file based, it was bod.
You know, there was actually,um, relatively speaking, very
(42:21):
little live.
Uh, I mean, there was livevideo.
I'm not saying there wasn't,but you know very little.
Um, certainly, that hasn't beenflipped.
I mean, still, the majority ofcontent is still file-based,
it's still VOD.
There's an asset that's sittingon an origin server somewhere
(42:42):
and being distributed, but let'sface it, you see, netflix who,
just a couple of years ago, Ican remember famously, this was
more than a couple of years ago.
You know I can rememberfamously this was more than a
couple of years ago.
But asking somebody whose nameI won't reveal, but somebody we
all know very well insideNetflix, at that time, you know,
(43:02):
why doesn't Netflix do live?
And the response that you knowthat I got back was I mean, it
was as firm as anything Netflixwill never do live.
All about the VOD, you know,sort of like ad.
You know Netflix will never doadvertising.
And then, next thing, you knowNetflix is doing advertising
very much and you know so, ontransforming, you know um, over
(43:34):
the next, you know, say, threeto five years, so kind of the
medium term horizon, uh, withthis really decided move to like
live live streaming.
You know, and and it's notgoing anywhere.
It's not like this is a fad.
Speaker 2 (43:45):
This isn't 3d video, remember that yeah, by the way,
the first part of my question isgot.
Speaker 1 (43:53):
I got it, where are we?
Speaker 2 (43:54):
going the.
The netflix comment lives nowin the archive with apple's
comment they would never sell aphone, right?
Speaker 1 (44:01):
yeah, exactly, yeah yeah.
Speaker 2 (44:03):
So you know I I think , and just looking at at
streaming and video, um, youknow things are happening there.
Live is is more and more a partof the fabric.
Exclusively live streamedevents are that way.
I mean you just look at this ispublic around Amazon.
But look at what they're doingwith Premier League.
(44:24):
I mean I don't think there's asport at the moment that they
aren't live streaming.
They're securing deals toexclusively stream events across
the spectrum of sports.
There is live linear right, sothere are fast services that are
increasing as well.
We see that taking off.
(44:44):
I think, just sort of generallyspeaking, we will see this
continued migration andpreference from consumers.
I think that's the mostimportant thing, never mind what
platforms are promoting what doconsumers like.
There's a whole host of reasons, I think, for consumers to
prefer to stream an event eithervideo on demand, but certainly
(45:10):
a live event rather thanwatching a broadcast television.
You know mobility, flexibility,it's extraordinary and the
selection of content is justkind of endless.
So I think we'll see that justcontinue to take off.
There are things like 4kstreaming, as you mentioned.
There's immersive experiences,you know AR and VR, those kinds
(45:31):
of things, maybe some of thembeing promoted by Meta and
others.
That, I think will you knowwill potentially see get more
traction.
I'll just say too, and this ismaybe the more boring part, we
see enterprise use cases as alsopicking up here where, as I
(45:53):
mentioned before, collaboration,iot, connected car security,
networking those will consumeresources at the edge, and
that's the kind of thing weexpect.
Speaker 1 (46:05):
So what does this mean, though, for how networks
are built?
Um, because that that's reallya heart of my, of my question,
and I'll and and I'll, and I'lladd another layer of context.
Um, I, I guess, a couple yearsago now, akamai bought linode.
Right now, uh, and they're, andthey've been very active out
(46:26):
promoting, uh, you know the theedge, edge computing.
I mean linode is their.
You know cloud edge, edgecomputing, I mean Linode is
their.
You know cloud edge computingplatform, and I mean we both
just came back from IBC.
You know a lot of.
You know Akamai is out thereand, by the way, other networks
too.
It's one of the trends that I,dan Rayburn and I talked about
(46:48):
in a couple of days ago in hislatest podcast episode.
Was this trend?
So then, my question is okay,if Akamai is building compute
into the edge and you knowthey're positioning it a little
bit different, I mean, you knoweverybody has their angle right,
but at the end of the day, thisis edge computing, you know at
(47:09):
the end of the day, you knowthat's what they're doing.
They're making hardwareavailable in a distributive
fashion that an Akamai customerand one can assume Tata and kind
of go down the list of all theobvious that they're probably
either doing it, are doing it,planning to do it, and just
haven't announced it this iscoming.
(47:30):
Are doing it, planning to do itand just haven't announced it.
You know like this is coming.
So what does this mean for theway that CDNs and networks are
built and the architectures?
You know?
What are your thoughts?
Speaker 2 (47:43):
on that.
Well, yeah, I mean, and Linodeis a good example to bring up,
at least from public statements.
They were, linode was maybe20-something sites.
At the time of the acquisition,akamai paid, I think, $900
million for the company.
They've announced plans sincethen to roll out.
(48:03):
Akamai has announced plans toroll out to 100-plus locations
with that technology, with thatedge computing stack from Linode
, with that technology, withthat edge computing stack from
Linode.
Back to the point about networks.
It seems to me that if you're,you know, really focused in on
(48:25):
edge computing, there's adefinition there that plays
right.
So it could be a premise edge,the device where some of those
edge computing use cases arerunning, because that's the most
efficient place to do it.
But it's not no longer thecentralized cloud.
We know that.
That's understood.
What the hyperscalers do webelieve that back to your point
about networks that it's theservice provider who has
(48:49):
unparalleled proximity scalereach service provider who has
unparalleled proximity scalereach that has to activate in
their network an edge computingcapability and it should be as
deeply embedded, as close to theusers as possible.
That's where we come in and Ithink others, if you look at the
hyperscalers Azure, aws, I meanthey do have a kind of edge
(49:12):
ambition and in many cases theybelieve that should be sitting
inside a service providernetwork.
Speaker 1 (49:20):
We agree with that.
Speaker 2 (49:21):
But you know, the service provider has to be a
collaborative partner if that'sgoing to happen.
And that's at the heart of whatwe do really well.
You know, you mentioned Alkama.
It's not clear to me whetherthey can really activate that in
the same way.
I think that's the, you know,the mystery that has to be
solved.
Yeah, it's fascinating.
Speaker 1 (49:44):
It's fascinating.
I think we can end on this note.
I won't really pose it as aquestion, but I'll give you a
chance to you know, either sayyes, you know, or comment
further.
But it seems like where Quiltis positioned really well is
that you know you're able tobring together the constituents
(50:07):
who you know.
There's obviously technologycomponents.
There are technology componentsthat you bring that are needed,
right, that a service providerneeds to be able to deliver.
You know, throughout this, youknow, as I used the word earlier
morass of networks andarchitectures, right,
(50:51):
no-transcript, it's public andwell-known.
You know Netflix has done anamazing job of building out
OpenConnect.
I mean, I think the last publicnumber, like whatever 80,000,
you know machines or pops or Idon't even know how it's defined
, but it's a huge number thatthey have deployed.
They've done an amazing job.
But that was hard and there'sstill networks who are like yep,
(51:15):
not in our network, you know,and so that's a challenge.
You reference Akamai.
You know like, like ah, youknow there's some where it's
kind of like oh, we're not tooexcited about maybe getting, you
know, having having acommercial CDN inside our
network.
So is this is mycharacterization correct where
(51:36):
you know you're bringingtechnology, but you're also able
to bridge some of these gaps.
Is that, you know, either toosimplistic a view or is that an
incorrect view?
Speaker 2 (51:47):
No, Mark, you're spot on.
Yeah, you're spot on and I'mglad you raised this point A
good way to maybe close up.
But yeah, the tech we bring ispowerful and open caching as an
open specification is powerfulfor the reasons we've done
really well.
(52:08):
That sets us apart is recognizeearly on the need to
collaborate deeply with,cooperatively with, service
providers and bring that, bringthe ecosystem, the value chain
together in that way.
I should say, quilt of theinspiration for the name was, to
this very point, the idea ofstitching together a fabric so
(52:30):
that it comes becomes a whole, acomplete, and I really do think
we've done that better thananyone else.
You know can witness theComcast announcement this week.
I think, that speaks to ourability here.
Speaker 1 (52:42):
Yeah, that's great.
I think it's a great way to endQuilt love the name.
You know you're stitchingtogether all of these, um, these
hard things and ultimatelyenabling the ecosystem to
deliver content, you know, moreefficiently, and you know that
has a positive impact on cost,positive impact on user
(53:04):
experience.
Uh, you know, just has apositive impact.
So, hey, mark, um, thank youfor joining me.
Uh, we went a bit long, but youknow, as I told you before, we
pushed record or pushed go live.
You know the way we run this isif the conversation's rolling,
we just keep rolling and at thepoint that you know it kind of
(53:27):
naturally winds down, then wewrap it up.
So really appreciate your timeand thank you for coming on.
Thanks.
Speaker 2 (53:34):
Mark, thanks for having me Take care.
This episode of Voices of Videois brought to you by NetInt
Technologies If you are lookingfor cutting-edge video encoding
solutions check out NetInt'sproducts at netintcom.
voices of video.
Voices of video.
The voices of video voices ofvideo okay, well, good morning
everyone.
We are back for another veryexciting edition of Voices of
Video, and I have with me todayMark Fisher from Quilt.
(00:29):
So first of all, mark, welcometo Voices of Video.
Speaker 2 (00:34):
Thanks, mark, good to be here with you.
Speaker 1 (00:36):
Yeah, yeah, so we were talking before we started,
you know, before we opened thesession, and it's been 10 years
since we crossed paths Now Ithink maybe it was a few years
before that.
Even We'll get into exactlywhat we're referencing, because
we're going to talk about thiskind of a cool thing that you
(00:57):
know, Quilt has been doingbeyond building your product.
So, anyway, a little teaserwe're dangling out there, right?
But, yes, welcome to the show.
And I just want to say foreverybody listening live, if you
have questions, this is afireside chat type format and we
(01:21):
want to get to your questionsand we're very happy to take
really any question and if wecan't answer it, we'll just say
so, right, yeah, so so feel freeto just, you know, type in some
questions in the chat and ourproducer, anita, will pick those
(01:41):
up and then you know we'llwe'll do our best to get to them
.
So with that you know, mark,why don't you say a few words
about who you are and what youdo at Quilt?
And then you know, I wouldn'tassume that everybody knows
exactly who Quilt is and whatyou do.
So let's start there.
Speaker 2 (02:04):
Yeah, great, and thanks for the invitation, mark,
to join you today.
So I've been working in variousstartups now for a number of
years, maybe tracing or trackingwith the big trends across the
internet Early 90s, thecommercialization of the
internet, internet serviceproviders, then to the mobile
(02:27):
Internet with a company, danger,that was acquired by Microsoft,
then to Internet of Things, acompany, jasper, which was
acquired by Cisco, and now atQuilt.
I'm in every case and now I'mresponsible for marketing
business development.
And now I'm responsible formarketing business development.
It's been, you know, a greatexperience across the board, and
(02:47):
what makes it great are thepeople that you are with, that
you work with, whether they'rein the company where you work or
across the industry.
That, for me, has always beenvery rewarding.
Yes, so Quilt and I agree withyou, not everyone would know us.
We obviously had big news thisweek, an announcement of a
partnership with Comcast, whichwe're very proud of and was
(03:09):
years Absolutely Spend some timetalking about that for sure.
So Quilt is a software and cloudservices company.
We're unique because we reallyare focused on edge computing as
a general category, but what wedo to go to market is unique.
(03:33):
In particular, we have a verystrategic relationship with
Cisco, who's an investor, andwith them we go out and partner
with big service providers andsometimes smaller service
providers, and we do that to putour software on compute
(03:55):
hardware deep inside theirnetworks, close to consumers or
businesses.
And once that's done, thatinfrastructure is in place, that
edge cloud, service provider,edge cloud then we turn, we
unite it, we federate it, if youwill, with our cloud services.
And then we turn to today bigcontent providers, some of the
(04:17):
biggest stream platforms in theworld, and we say, hey, we've
got a better way to deliver yourcontent live streams, vod or,
in some cases, just softwaredownloads.
And we can demonstrate thedifference by our performance,
where we routinely outperformevery other commercial CDN
because we have this advantageof being closer.
(04:38):
It's sort of that simple.
At the beginning of what we'redoing, we're taking that use
case as the foundation.
That's what gets us in the doorwith these big service
providers, like a Comcast whowe've announced this week, and
then we're going to turn ourattention to other edge
computing use cases.
We can talk about that as timegoes on, yeah, yeah, amazing.
Speaker 1 (05:03):
Remind me, by the way , when was the company founded?
What year was quilt?
Speaker 2 (05:08):
yeah, so we've got some history.
Uh, we were founded in 2010 umthat's right the company over 12
years now.
we've, you know, it's been quitea journey and I should say,
before talking about comcast,that, um, you know, one of the
things we recognized early on,because we believe deeply in
(05:28):
this notion of deep partnerships, collaborative partnerships
with service providers that wecan then turn and use as an
infrastructure and network to doall sorts of things today
content delivery.
We were one of the foundingmembers of what was at the time
the Streaming Video Alliance.
(05:49):
It's now the Streaming VideoTechnology Alliance, and we were
very keen to be the catalyst,one of the catalyst companies
for that, because we believethat the industry needed to get
together in a collaborative wayto build, among other things,
the specifications, thestandards for things like open
(06:11):
caching, which is the foundation.
With open caching and our workthere, then the likelihood that
they'll you know as time goes onembrace this, endorse this,
deploy it and support it is thatmuch better and that's proven
(06:33):
to be true.
Speaker 1 (06:35):
Yeah, let's talk about open caching.
Why don't you give you know theone-minute primer on what open
caching is, and I also would bevery interested in having you
tell the evolution of you know.
Explain the evolution of opencaching from when the company
was founded to now, because Ithink it's fair to say and I'm
(06:59):
not going to take words out ofyour mouth but in 2010, networks
looked maybe maybe verydifferent.
Is networks looked maybe verydifferent goes too far, but
looked different.
The way content was deliveredwas different.
In other words, 2010,.
There was a different problemthan there is today in 2024, so
(07:22):
nearly 15 years later.
So explain open caching andthen you know how the technology
and its application has evolved.
Speaker 2 (07:31):
Yeah, happy to do that.
So, and just for background, onQuilt and I'm sure some of your
viewers will know this andunderstand it perhaps better
than I but we began with thisnotion of a product around
transparent caching.
Without going into that indetail, that was a clever way to
help service providers moreefficiently deliver content like
(07:53):
big software downloads andstreaming video in their
networks.
However, that depended onunencrypted traffic and it was
clear to us pretty quickly thatyou couldn't simply depend on
that.
You know streams in the clearthat could be then candidates
for transparent caching.
(08:14):
So with that in mind, and toyour point about open caching,
we had the vision for serviceprovider direct participation in
the value chain of contentdelivery.
Obviously, there were CDNs atthe time Akamai for example who
were doing this in a differentway, and we recognized, partly
(08:36):
because we knew serviceproviders and spent a lot of
time with them, that I'll callit the traditional CDN
architecture and methodologyoffered very little
collaboration with serviceproviders, and it's the way
things have developed.
It's the fault of no one, butthere was a better way.
(08:58):
So I'll just give you theextreme example We've had
recently NFL games in the US,thursday Night Football, for
example.
That's publicly known.
Amazon delivers that throughstreaming.
Peacock delivered famously someNFL games earlier in the year
and then just recently In thetraditional method.
A big CDN, a commercial CDN,would deliver literally millions
(09:25):
of redundant unicast streamsinto the exchange points,
peering points, of a majorservice provider like Comcast.
And those would trace their paththrough the core, out through
distribution and then into theaccess network, into viewers
homes.
(09:45):
That creates a reallyinefficient traffic flow and
even contact.
During the earlier in januarythis year, when the nfl game the
playoff game was was streamedvia peacock.
They said it was the biggestday on the internet for them
ever.
That one event.
Speaker 1 (10:06):
Amazing.
Something like 16 million rightConcurrent.
Speaker 2 (10:11):
Correct, it was the largest streamed event in the US
so far and I'm coming back atopen cash and so sorry for the
journey here, but we recognizethere's a better way and and in
that in in looking at this, wesaid you know, if we could get a
service provider to be aparticipant in value stream,
deploy the caching technologydeep in their network in a very
(10:35):
distributed way and close tousers, we could send and it's
literally a simple you know oneor two Unicast streams into the
network.
They reach those caches and thenfrom their fan out to the
user's homes.
As many as I want to watch.
It could be 16 million, itcould be 100 million for the
(10:55):
super and the offload benefit tothe service provider is massive
, Massive.
99% of the traffic vanishes offthe core.
And they don't have to build abigger network through brute
force.
They can simply depend on opencaching.
We took what was a networktransport problem or use case.
(11:17):
We transformed it into an edgecomputing use case.
Edge computing wins inefficiency and economy every
time, but we recognize that todo this, we have to get the
industry together, sit at atable together, start to
translate and build thespecifications that would allow
(11:38):
this to happen in such a waythat everyone at the table and I
mean by that service providers,content providers and even CDNs
everyone's a participant inthat, such that, when it's done,
everyone can look at it and say, yeah, that's exactly what I
wanted, let's do this, andthat's what happened.
Yeah, it was really.
Speaker 1 (12:00):
I know it was a ton of work.
You know, to build an industryforum is.
I mean, if someone has neverdone it, they just you just have
to do it to understand how hardit is and then not only to
build it but to keep it runningand then to have it achieve its
mission.
You know, I think and you knowwe don't have to talk about all
(12:22):
the failed attempts and thegroups that started and then
kind of flamed out, butcongratulations, because it was
a huge initiative for a vendorto take on and I was fortunate
and honored to be in some ofthose very first In fact I think
I was at the very first meetingand some of the other
(12:43):
subsequent meetings.
First, in fact, I think I wasat the very first meeting and
some of the other subsequentmeetings.
So it was really cool to seehow the SVA and then now the
SVTA, you know, has evolved.
Speaker 2 (12:52):
Yeah, and just to add to that quickly, you know we
had, at the time of theannouncement of the formation,
we had some outstanding foundingcompanies right Among those.
Speaker 1 (13:04):
That was amazing.
Speaker 2 (13:06):
Among those was Comcast right, verizon level
three.
Now Rumen Fox was on board.
Speaker 1 (13:14):
Yeah.
Speaker 2 (13:14):
Baseball.
You may remember Joe, andZarrillo was the-.
Yeah, yeah.
Speaker 1 (13:20):
Of course, of course.
Speaker 2 (13:22):
It's not the result of any one company or one
person's vision.
It's the collective power of agroup saying let's make this
better.
And I should say the SVTA, nowrun by Jason Thiebaud.
It's got a huge remit.
The domain of opportunity anduse cases and problems they're
(13:48):
looking at is pretty vast, andthey've done good work across a
wide variety of issues that theindustry faces, and one of the
unique things we felt about theSBTA is that it wasn't a narrow
audience of people trying tosolve a narrow problem there's
room for that but this was thefull ecosystem saying let's look
(14:12):
at each of these things.
It could be live streaming, itcould be quality measurement, it
could be advertising a reallywide range of things and so I'm
really happy to see howeffective it is yeah, yeah, well
, let's so.
Speaker 1 (14:26):
So tell me this for those who aren't as familiar or
haven't deployed Quilt or, youknow, haven't talked to you
directly, tell us is Quilt asoftware company and you know
you're deploying your, yoursolution, you're licensing your
solution that then an operatoris deploying?
(14:47):
Are you actually deployinghardware?
What is your model?
And you know?
I think that would be good,yeah absolutely.
Speaker 2 (14:55):
Thank you and yeah, so in simple terms, and I'll
describe the method by which wedeploy, maybe that'll help
illustrate.
We're a software and cloudservices company.
So let's first take the side ofthe equation where we're
working with an operator.
We create a partnership likeComcast.
We won't buy the hardware.
(15:18):
We work with Cisco and in somecases they provide their
hardware UCS into the solution,or the service provider can
bring its own hardware.
Either way is fine.
We then, and they deploy thatright, so they provide the space
, the power, the networking, youknow the IP that goes into that
(15:40):
.
They put it as deep in theirnetwork as possible.
This is something we work withthem in advance on the design.
We then put our edge softwarethere OS, and then the edge
applications that we offer.
We web connect that, so weconnect that to our cloud, so we
(16:01):
cloud enable each of thosenodes and just as an example say
, for example, with Verizon,there would be hundreds of those
nodes across their network tosupport.
And then we in every case, forevery operator and there are,
you know, 180 plus operatorsworldwide who support this
today- worldwide who support?
(16:22):
this today.
We unite those or often it'sused, the term's used.
We federate those onto a singleunified platform where the
control plane is in the cloudthat we run and then we can
onboard a content provider.
It's publicly known thatPeacock, for example, is on the
platform and with that singleAPI interface, which is a SVTA
(16:50):
open caching standard, we canbring that content provider onto
the platform and then expose tothem, make available to them,
the universe of federatedservice providers that are using
our open caching technology.
Interesting, yeah, so that'sthe short version.
Hopefully that.
Speaker 1 (17:10):
Okay, one of the things, one of the questions I
have, because it's very much apersonal interest and for NetEnt
, you know we're gettinginvolved more and more as
hardware is making its way.
As hardware is making its way,um, dedicated hardware is making
its way into the video networkand so I'm speaking of, you know
(17:31):
, gpu or vpu to do real-timemuch higher capacity video
encoding, transcoding, that's,um, you know, not bound by a CPU
so much, much more efficient.
Does the open caching standardenable, for example, now,
(17:53):
encoders to be at the edge ofthe network, so that you're
actually even doing the encodingin real time at the edge of the
network?
Can you comment on that?
Is that built into the standard?
Is anybody deploying that?
Are they thinking aboutdeploying it or is it not you?
Speaker 2 (18:10):
know possible.
Well, it's certainly possible.
Let me approach it from acouple of perspectives.
So we don't have anyone who'sdoing that today.
We've had that conversationbecause it's of interest to some
.
For the reasons you're saying,it could be a more efficient way
to do this.
Um, open caching doesn't?
(18:31):
The specifications at themoment don't explicitly address
such a use case, but there'sthere's no reason why it
couldn't be.
Speaker 1 (18:38):
You know, part of the suite of use cases that are in
other words, rather than, ratherthan pulling a stored uh,
either a stored asset or a astream being the input, it could
, literally the input could bean encoder.
Right, it's just sitting thereand it's producing.
You know it's outputting theappropriate files in real time,
(19:00):
you know the appropriate encodesfor whatever device.
Speaker 2 (19:03):
You know it's is uh being supported, I guess, or
devices uh, yeah, yeah, exactlyright, and, as I said that there
have been a number of peoplewho've talked about it with
interest, um, yeah but the otherone just throwing you know new
ideas.
Uh uh, gaming companies havetalked about doing rendering at
the edge.
I think it's kind of a similarnotion which is, yeah, you can
(19:25):
do it there more efficiently andfaster, so these things are
coming, I think and there's noreason why open cache and can't
support you, you, you, no doubt.
Speaker 1 (19:37):
uh, netflix knows who you are, but you know you
should be talking to netflixbecause obviously it's very
public.
They are building a cloudgaming network and it would just
seem to me with their OpenConnect platform, which of
course they write a lot about.
So this is all public is kindof the largest edge distributed
(20:00):
network, probably in thebusiness, and maybe not, but at
least with my limitedunderstanding.
So, yeah, very interesting welllet me comment on that just
generally.
Speaker 2 (20:13):
This is not about netflix specifically, but just
generally.
One of the things that we'veheard from service providers
telcos, broadband providers,cable providers, mobile
providers is this idea that, um,a lot of people, a lot of and
you can imagine, I'm sure youknow them all but a lot of
companies some hyperscalers,others with streaming services
(20:36):
want to get into those networksand want to be deployed deeply
with their, I'll say devices,their edge delivery devices.
But it seems to be a consensusamong service providers that
they're saying, yeah, I can't dothat for 10, 20, 30 people.
It would be a nightmare tomanage.
That's a lot of hardware.
(20:56):
That's exactly right, and so atleast several that we've spoken
to, who are major serviceproviders, are saying we'd like
to have this capability on aunified platform from one
partner that we can then offerto all, and and that's at least
(21:18):
um a sense that we're gettingabout open caching, because it,
as we've already said, it's it'sa industry specification, yeah,
adopted as a standard by theitf, uh, and it's open, and so
so this is something that wethink works in our favor so far.
Speaker 1 (21:37):
Yeah, it's great.
Well, this is the perfectlead-in, because you referenced
the Comcast announcement thatcame out.
That was Tuesday, I believe, sowe're recording this on
Thursday, so a couple days agoago now.
Congratulations, uh, uh.
We even have some listeners whohave, uh, who have posed some
questions.
So, um, I I want to talk aboutthis, and one of the questions
(22:02):
that I have and and I saw in thechat, is you know, maybe, maybe
you can explain in layman'sterms, you know, how, how this
is going to improve thestreaming experience for Comcast
customers?
And, like, my question is sortof a follow-up is is this a
Comcast customer of the Cape?
(22:22):
Because you know one of thethings we always, when we say
Comcast, are we talking thecable network, are we talking
the you know the IP over QAM, orare we talking the open
Internet?
So maybe you can even clarifythat.
But just in layman's terms, howis this going to improve the
experience?
Comcast didn't do this justbecause they needed a press
(22:43):
release, so there's a reasonthat they're adopting.
Speaker 2 (22:50):
Yeah, and that's exactly right.
Um, and you can imagine now,they were one of the founding uh
members of the streaming videotech.
Speaker 1 (22:58):
I remember that yeah, so we we've got some.
You've had a long deep.
We spent time with them, youknow but very specifically.
Speaker 2 (23:06):
So this is comcast, the you know cable provider okay
, okay.
Interesting.
It's the network, yes, and sothey're a national ISP three
plus million broadband users.
They're very, very big in theUS, one of the biggest cable
providers in the world, and, asI mentioned earlier that we will
(23:31):
have compute hardware placeddeep in their network in a
highly distributed architecture.
It will be the most distributedCDN if you will ever.
And on that hardware that theyprovide and they run and own and
(23:52):
operate, we will put oursoftware, we will cloud connect
it and then we will provide thecontrol plane that unites that
into the rest of this federatednetwork I've been talking about.
So why did they do this?
I think and there's a sorry forthe just like straight up
marketing promotion here, butthere's a great video on our
site in which our CEO, alainMayor, interviews network
(24:15):
officer Elad and he talks aboutyou know the dynamics, the
forces that are driving them tothis kind of outcome, a
partnership with us, and theyhave other announcements about
virtualization.
But specifically, they'retrying to activate their edge
(24:35):
because they know it's valuableand it's of interest to many in
the industry, not just streamingproviders In such a way that
it's unified, that it can bemade available to anyone who
wants to make use of it, to makeuse of it, and it would be then
used for a whole range of usecases, not just streaming, but
potentially gaming.
(24:56):
Other consumer use cases.
There are enterprise use casesright that, for example,
networking, security,collaboration, iot, connected
cards.
There's a really broad range ofedge computing use cases that I
think.
Well, in their case, I think wesee with them coming and their
(25:21):
goal is to create the foundationon which all that stuff can be
built and run efficiently.
Speaker 1 (25:27):
And so this is hopefully that's helpful.
This is what's driving, I think, their effort here.
Can you quantify?
You said it's the largestdistributed CDN or network.
What does that mean?
Is that based on number ofmachines?
Is it number of locations?
Is it capacity?
What exactly does that mean?
The largest?
Speaker 2 (25:47):
Yeah thank you, it's good to be clear about this.
So, in our view, when we talkabout distributed, we're
referring to some people call iton net distribution.
We often refer to it as deeplyembedded caches of who has the
(26:09):
most deeply embedded caches.
I'm putting aside capacity forthe moment, because we have more
than enough capacity to deliveragainst the demand we and the
customer demand we have today,but it's the proximity and the
distribution that gives usadvantage, right?
So we're inside the network,we're downstream of peering
points, exchange points that canbe congested, but, more
(26:31):
importantly, we're deep in thenetwork, over on the access side
, past the core.
We're downstream of peeringpoints, exchange points that can
be congested, but, moreimportantly, we're deep in the
network.
You know, over on the accessside, past the core, past
aggregation, sitting on theaccess side.
So there's nothing closer,there's nothing with more
proximity than we have.
And if you were to aggregateall of these in this federated
CDN that we have, all of thesedeeply embedded caches, nobody
(26:54):
has.
This is thousands of cachesthat we have.
Yeah.
Speaker 1 (26:59):
Yeah.
Speaker 2 (27:01):
I'm trying to.
Speaker 1 (27:03):
Okay, yeah, yeah, definitely that's helpful.
You know, I recall that Comcastrecently launched an enhanced
4K service, and maybe that wasaround the Olympics, I'm not
remembering now exactly, but I'mjust curious is there any
(27:24):
relation there or can you tipyour hat towards?
Is this now going to enablesome, shall we say, say, higher
quality experiences?
Because, let's face it, whenyou talk about, you know whether
it's quote enhanced 4k or, youknow, xr, vr, even gaming for
that matter the big challenge,the big hurdle, is really
(27:48):
network capacity.
At the end of the day, um, soyou know, people say, you know
how come.
You know service x, service y,service z, you know they don't
have more 4k or they don't.
You know, you know they're notstreaming with better quality
and it's like, well it?
It's pretty simple it comesdown to capacity and cost you
know.
(28:09):
So yeah, I curious.
Can you comment on?
Is there any you know?
Were you an enabler to that oryou know that announcement?
Speaker 2 (28:18):
Well, 4k is one of the use cases that's enabled, if
you will, by what we're doing.
Absolutely, you know.
I don't think it was the driver, for the reasons I've already
said.
They have a vision for edgecomputing that transcends 4K or
any one thing.
Speaker 1 (28:37):
But yeah, 4k would be an example.
Speaker 2 (28:39):
But back to your earlier point, and I completely
agree capacity is the scarceresource here.
As we look at what's happeningwith the use of internet and
streaming, just look narrowly atlive streaming.
Uh, you know, we.
We say 16 million live streams.
That's a us record.
You know the the super bowl isa hundred million users.
(29:02):
Do we want to somedayexclusively stream that event?
Maybe, and so it's.
You're exactly right, the, thecapacity to do it.
It becomes the constraint and,as I said, that's what we solve
for when we work with serviceproviders, like we do, because
we're offloading, you know, 99%of the core traffic associated
(29:23):
with a big event.
That's the way you get to 100million.
I mean big events, right, thecricket finals, that's a billion
viewers, and if they all wantto stream it, that's a problem.
Speaker 1 (29:39):
Yeah, definitely.
Well, I am looking here in thechat and there are a number of
questions that came in and thenare coming in, so why don't we
switch here?
It continues.
Many of them are in thedirection we're already talking,
so yeah, so okay, what are themain gaps?
(30:02):
To bridge the gap of the bigmigration of broadcast to IP?
Ip Boy, this is an interestingquestion, but you're working
primarily with service providers, so I think you're in a good
position.
Just comment generally.
Speaker 2 (30:20):
Yeah, generally and there's several parts to this at
least that we see so many ofthese big service providers,
cable companies, for example,they're already going to all IP
services, right.
So, for example, I mean I'm aComcast customer at home.
When I watch Xfinity, I'mstreaming it, right, it's not
(30:42):
the broadcast signal that I'mgetting, I'm streaming from
their platform.
Others are doing the same thing, and when we're deployed in one
of these networks, and whenwe're when we're deployed in one
of these networks, we cansupport that streaming of their
managed content.
We do that today with Verizon,fios TV, for example.
So we help enable that.
I'm kind of doing thisresponding in part to your, to
(31:04):
your question.
The larger question of themassive transformation from
broadcast to IP is certainlyunderway.
We see those trends, but itwill take time and, as I said,
(31:26):
capacity is one of theconstraints that needs to be
solved for to get us to theother side.
You know it'll take years, forsure, right, for everyone to
stop using broadcast televisionand start streaming, um, but but
it's, it's an unstoppable forcein the industry, it's
inevitable, and so I think it'sjust.
You know, so I think I, I Iwould.
Speaker 1 (31:46):
I would like you to comment, though, because I think
the question, uh, is okay.
All all of what you said istrue and you know the trends and
, like you say a lot really, allof these networks, they're
already IP networks, so you knowone of my.
So my initial response to aquestion like this is like well,
what gap are you talking about?
(32:08):
Because, fundamentally, otherthan I don't know, you know some
really remote regions of theworld, or you know where there
still are significant legacybuild-outs, like I don't know in
the Americas who isn't runningan IP network, you know.
So now that's my response.
I'm sure others, who are muchcloser to the details, could say
(32:30):
, oh, but Mark, you know.
I'm sure others who are muchcloser to the details could say,
(32:59):
oh, but Mark, controllednetwork, my closed network, and
then I have my inner theinternet right, because that is
still true and even insideComcast, right, you know there,
there there's sort of twonetworks.
I mean not sort of.
There are two networks you knowthat you interact with.
So I'm wondering do you have anyperspective of you know like at
what point, or will they nevermerge for whatever technical or
(33:23):
political reasons, or you knowwhatever.
Speaker 2 (33:25):
Well, again, I think it's going to be a long, long
time before somebody shuts offthe broadcast network at one of
these big companies.
Speaker 1 (33:35):
For example.
Speaker 2 (33:35):
The other gap, though just to continue on the list
would be solving for ultra-lowlatency delivery.
So we want the signal, that'sright.
If someone's streaming toarrive at the same time the
broadcast signal does, so thatyou know, the classic problem of
social media spoiler doesn'ttake place.
I think we're getting there andwe're working with one of the
(33:59):
largest streaming platforms inthe world on a ultra low latency
protocol, awesome UDP basedthat is delivering the signal
within you know, a few seconds,five seconds.
So we're starting to see thatpart of it being, because it
used to be right, maybe 60seconds, which was an
(34:21):
unacceptable delay, and peoplewouldn't consider that if they
had you know, money riding onthe game, but we're getting
there on that front.
But yeah hopefully the problem,you know, the transformation,
the migration.
It'll take time.
Speaker 1 (34:37):
Yeah, yeah, for sure, okay.
Well, a question that came inwas about WebRTC and
specifically the person isasking you know, well, I'll just
read the question, because it'spretty well formed what do you
think about the future ofstreaming via WebRTC and then
its impact on CDNinfrastructures?
(34:59):
And they go on to say andtranscoding.
So yeah, webrtc, is it playinga factor?
Will it never play a factor?
Is everything going to go toWebRTC?
What's your view there?
Speaker 2 (35:17):
Yeah, I mean, I think this is promising.
I'm afraid I'm going to answerit in the more general case,
because I've used the example ofone of our customers who's a
global streaming platform andthey're using a UDP-based stack
for live streaming.
That gets us to a few secondsof delivery, that is, latency
(35:38):
between the broadcast and thestream.
If it's going to be WebRTC,great, we can support all of
this.
So we're looking for you knowwhat our customers want, as
opposed to for us, you know,promoting a particular path or a
very specific stack.
(35:59):
Beyond that, I, you know, Iwill see, we'll see where this,
where this goes.
Speaker 1 (36:05):
Yeah, exactly, exactly it's.
It's a really exciting space.
It's a really exciting spaceand I think you know.
Just my personal commentary isI think one of the challenges
excuse me that we have in theindustry and when we're having
these sorts of conversations isthat there is no sort of hand
(36:25):
wave.
You know and I'm picking onWebRTC here but there almost
never is a scenario where it'slike, well, webrtc is going to
solve that, or XYZ is going tosolve that, or oh, this is going
to you know, sort of that.
And oftentimes I find I'm inconversations where questions
are framed or where it's kind oflike.
(36:47):
It's like, oh, but isn't youknow, fill in the blank,
technology going to solve that.
And it's like, well, let's zoomout.
And first of all, you have tounderstand the incredibly
complex architectures of howcontent is delivered from you
(37:09):
know, from origin, from source,from you know, whatever the
topology, whatever the topologyis, to the viewer.
In one scenario, it actuallymight be true that WebRTC
largely is going to be the waythe content is delivered, but
for that very same stream, therecould be another scenario based
(37:30):
on a different geographicproximity and the networks that
are involved and the devicesthat are involved, that WebRTC
will never work, you know, andthere can be a third, and
there's a fourth and a fifth anda sixth.
So I make this observation morefor listeners who are listening
, you know, to conversationslike this or just having other
(37:50):
conversations throughout theindustry, and it's very natural
and it's perfectly reasonable tobe really leaning in like, hey,
what technology should we belooking at, building on, maybe
even doubling down on?
But regretfully, for all theyears that I've worked in the
space, I have never seen, and Idon't see in the future, a
(38:14):
scenario where a technology iswant to debate.
You know, feel free to commenton LinkedIn, you know, let's
(38:38):
talk about it.
So, all right.
Speaker 2 (38:41):
Let's move on, Completely agree.
Speaker 1 (38:42):
Mark, yeah, yeah, yeah, Okay, let's move on.
So all right.
So here's an interestingquestion.
All right, so here's aninteresting question how can
hundreds of gigabytes of data beused for real-time analytics?
What can the best turnaroundexpect?
(39:02):
How can hundreds of data beused for real-time analytics?
Speaker 2 (39:24):
Do you know what this question means use edge
computing as a resource tohandle processing of data in
proximity to the device that'sproducing it, because that's
more efficient and moreresponsive than centralized
cloud computing?
Yeah, the answer is verypossibly yes, and at least there
(39:48):
are definitely use cases aroundedge computing where the
assumption is we want to processanalytic data from a device as
close to that device in thecloud edge cloud as close to the
device as possible.
Uh, so so?
we're not doing that todaythat's not a current um use case
we're supporting, but I thinkit's okay there, in theory it's
it's one that's possible yeah Idon't know if that's the
(40:10):
question okay, all right, well,um, well, feel free.
Speaker 1 (40:13):
I know Mark is very accessible on LinkedIn.
I'm accessible on LinkedIn, soyou know.
If we don't hit your questionor if you have a question that
you weren't able to raise, yeah,hit us up.
Ok, are you developing?
This person wants to know ifyou're developing the Gulf
Cooperation Council countries,so I think it's like Saudi
(40:35):
Arabia, bahrain.
So yeah, are you you know?
Do you have any deploymentsthere or anything coming up, or
can you hint something coming up?
Speaker 2 (40:47):
Yeah, I can say what's public, which is we have.
We announced some time ago apartnership with STC, saudi
Telecom.
Speaker 1 (40:53):
Okay.
Speaker 2 (40:54):
And we are deployed there and they have announced
again publicly that they aredelivering content throughout
the region.
I can't name exactly thecountries within that region,
but they're referring broadly tothe Middle East.
So yes, we're there with thatpartnership and it're referring
broadly to the Middle East.
So so, yes, we're there withthat partnership and it's active
(41:15):
.
Speaker 1 (41:15):
Awesome, yeah, yeah, that's great, that's great.
Okay, um, well, let's, uh,let's, you know, let's wrap up,
um, where I have a two-partquestion for you and I won't do
the usual, you know, crystalball.
Where are we going to be infive years?
(41:36):
I, I, personally, I hate thosequestions, yeah, but but I do
have a two-part question andit's and, and I'll start I'll.
I'll start with the first partand that is where.
How are networks?
You know, I referenced it whenyou started the company, when
(41:57):
Quilt was founded in 2010.
It was.
You know, network architectureslook different.
You know, largely.
You know 95% of, or what youknow.
I'm just guessing.
I don't know the real number,but it's probably in that range
of content being streamed.
It was file based, it was bod.
You know, there was actually,um, relatively speaking, very
(42:21):
little live.
Uh, I mean, there was livevideo.
I'm not saying there wasn't,but you know very little.
Um, certainly, that hasn't beenflipped.
I mean, still, the majority ofcontent is still file-based,
it's still VOD.
There's an asset that's sittingon an origin server somewhere
(42:42):
and being distributed, but let'sface it, you see, netflix who,
just a couple of years ago, Ican remember famously, this was
more than a couple of years ago.
You know I can rememberfamously this was more than a
couple of years ago.
But asking somebody whose nameI won't reveal, but somebody we
all know very well insideNetflix, at that time, you know,
(43:02):
why doesn't Netflix do live?
And the response that you knowthat I got back was I mean, it
was as firm as anything Netflixwill never do live.
All about the VOD, you know,sort of like ad.
You know Netflix will never doadvertising.
And then, next thing, you knowNetflix is doing advertising
very much and you know so, ontransforming, you know um, over
(43:34):
the next, you know, say, threeto five years, so kind of the
medium term horizon, uh, withthis really decided move to like
live live streaming.
You know, and and it's notgoing anywhere.
It's not like this is a fad.
Speaker 2 (43:45):
This isn't 3d video, remember that yeah, by the way,
the first part of my question isgot.
Speaker 1 (43:53):
I got it, where are we?
Speaker 2 (43:54):
going the.
The netflix comment lives nowin the archive with apple's
comment they would never sell aphone, right?
Speaker 1 (44:01):
yeah, exactly, yeah yeah.
Speaker 2 (44:03):
So you know I I think , and just looking at at
streaming and video, um, youknow things are happening there.
Live is is more and more a partof the fabric.
Exclusively live streamedevents are that way.
I mean you just look at this ispublic around Amazon.
But look at what they're doingwith Premier League.
(44:24):
I mean I don't think there's asport at the moment that they
aren't live streaming.
They're securing deals toexclusively stream events across
the spectrum of sports.
There is live linear right, sothere are fast services that are
increasing as well.
We see that taking off.
(44:44):
I think, just sort of generallyspeaking, we will see this
continued migration andpreference from consumers.
I think that's the mostimportant thing, never mind what
platforms are promoting what doconsumers like.
There's a whole host of reasons, I think, for consumers to
prefer to stream an event eithervideo on demand, but certainly
(45:10):
a live event rather thanwatching a broadcast television.
You know mobility, flexibility,it's extraordinary and the
selection of content is justkind of endless.
So I think we'll see that justcontinue to take off.
There are things like 4kstreaming, as you mentioned.
There's immersive experiences,you know AR and VR, those kinds
(45:31):
of things, maybe some of thembeing promoted by Meta and
others.
That, I think will you knowwill potentially see get more
traction.
I'll just say too, and this ismaybe the more boring part, we
see enterprise use cases as alsopicking up here where, as I
(45:53):
mentioned before, collaboration,iot, connected car security,
networking those will consumeresources at the edge, and
that's the kind of thing weexpect.
Speaker 1 (46:05):
So what does this mean, though, for how networks
are built?
Um, because that that's reallya heart of my, of my question,
and I'll and and I'll, and I'lladd another layer of context.
Um, I, I guess, a couple yearsago now, akamai bought linode.
Right now, uh, and they're, andthey've been very active out
(46:26):
promoting, uh, you know the theedge, edge computing.
I mean linode is their.
You know cloud edge, edgecomputing, I mean Linode is
their.
You know cloud edge computingplatform, and I mean we both
just came back from IBC.
You know a lot of.
You know Akamai is out thereand, by the way, other networks
too.
It's one of the trends that I,dan Rayburn and I talked about
(46:48):
in a couple of days ago in hislatest podcast episode.
Was this trend?
So then, my question is okay,if Akamai is building compute
into the edge and you knowthey're positioning it a little
bit different, I mean, you knoweverybody has their angle right,
but at the end of the day, thisis edge computing, you know at
(47:09):
the end of the day, you knowthat's what they're doing.
They're making hardwareavailable in a distributive
fashion that an Akamai customerand one can assume Tata and kind
of go down the list of all theobvious that they're probably
either doing it, are doing it,planning to do it, and just
haven't announced it this iscoming.
(47:30):
Are doing it, planning to do itand just haven't announced it.
You know like this is coming.
So what does this mean for theway that CDNs and networks are
built and the architectures?
You know?
What are your thoughts?
Speaker 2 (47:43):
on that.
Well, yeah, I mean, and Linodeis a good example to bring up,
at least from public statements.
They were, linode was maybe20-something sites.
At the time of the acquisition,akamai paid, I think, $900
million for the company.
They've announced plans sincethen to roll out.
(48:03):
Akamai has announced plans toroll out to 100-plus locations
with that technology, with thatedge computing stack from Linode
, with that technology, withthat edge computing stack from
Linode.
Back to the point about networks.
It seems to me that if you're,you know, really focused in on
(48:25):
edge computing, there's adefinition there that plays
right.
So it could be a premise edge,the device where some of those
edge computing use cases arerunning, because that's the most
efficient place to do it.
But it's not no longer thecentralized cloud.
We know that.
That's understood.
What the hyperscalers do webelieve that back to your point
about networks that it's theservice provider who has
(48:49):
unparalleled proximity scalereach service provider who has
unparalleled proximity scalereach that has to activate in
their network an edge computingcapability and it should be as
deeply embedded, as close to theusers as possible.
That's where we come in and Ithink others, if you look at the
hyperscalers Azure, aws, I meanthey do have a kind of edge
(49:12):
ambition and in many cases theybelieve that should be sitting
inside a service providernetwork.
Speaker 1 (49:20):
We agree with that.
Speaker 2 (49:21):
But you know, the service provider has to be a
collaborative partner if that'sgoing to happen.
And that's at the heart of whatwe do really well.
You know, you mentioned Alkama.
It's not clear to me whetherthey can really activate that in
the same way.
I think that's the, you know,the mystery that has to be
solved.
Yeah, it's fascinating.
Speaker 1 (49:44):
It's fascinating.
I think we can end on this note.
I won't really pose it as aquestion, but I'll give you a
chance to you know, either sayyes, you know, or comment
further.
But it seems like where Quiltis positioned really well is
that you know you're able tobring together the constituents
(50:07):
who you know.
There's obviously technologycomponents.
There are technology componentsthat you bring that are needed,
right, that a service providerneeds to be able to deliver.
You know, throughout this, youknow, as I used the word earlier
morass of networks andarchitectures, right,
(50:51):
no-transcript, it's public andwell-known.
You know Netflix has done anamazing job of building out
OpenConnect.
I mean, I think the last publicnumber, like whatever 80,000,
you know machines or pops or Idon't even know how it's defined
, but it's a huge number thatthey have deployed.
They've done an amazing job.
But that was hard and there'sstill networks who are like yep,
(51:15):
not in our network, you know,and so that's a challenge.
You reference Akamai.
You know like, like ah, youknow there's some where it's
kind of like oh, we're not tooexcited about maybe getting, you
know, having having acommercial CDN inside our
network.
So is this is mycharacterization correct where
(51:36):
you know you're bringingtechnology, but you're also able
to bridge some of these gaps.
Is that, you know, either toosimplistic a view or is that an
incorrect view?
Speaker 2 (51:47):
No, Mark, you're spot on.
Yeah, you're spot on and I'mglad you raised this point A
good way to maybe close up.
But yeah, the tech we bring ispowerful and open caching as an
open specification is powerfulfor the reasons we've done
really well.
(52:08):
That sets us apart is recognizeearly on the need to
collaborate deeply with,cooperatively with, service
providers and bring that, bringthe ecosystem, the value chain
together in that way.
I should say, quilt of theinspiration for the name was, to
this very point, the idea ofstitching together a fabric so
(52:30):
that it comes becomes a whole, acomplete, and I really do think
we've done that better thananyone else.
You know can witness theComcast announcement this week.
I think, that speaks to ourability here.
Speaker 1 (52:42):
Yeah, that's great.
I think it's a great way to endQuilt love the name.
You know you're stitchingtogether all of these, um, these
hard things and ultimatelyenabling the ecosystem to
deliver content, you know, moreefficiently, and you know that
has a positive impact on cost,positive impact on user
(53:04):
experience.
Uh, you know, just has apositive impact.
So, hey, mark, um, thank youfor joining me.
Uh, we went a bit long, but youknow, as I told you before, we
pushed record or pushed go live.
You know the way we run this isif the conversation's rolling,
we just keep rolling and at thepoint that you know it kind of
(53:27):
naturally winds down, then wewrap it up.
So really appreciate your timeand thank you for coming on.
Thanks.
Speaker 2 (53:34):
Mark, thanks for having me Take care.
This episode of Voices of Videois brought to you by NetInt
Technologies If you are lookingfor cutting-edge video encoding
solutions check out NetInt'sproducts at netintcom.
Popular Podcasts
True Crime Tonight
If you eat, sleep, and breathe true crime, TRUE CRIME TONIGHT is serving up your nightly fix. Five nights a week, KT STUDIOS & iHEART RADIO invite listeners to pull up a seat for an unfiltered look at the biggest cases making headlines, celebrity scandals, and the trials everyone is watching. With a mix of expert analysis, hot takes, and listener call-ins, TRUE CRIME TONIGHT goes beyond the headlines to uncover the twists, turns, and unanswered questions that keep us all obsessed—because, at TRUE CRIME TONIGHT, there’s a seat for everyone. Whether breaking down crime scene forensics, scrutinizing serial killers, or debating the most binge-worthy true crime docs, True Crime Tonight is the fresh, fast-paced, and slightly addictive home for true crime lovers.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.