March 1, 2023 • 36 mins
In this episode, Eveline Oehrlich is joined by Haseeb Budhani to discuss all things that are essential in regards to Kubernetes.
Haseeb is a Co-founder and CEO at Rafay Systems. The Rafay Kubernetes Operations Platform is a turnkey offering that automates Kubernetes cluster management, modern application deployment and operations at scale.
Enjoy the Humans of DevOps Podcast? We’re incredibly grateful to be voted one of the Best 25 DevOps Podcasts by Feedspot.
Want access to more DevOps-focused content and learning? When you join SKILup IT Learning you gain the tools, resources and knowledge to help your organization adapt and respond to the challenges of today.
Have questions, feedback or just want to chat about the podcast? Send us an email at podcast@devopsinstitute.com
Haseeb is a Co-founder and CEO at Rafay Systems. The Rafay Kubernetes Operations Platform is a turnkey offering that automates Kubernetes cluster management, modern application deployment and operations at scale.
Enjoy the Humans of DevOps Podcast? We’re incredibly grateful to be voted one of the Best 25 DevOps Podcasts by Feedspot.
Want access to more DevOps-focused content and learning? When you join SKILup IT Learning you gain the tools, resources and knowledge to help your organization adapt and respond to the challenges of today.
Have questions, feedback or just want to chat about the podcast? Send us an email at podcast@devopsinstitute.com
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Narrator (00:02):
You're listening to the Humans of DevOps Podcast, a
podcast focused on advancing thehumans of DevOps through skills,
knowledge, ideas and learning,or the SKIL framework.
Haseeb Budhani (00:16):
VR, even in this economy at a point where there's
not enough talent available inthis space, who understands and
can operate Kubernetes not justbring up the cluster right?
Again, these are simple things,like truly running this at an
enterprise scale is a very hardscale.
Eveline Oehrlich (00:33):
Welcome to the Humans of DevOps Podcast. I'm
Evelen Oehrlich, Chief ResearchOfficer at DevOps Institute. The
name Kubernetes originates fromGreek meaning Helmsman, or
pilot. I've also heardKubernetes, often described as
the Linux of the cloud. Andtoday is the most popular
(00:54):
container orchestration platformfor multiple reasons. Kubernetes
burst onto the it developerscene in 2014, when Google
released it as an open sourceversion of its Borg technology,
which which they developed as away to run 1000s of jobs and
applications across multipleclusters and machines. Since
(01:15):
then, the technology actuallyhas spread far and wide. It is a
key part of managingapplications in data via
containers, and Gartner projectsto reach this technology to
reach 944 million by 2024.Today, I'm excited to have with
(01:35):
us Haseeb Budhani, who is cofounder and CEO at Rafay
Systems. Hello, welcome to thepodcast. Haseeb.
Haseeb Budhani (01:44):
Hi, everyone. Nice to talk to you. And thank
you for having me.
Eveline Oehrlich (01:47):
Yes, I did some googling you and looked at
your of course, LinkedInprofile, you have done a lot of
things I had to scroll andscroll and scroll, all these
different types of things youwill have been doing maybe give
us a cliff notes of what haveyou been up to before you were
(02:09):
advice systems, the CEO andfounder?
Haseeb Budhani (02:14):
Sure, since undergrad, I have been fortunate
enough. In fact, while I was Iwas a senior in college, I did a
number of internships in my inmy time at USC. And the very
last internship I did was with acompany that was writing
(02:36):
shopping cart software. This isback this fall of 99. Not going
to exactly how old I am. Thatwas my senior in college. And
yeah, these guys were doing somecrazy things, writing this thing
called a shopping cart. And theywere desperately hiring
(02:57):
engineers. And they give meYeah. And it was a lot of fun.
It was it was crazy. So beforethen I'd worked at Cisco and
done a set internship at Ciscoand an internship at Ericsson.
And I had a sense for how thosecompanies weren't, these are big
companies. But this othercompany, it was just crazy.
These people didn't sleep. Theyjust kept burning code. And it
(03:17):
was awesome. So having once Ifinished that internship, then I
started applying for four fulltime jobs. I just wanted to work
at startups. I just wanted tofind companies and I didn't have
the concept or appreciation as a21 year or for for stock or
anything had no idea I caredabout was about these these
companies. Well, you know, theypay well, they don't pay as well
(03:39):
as a Cisco perhaps. But wow, youget to work on some amazing
things. And that's what I did.My first job out of college was
at a start as a startup, whichwas called Publix. And that
solve for single sign on. It'sone of the first companies that
were doing single sign on in ourindustry. So it was not a thing
at the time, but they were doingit under another startup. And in
(03:59):
the process. I think I just youknow, good or bad, I learned a
lot of a lot of differentthings. And that's been you
know, it was a lot of fun. Andtruly, I believe that had I not
done those things, I would nothave had the opportunity to work
on startups that last 1012 yearsthat I've been, you know, in
some fashion of founder orwhatever. Yeah, just small
(04:24):
decisions, or very smalldecisions can put you on these
paths. And I think back a lot tothat one specific afternoon when
we had this career day. And Idecided to talk to these crazy
people back at USC and thatbecame you know, my pad had a
mountain that hadn't taken a jobat Cisco because I had an offer
from Cisco when you when yougraduate come work here and I
didn't take it. Yeah, I alwaysworked at startups. Only work at
(04:48):
large companies through somepathway positions that have not
worked at I've never it's been along time since I've actually
applied for a job in a largecompany. That's become my my
personality but the The mostimportant thing I do here, or I
require companies is, as a CEO,my primary job is sales. Like my
job is to give our customers theclarity as to why the company
(05:15):
exists, why raffia exists? Well,like any other company, when
I've worked at why does thiscompany need to exist, rather,
and how this can make their lifebetter? By getting people to
their point of clarity so thatthey can go, I see how I could
use it. There's nothing like it.I mean, that's, that's what
drives me. Right? So I enjoythat more than anything else.
And particularly if you havethe, you know, somewhat of a
(05:36):
technical background, I happento have one. It's, it's great,
right? So anybody working on newtechnologies, we all have to
understand it. I mean, we cancreate technologies, we can sell
it, it doesn't really matter.It's like a tree that fell in
the forest, or it doesn'tmatter. But being able to build
something, and then being ableto articulate why. And why
(05:56):
should you care? Right, that is,arguably the more important
thing. And if you can figurethat out, you know, you're gonna
have a lot of fun in thisindustry.
Eveline Oehrlich (06:05):
That leads me to one of the factoids, I found
rough AI systems was actuallynamed as one of the hardest one
of the 10 Hottest humanitystartups of 2022. By CNR. It's a
channel magazine, but that'sgreat. So two questions for you.
(06:25):
What does Rafay Systems do?First, and what has put you on
the list of the top 10 startupsand like you said, of course,
your startup, your passion forstartups, probably has something
to do with it. But help help ourlisteners, as those are many of
them are developers, I owe, youknow, geeks and good geeks. Help
(06:48):
them understand what does RafaySystems do? And how did you make
it to this top 10 startups.
Haseeb Budhani (06:56):
So Rafay is in the Kubernetes management space.
Kubernetes is, oh my god,there's like hundreds of
companies, something inKubernetes. It's a very noisy,
very busy space. But when Ithink about what we do, like,
fundamentally, we solve a peopleproblem in this company. Every
(07:17):
enterprise, we telco, anycompany that has software that
they're deploying, goingforward, it will be
containerized, or some function,because they want move fast as a
company. And in many cases,majority of cases for companies
will pick Kubernetes, as ourorchestration engine, you
(07:38):
describe Kubernetes has been thefastest growing container
orchestration platform outthere. It's it's the de facto
standard, if not the standardpresently. But it brings with it
a number of complications. Justbecause you have an
orchestration engine doesn'tmean you as an enterprise can
consume it easily. You have anumber of constituencies inside
your company, there's security,there's operations as
(07:59):
developers, there's all theseother different people and
multiple developmentorganizations inside every
enterprise. They all have towork together somehow. It's a
people problem, everything is apeople problem. And to that end,
what you've what you'veattempted to solve for, and it
seems like you've done a reallygood job, and you have a really
nice roster of customers who usethe product daily, is that
(08:21):
they've thought through what ittakes for an enterprise to
really build shared servicesplatform for Kubernetes. How do
you deliver Kubernetes as aservice inside your enterprise?
And what that means is not justthe material cluster, that's the
easy part, building a cluster orcreating a Kubernetes cluster.
That's not the issue. That's asolved problem. In my mind, the
(08:42):
real problem is, who has accessto what do we know what they're
doing? Can we do thisconsistently? Can we can we
update patches consistentlyacross our fleet of clusters?
can I provide different levelsof access to different teams? If
certain teams have networkingrequirements that are different
from other network teams? Howabout how do I make that happen?
(09:03):
And how do I do all of thiscentrally with the right level
of governance? This is howenterprises pay. If you solve
that problem in the context ofKubernetes, of course, it's
solving a technology problem.But really, you're solving a
people problem. Because theskill set is it takes time to
build a skill set. We are evenin this economy at a point where
(09:24):
there's not enough talentavailable in this space, who
understands and can operateKubernetes not just bring up a
cluster right? Again, these aresimple things like truly running
this at an enterprise scale is avery hard skill. Every
enterprise is looking for thosepeople and they're not able to
find them. And what we aretelling them is we will help you
(09:44):
augment your teams with softwareautomation. That's what we sell.
Fundamentally what we're sellingis automation that augments an
existing team. And the beauty ofthe right automation is that a
the enterprise they get up andrunning now. They don't need to
wait a while to hire Are peopleand then build a platform
because we sell them a platform.But to me the more important
thing, and this, I think, longterm is the right way to think
(10:06):
about any technology, the peoplein the organization that are in
our customer, who are notexperts at Kubernetes, by
working with our product, and byworking with our support
organization, they actuallyended up becoming experts. If
you can, in sort of, indirectlyor perhaps as a byproduct, get
(10:27):
people, you know, sort of, youknow, adept at Kubernetes, and
all the things that happenedaround it, right. That is, that
is generally good for our forour community, all right,
industry, right. So we sell agreat product, the enterprise is
happy, because they can now movemuch faster, right? They're TCO
is lower because they can dothis today, etc. They don't need
(10:48):
to wait to hire another 510 15people. But the existing IT
engineers and I'm using the wordit very loosely, we call them
DevOps, right depends on the onthe function they have. And then
broadly speaking it, they allget to learn this new
technology, which is going to bewith us for at least 10 years,
if not going to be all over timeafter learn this. And yeah, we
(11:10):
take pride in saying that ourcustomers, their engineers, you
know, months into ourengagement, Rafi, yes, the
enterprise is better off, butthe engineers are better off to
Eveline Oehrlich (11:20):
Yeah, we'll get to the future, hold that
thought on where this is going.Because that I want to dive into
that a little bit deeper. Butback to what you were just
saying. So we do now 22 was agreat year for Kubernetes. And
we've had in our organization,lots of questions for upskilling
in this topic. While it wasinitially viewed as something
(11:42):
only really large enterprisescould benefit from, we know that
it has improved in usability,most likely because as you said,
skills have gone up. But thereare still some challenges. Now,
there are some technicalchallenges with it. What would
you say if you think about yourclients, those you speak to
every day? What are some of thebiggest challenges leveraging
(12:05):
Kubernetes these days, and justfor right now just focus on the
technology in itself? Because Iknow there's a few. And then
we'll move further on once we'redone with that towards the you
already said that, which ismusic to my ears, the skills and
the skill development and thereduction of toil and all of
that? Well, we'll get to thatright now. Let's just focus and
hone in a little bit on thetechnical challenges you you
(12:28):
see.
Haseeb Budhani (12:29):
Yeah, absolutely. So many people ask,
you know, what does Rafay do,and I use the phrase Kubernetes
management, I didn't use thephrase Kubernetes for the
following reason. In my mind,the the biggest player in the
Kubernetes space is AWS, theyprovide an engine or product
(12:53):
called COVID PKS elasticKubernetes. Service, which is,
as far as I know, the most usedKubernetes offering right now.
So Amazon is the Kubernetescompany. So then what is rapid?
So once a customer decides I'mgoing to use maybe Amazon's
(13:13):
Kubernetes, maybe somethingelse, Azure, then they start a
journey, where they have to nowfigure out, okay, how do I
automate the provisioning ofthese clusters? There's
automation for that, and thenpeople who know different types
of technologies, right,TerraForm, etc, then I need to
understand how to upgrade thesethings. Okay, there's automation
(13:35):
for that, that you have to kindof figure out, then I need to
understand what are thecomponents that need to run on
this Kubernetes cluster, so thatmy applications can consume it.
They're all sort of raw out ofthe gate. So then you have to
learn these things. ThenSecurity says, Well, you really
need to make sure the rightpeople have access to the right
thing. So we have to think aboutwhat based access control and
the right level of identity andthe right level of access, tie
(13:57):
this back to the enterpriseSingle Sign On system. We should
really audit everything. who'sdoing what? Okay. All right,
let's go figure that out. Whilewe're on the point of
Kubernetes, is deployapplications. So case, we should
have pipelines of some sort,connecting back to maybe GitHub
or GitLab. Right? So we got tofigure that out. Hey, we have
(14:18):
certificates that we're pushinginto these clusters, for, you
know, TLS termination, we reallyshould think about some sort of
secrets management. Okay, well,let's go figure that out. And
chargebacks, service mesh,network policies and Kubernetes
policies, which is differentfrom their policies, and then
(14:39):
really, developers need access.So we should really think about
it the right developerexperience. And so each of these
things, and there's, dependingon the situation, there could be
many other things to do. Andthis is a challenge right now in
this industry, right? Nobody'swritten a book. I'm sure there's
a Kubernetes for Dummies, I'msure there is. But but it
doesn't talk about what it takesto not just get the toy up and
(15:02):
running. But to truly consumethis as an enterprise. By that
framework, what I just describedas a is a function of mine just
experience working withcustomers. But you know, there's
no Bible here. Maybe you shouldwrite one, I don't know. But
this is the challenge, right?And we're and the worry I have
right now is that whenenterprises jumped into this,
they don't know all thesethings. Now, could they know,
(15:23):
they just started, right? Andinitially, everything seems
easy. What's the big deal? Ijust go to the console, I build
a cluster, boom, boom, boom,it's working. See, I can do
Kotlin. But then start the realchallenge. Right, and now it
takes a year, who knows how longit's going to take depends on
the size of the company. Andthat's not okay. Right? This is
this is a detriment, right? Thisis going to get in the way of
(15:45):
progress. Right. So obviously,when when you see gaps startups
come about, and Rafi solve thatproblem, our job is a, we're
going to make for example, Eks,or Azure Kubernetes, which is
AKs, we're gonna make it out ofthe box enterprise ready, and
we're gonna help you manage nnumber of views across clouds if
you want. So we solve thatspecific gap. But but this is a
(16:07):
really hot technology promise.So our customers are pretty
sophisticated engineers, andthey're all very smart people.
This is what I actually reallyenjoy about this, this specific
job. I'm not trying to, I don'tneed to convince anybody of the
problem, right? We have openconversations, and they get to
the point of clarity, I get whatthe gaps are. But but sometimes
(16:28):
they don't know what the gapsare, unless they experience
that. But experiential learningtakes time. Right? You're going
to take a year, 18 months tofigure out what you don't know,
or what is missing. But yourcompany just launched 18 months,
this is the technology issue,right now in our industry. But
we're a small company,relatively speaking. So it's not
like I can I can solve this foreverybody. But this is what I
(16:48):
think about all day long. Bythis. Sometimes we as an
industry have the you know, wedo something that is bad, which
is reserved, you know,trivialize the the, you know,
the complexity. Everything iseasy. Look how easy it is look,
one command, and it's all done.Now, let's just have real
conversations about yes, indeed,this is hard. That's okay.
(17:10):
Because it's hard doesn't meanit's bad. It just means I need
to step back and think about howam I going to solve for the
complexity? I think it's betterto have that open conversation
versus try to kind of, you know,hide this or not hide as hide is
a strong word, but just, youknow, trivializing this or, or
trying to minimalize This is notokay. Right. And I truly believe
(17:31):
that this is hurting a lot ofenterprises or companies in
general, who are starting tojoin it because they are not
aware of how complex thattechnology is. And if they were,
they would make betterdecisions. And maybe some of
them will say, this is not forme, that's okay. But still, we
have to give them that clarityupfront, it's complex, here are
ways to solve it. That will overkind of, you know, at the macro
(17:53):
level that will help thisindustry more than a specific
vendor who's trying to get to asale. Don't think like that, I
think at the at the macro level,give people that education and
that clarity, and I think that'sgood for all.
Narrator (18:09):
SKILup IT Learning is the perfect online destination
to learn about DevOps anddigital transformation anytime,
anywhere. A digital learningplatform provides you with all
the resources you need toupskill and learn about these
important topics, includingexpert led course videos, access
to certification, prep courses,and a community of supportive
(18:31):
peers. A subscription basedmodel makes it easy for you to
gain the skills needed forsuccess in the modern workplace.
Visit DevOpsInstitute.com nowand explore our different
learning plans, get 30% off,SKILup IT Learning Plus now,
claim your offer and startlearning today with code Feb. 30
off, that's F E B three zero, of f.
Eveline Oehrlich (18:56):
And I think as technologists, I've seen this
over and over in conversationswith CIOs and VPs where they
have been promised by theirteams that oh, yeah, this shiny
new object can help usimmediately to increase flow,
velocity, speed, quality,whatever. And those executives
(19:16):
promise that onto the partner inbusiness, and six months into
the journey or 12 months, maybelonger even then they are all
disappointed because it hasn'tachieved the outcome they have
been thinking about right. And Ithink you hit as we say in
Germany, the nail right on itshead where we sit everywhere.
(19:38):
Yeah, we you need to berealistic about setting
expectations and making surethat we have these
conversations. Sometimes they'rehard. You did mention a few
things earlier, around thechallenges relative to people
and organizations, right you andin your examples, I could hear
some of that because we've stillsilos now Add sometimes a bad
(20:00):
word because we have experts incertain areas, right? Even
though we have dev SEC ops, thesecurity team is not yet as
integrated as it should be. Andso my curiosity is around people
and organizational challengesrelative to Kubernetes. In
bringing in such technology,what do you see your clients do?
And what, what's what works?Because many of our listeners,
(20:24):
I'm sure say, Yeah, okay, we'vedone this, but it fell flat on
its head, and it hasn't failed.Because it's Cuban Eddie's which
didn't work. It failed forcultural reasons.
Haseeb Budhani (20:37):
Yeah, there's, there's some influence we have.
And there's, of course, youknow, as a vendor, some
influences we don't have when weget to see a lot of stuff. What
one investment we've made inthis company, which is perhaps
unorthodox is that we built aservice delivery team, in the
company. And the point of thisorganization is, when a customer
(20:59):
says, people do talk to them, wesort of, you know, agree that
they should try the product, andthen we try the product. And so
this is pretty awesome, I shouldbuy it. And then they buy it.
And then what, right, justbecause you did a POC doesn't
mean it's it's plugged into yourinto your company, right? That
that's a process, right? A lotof people have to be educated to
bring developers and say, Hey,here's a, here's a platform, and
(21:22):
you want to think about ABC. SoSo we saw that happen again, and
again, where the where thecustomers would basically go
through this internal process,right, where they try to build a
framework, and they try toinvite developers and try to
sort of, you know, convince thedevelopers internally, hey,
don't do this yourself anymore.It's waste of your time, here's
(21:43):
a better platform, and we'regonna give you all the
automations. So we decided, youknow, what, we're gonna build a
program. We're just going to doit for you, for you, Mr.
Customer. So we build theprogram. So we kind of walk once
the deal is almost done, orwhatever, it's in contracts, we
say, hey, look, we have a team,they're gonna engage with you.
And they're gonna help you getto that finish line, whatever
(22:03):
that means to you. Right? Itcould be bring these 10 teams
over to eat gas, perhaps Right?Or whatever, or on prem
Kubernetes doesn't. So we didthis for the following reason,
because it is hard, right? Like,everything boils down to, this
is new. I know that we have beentalking about Kubernetes as a
community for I don't even know,whatever, whatever, no, six
years something like that sucks.Yeah, I meet customers all the
(22:26):
time, who say, Oh, I've beenworking on this for seven years?
And I'd actually tell themhonestly, seven years ago, I had
no idea. I never heard of it. Iheard about Kubernetes, five
years. All right. Okay. So it'snew, right? Even now, many,
many, many enterprises who areworking on Kubernetes are
probably in the first year ofusing. So how can they know?
(22:46):
Right? These are things that arenot obvious, right? How do you
build a practice? aroundKubernetes? And this is where we
see a lot of toil, right? Thisis where we see projects fail.
in companies where they buysomething, and then well,
nothing happens. Right. And it'snot that the product is good or
bad, right? You've seen enoughof our competitors kind of go
through this and we're learningfrom, you know, you know, we're
(23:08):
standing on the shoulders ofgiants, as they say, and we're
learning from others who havecome before us. And the biggest
mistake, I see the two mistakesI see that have been made in
this industry by other vendors.One is that they really focus on
Kubernetes and not Kubernetesmanagement. And this is what we
discussed earlier in the call,what is the distinction between
just a cluster and all the otherthings that need to happen. And
(23:30):
the second one was they did notinvest in helping their
customers. They relied on, youknow, systems integrators or
somebody else your problem, gofigure. And that's a hit or
miss. Don't do that be happy tohelp our customers get up and
running. Selfishly, obviously,because you want to make money
and you don't want to see churn.But but more importantly,
(23:50):
because our customers spend two,three months doing a POC, they
spent money, we want to makesure they get value out of the
money, and not just park oursolutions on our shelves or
whatever. Right? This issomething that has been really
important. It's been a massivegame changer for us as a
company. Because yeah, thatcustomers actually use a product
and they're happy. And happycustomers will people who are
(24:13):
right, because they're solving aproblem. But this this is I'm
telling you the story becauseall of this ties back to the
complexity, the the the unknownsin this space, what do I have to
do? Right? And if we've seenthis enough times are we have
enough customers, if we canessentially educate our
(24:34):
customers and we don't chargefor this, right? This is part of
our our engagement, right? We'regonna help you come up with a
framework that we've seen workand I don't know 10s and 10s of
other companies who want to dosomething else that's okay. But
here's what others have done.Here's how their standard
operating procedure or centeroperating model ends up being as
(24:56):
it relates to Kubernetesmanagement and you know, please
consider taking these actionsunder reality is most customer
said, This is awesome. This isgoing to make my life easy. I
don't need to invent a process,you are giving me a process. We
do this because of the problemyou described, which is, yeah,
complexity. So many things thatare not known people are
learning on the job, becausewell, how can they not happen to
(25:16):
know before? Right? We can'texpect them to know these
things. Right? Now. It'spossible. They've never done
this. It's unfair to expectthis, from engineers who tell
perhaps a year ago, we'reworking on your cloud, or even
VMware infrastructure in ourdata center, we can't expect
them to become PhDs inKubernetes, or that's completely
unfair. We have to help. Right?And we do. So I love
Eveline Oehrlich (25:43):
it. So there's tons of knowledge transfer
during those projects, I'massuming or do during the
activities. Now I've heard,we've heard platform
engineering. And we've seenquite a few organizations adopt
platform engineering, when wedid a survey on sre. We saw that
a very, let's say, notpredominant yet, but a very
(26:07):
popular way to shape and framean organization. And then the
other term is developerexperience. Those are two terms,
which every day we get aquestion on. Okay, how do we
start? What do we do? I wouldlike you to just kind of reflect
a little bit on the two terms,developer experience, of course,
(26:29):
and then the platformengineering, particularly there,
maybe drill down a little bitinto the pros and cons, what do
you see, relative to platformengineering?
Haseeb Budhani (26:40):
Yeah, so our customer, in pretty much every
account is a platform engineer,to be sent to. And we look for
them. Sometimes the name of theteam perhaps is not, not from
engineering. I mean, you know,they may call themselves
infrastructure engineering, orcloud operations or something,
but but they are the ones whoare responsible for delivering,
(27:04):
essentially, these platformties, concepts, and Kubernetes
is one of them, because theydon't want to have and maybe we
can take a step back and thinkabout why is this even
happening? Why does a batonexist, which, which is a, which
is a really interesting thing tothink about what has happened
beforehand, at least in thecontext of Kubernetes, it
(27:27):
applies elsewhere, as well, I'msure it's clearly applies to
cloud, what's happened is somedevelopment team, they decide,
hey, these container things arepretty awesome. They allow us to
move very fast. So you knowwhat, we're gonna do this, we're
just gonna do this containersaying and they go to it, or at
that time, maybe cloudengineering, and this entity
(27:48):
isn't anything like Kubernetes.And they probably say, No, this
is not something that we providetoday, they go, no problem, I'll
just do it myself. But thesedevelopers essentially go to the
cloud and their developers, butthey'll figure it out. They go
provision infrastructure,Kubernetes. And they now run
their applications. Okay, now asecond team shows up and says,
Well, this is really cool, weshould do the same. And then a
(28:09):
third one. Now, what's going tohappen is the skill set is going
to perhaps vary between theseand teams. But here's the most
important thing. You are nowmissing resources, by
definition, because threedifferent teams are doing
business, in a sense is not goodfor your business. Somebody
should do this, if indeedmultiple teams needed, it should
(28:30):
happen separately, is simple. Itshould happen centrally, why
would you have every team dothis, because then the processes
are, are not standardized. We noone will use one methodology and
one will use another and why letthem do the thing you pay them
for, which is not Kubernetes.It's the application, let them
write the app, you do theKubernetes for them. In comes
(28:53):
about from engineering. But nowthe platform engineering
organization is not just solvingfor, I just need an app
deployed, they have to solve forthe other enterprise problems
that we talked about earlier.Right, because they're part of
the IT function. And they haveto think in terms of risk and
compliance and security. Allthese other things have to be
thought about as well thatperhaps the developers will not
prioritized because they want towrite their application. This is
(29:15):
why platform engineering teamsexist. And these are the
customers that we sell to and wemake them successful. But here's
the thing, we have to reallyunderstand who is our customers
customer? So our customersplatform engineering, who is
their customer, the developer,if there are no developers who
want to use containers andKubernetes. Well, there's no
(29:37):
need for any of this stuff.Right? What is the point of
Kubernetes? If no, nobody's evergoing to run things, right? So
the developer has to actually bereally happy with the experience
that they get when they interactwith this new platform from
vendor x. Right. Let's not bringit off into this. So it behooves
us to understand what is thedeveloper need? Because if
(29:57):
they're not happy, well, end ofthe day, there is no sale here.
So we have to invest time inthis. And we have, what do they
need? What is their pipeline ofchoice? What is their experience
expectation around debugging ofapplications? What is the
experience expectation aroundself service, right, some things
they want to do themselves. Andsometimes they want to do a lot
(30:19):
of things themselves, and thecompany is okay with it.
Sometimes the company makes adecision, these things platform
or do these other thingsdevelopers can do. And look at
turns out, we have to be in themidst of that, if we are not,
we're very, very ignoring,actually the most important
constituent here, the developer,who is not our direct customer,
(30:40):
and that's okay. But he's ourcustomer, customer, she or they
are our customers customer. Andwe have to understand what it
takes for the platform team tomake their life better. And we
spend a lot we spend inordinateamount of time, try to make that
better and better and better. Sotoday, we be proud a lot of
capabilities to make developersreally happy in our platform.
(31:04):
And we continue to makeinvestments like as an example
var, spending a lot of timelooking at something like
backstage that I'm sure you'resure you've seen backstages? Or
is an overnight sensation in ourindustry to understand what does
it take to help teams Aaronbackstages service in house and
then be, you know, make it easyfor them to use the Kubernetes
(31:28):
or other plugins that they needto backstage and then allow them
to build their own dashboardswith backstage? Yeah, this is
like the two words you describemap like from engineering and
developer experience. These arethe two equally weighted, you
know, core tenets of ourplatform. If they're not, we're
never going to be a lastingcompany. You have soccer balls.
Eveline Oehrlich (31:51):
Right, beautiful. I have two more
questions. One is technical oneis fun. So let's do the
technical first, and we'll go tothe fun. You alluded a little
bit to the future of Kubernetes.So if you have to bring out that
crystal ball, what do you see inthe Kubernetes future?
Haseeb Budhani (32:11):
I think Kubernetes is going to be here
for a while. So so the future ofapplications, or any any modern
application, you know, in thenear term is going to have in my
opinion, you know, somefunctions component like a
lambda component in theapplication or some
(32:32):
microservices, there'll be somecontainer containers. And then
there'll be some managedservices, like they'll be using
rds, and AWS or whatever, right?So this was maybe these Kafka,
these are services that sort ofoutside of you. And my view of
the world is that the rightinfrastructure management
(32:53):
offering, I said, infrastructuremanagement, like Kubernetes is
going to help customersorchestrate all of these things.
You must, because you have tohelp truly help somebody, you
know, deploy and operate anapplication and then enable
developers to do this pilot. Soplatform and developers, we're
going to be the two incidents.Today, RAF is a Kubernetes
(33:15):
company. Five years from now, itshould not be recognized. Fibers
mineral raw fish should be amodern application
infrastructure company. Right.
Eveline Oehrlich (33:27):
Yeah, holistically, you're saying I
love that the holisticmanagement of infrastructure,
towards agility and whateverelse is necessary at the
organization's.
Haseeb Budhani (33:39):
I think that's where the mark is going for
sure. And you have to you haveto go there to work today.
Eveline Oehrlich (33:45):
Yep. Excellent. Now, the fun
question, what do you do? Yes.What do you find? If you don't
talk to customers, or you don'tthink about all these very geeky
things? What do you do for fun?
Haseeb Budhani (33:59):
Who has the time? We have two kids. You have
a 13 year old and a 10 yearoldso there's enough to this
that they have that we sort ofyou know, recently as you drive
them around, right, that's yourjob. Right. You're the
chauffeurs are here to bear. Youknow, we, you know, my son was
10 the a few months ago decidedhe's when I was a kid, I used to
(34:22):
play squash. He didn't like thelike badminton. As a family big
to play badminton at Stanford. Ilive in Menlo Park seven blocks
away from Stanford. Do that, youknow, these are the kinds of
things we do right. Everythingis outside of work. It's the
family. I mean, why do we do allthis work? Why do why do we work
this hard? It's because of thefamily so you know, all the rest
(34:45):
of the time goes.
Eveline Oehrlich (34:46):
Excellent. That sounds like me. I spend a
lot of time on the soccer fieldswith my two daughters. When
Yeah, lots of watching soccerand feeling sad about myself
that I'm so old and I couldn'tplay it anymore.
Haseeb Budhani (35:00):
Soccer last time I played I'm in my mid 40s
minor. My knees will give out.
Eveline Oehrlich (35:08):
I know. Now we're watching it on TV and the
last this last session or thelast championship was quite
exciting. Anyway, this has beengreat has said thank you so much
for joining me today on humansof DevOps podcast. I really
appreciate all of your insightsand I know our listeners do that
as well. Again, we've beentalking to Haseeb Budhani, co
(35:32):
founder and CEO of RafaySystems. Humans of DevOps
Podcast is produced by DevOpsInstitute. Our audio production
team includes Julia Papp andBrendan Lay. I'm the Humans of
DevOps Podcast executiveproducer evolutionarily, if you
would like to join us on apodcast, please contact us at
humans of DevOps podcast atDevOps institute.com. That's a
(35:55):
mouthful. I know. I'm EvelineOehrlich. Talk to you soon.
Narrator (36:01):
Thanks for listening to this episode of the Humans of
DevOps Podcast. Don't forget tojoin our global community to get
access to even more greatresources like this. Until next
time, remember, you are part ofsomething bigger than yourself.
You belong
You're listening to the Humans of DevOps Podcast, a
podcast focused on advancing thehumans of DevOps through skills,
knowledge, ideas and learning,or the SKIL framework.
Haseeb Budhani (00:16):
VR, even in this economy at a point where there's
not enough talent available inthis space, who understands and
can operate Kubernetes not justbring up the cluster right?
Again, these are simple things,like truly running this at an
enterprise scale is a very hardscale.
Eveline Oehrlich (00:33):
Welcome to the Humans of DevOps Podcast. I'm
Evelen Oehrlich, Chief ResearchOfficer at DevOps Institute. The
name Kubernetes originates fromGreek meaning Helmsman, or
pilot. I've also heardKubernetes, often described as
the Linux of the cloud. Andtoday is the most popular
(00:54):
container orchestration platformfor multiple reasons. Kubernetes
burst onto the it developerscene in 2014, when Google
released it as an open sourceversion of its Borg technology,
which which they developed as away to run 1000s of jobs and
applications across multipleclusters and machines. Since
(01:15):
then, the technology actuallyhas spread far and wide. It is a
key part of managingapplications in data via
containers, and Gartner projectsto reach this technology to
reach 944 million by 2024.Today, I'm excited to have with
(01:35):
us Haseeb Budhani, who is cofounder and CEO at Rafay
Systems. Hello, welcome to thepodcast. Haseeb.
Haseeb Budhani (01:44):
Hi, everyone. Nice to talk to you. And thank
you for having me.
Eveline Oehrlich (01:47):
Yes, I did some googling you and looked at
your of course, LinkedInprofile, you have done a lot of
things I had to scroll andscroll and scroll, all these
different types of things youwill have been doing maybe give
us a cliff notes of what haveyou been up to before you were
(02:09):
advice systems, the CEO andfounder?
Haseeb Budhani (02:14):
Sure, since undergrad, I have been fortunate
enough. In fact, while I was Iwas a senior in college, I did a
number of internships in my inmy time at USC. And the very
last internship I did was with acompany that was writing
(02:36):
shopping cart software. This isback this fall of 99. Not going
to exactly how old I am. Thatwas my senior in college. And
yeah, these guys were doing somecrazy things, writing this thing
called a shopping cart. And theywere desperately hiring
(02:57):
engineers. And they give meYeah. And it was a lot of fun.
It was it was crazy. So beforethen I'd worked at Cisco and
done a set internship at Ciscoand an internship at Ericsson.
And I had a sense for how thosecompanies weren't, these are big
companies. But this othercompany, it was just crazy.
These people didn't sleep. Theyjust kept burning code. And it
(03:17):
was awesome. So having once Ifinished that internship, then I
started applying for four fulltime jobs. I just wanted to work
at startups. I just wanted tofind companies and I didn't have
the concept or appreciation as a21 year or for for stock or
anything had no idea I caredabout was about these these
companies. Well, you know, theypay well, they don't pay as well
(03:39):
as a Cisco perhaps. But wow, youget to work on some amazing
things. And that's what I did.My first job out of college was
at a start as a startup, whichwas called Publix. And that
solve for single sign on. It'sone of the first companies that
were doing single sign on in ourindustry. So it was not a thing
at the time, but they were doingit under another startup. And in
(03:59):
the process. I think I just youknow, good or bad, I learned a
lot of a lot of differentthings. And that's been you
know, it was a lot of fun. Andtruly, I believe that had I not
done those things, I would nothave had the opportunity to work
on startups that last 1012 yearsthat I've been, you know, in
some fashion of founder orwhatever. Yeah, just small
(04:24):
decisions, or very smalldecisions can put you on these
paths. And I think back a lot tothat one specific afternoon when
we had this career day. And Idecided to talk to these crazy
people back at USC and thatbecame you know, my pad had a
mountain that hadn't taken a jobat Cisco because I had an offer
from Cisco when you when yougraduate come work here and I
didn't take it. Yeah, I alwaysworked at startups. Only work at
(04:48):
large companies through somepathway positions that have not
worked at I've never it's been along time since I've actually
applied for a job in a largecompany. That's become my my
personality but the The mostimportant thing I do here, or I
require companies is, as a CEO,my primary job is sales. Like my
job is to give our customers theclarity as to why the company
(05:15):
exists, why raffia exists? Well,like any other company, when
I've worked at why does thiscompany need to exist, rather,
and how this can make their lifebetter? By getting people to
their point of clarity so thatthey can go, I see how I could
use it. There's nothing like it.I mean, that's, that's what
drives me. Right? So I enjoythat more than anything else.
And particularly if you havethe, you know, somewhat of a
(05:36):
technical background, I happento have one. It's, it's great,
right? So anybody working on newtechnologies, we all have to
understand it. I mean, we cancreate technologies, we can sell
it, it doesn't really matter.It's like a tree that fell in
the forest, or it doesn'tmatter. But being able to build
something, and then being ableto articulate why. And why
(05:56):
should you care? Right, that is,arguably the more important
thing. And if you can figurethat out, you know, you're gonna
have a lot of fun in thisindustry.
Eveline Oehrlich (06:05):
That leads me to one of the factoids, I found
rough AI systems was actuallynamed as one of the hardest one
of the 10 Hottest humanitystartups of 2022. By CNR. It's a
channel magazine, but that'sgreat. So two questions for you.
(06:25):
What does Rafay Systems do?First, and what has put you on
the list of the top 10 startupsand like you said, of course,
your startup, your passion forstartups, probably has something
to do with it. But help help ourlisteners, as those are many of
them are developers, I owe, youknow, geeks and good geeks. Help
(06:48):
them understand what does RafaySystems do? And how did you make
it to this top 10 startups.
Haseeb Budhani (06:56):
So Rafay is in the Kubernetes management space.
Kubernetes is, oh my god,there's like hundreds of
companies, something inKubernetes. It's a very noisy,
very busy space. But when Ithink about what we do, like,
fundamentally, we solve a peopleproblem in this company. Every
(07:17):
enterprise, we telco, anycompany that has software that
they're deploying, goingforward, it will be
containerized, or some function,because they want move fast as a
company. And in many cases,majority of cases for companies
will pick Kubernetes, as ourorchestration engine, you
(07:38):
describe Kubernetes has been thefastest growing container
orchestration platform outthere. It's it's the de facto
standard, if not the standardpresently. But it brings with it
a number of complications. Justbecause you have an
orchestration engine doesn'tmean you as an enterprise can
consume it easily. You have anumber of constituencies inside
your company, there's security,there's operations as
(07:59):
developers, there's all theseother different people and
multiple developmentorganizations inside every
enterprise. They all have towork together somehow. It's a
people problem, everything is apeople problem. And to that end,
what you've what you'veattempted to solve for, and it
seems like you've done a reallygood job, and you have a really
nice roster of customers who usethe product daily, is that
(08:21):
they've thought through what ittakes for an enterprise to
really build shared servicesplatform for Kubernetes. How do
you deliver Kubernetes as aservice inside your enterprise?
And what that means is not justthe material cluster, that's the
easy part, building a cluster orcreating a Kubernetes cluster.
That's not the issue. That's asolved problem. In my mind, the
(08:42):
real problem is, who has accessto what do we know what they're
doing? Can we do thisconsistently? Can we can we
update patches consistentlyacross our fleet of clusters?
can I provide different levelsof access to different teams? If
certain teams have networkingrequirements that are different
from other network teams? Howabout how do I make that happen?
(09:03):
And how do I do all of thiscentrally with the right level
of governance? This is howenterprises pay. If you solve
that problem in the context ofKubernetes, of course, it's
solving a technology problem.But really, you're solving a
people problem. Because theskill set is it takes time to
build a skill set. We are evenin this economy at a point where
(09:24):
there's not enough talentavailable in this space, who
understands and can operateKubernetes not just bring up a
cluster right? Again, these aresimple things like truly running
this at an enterprise scale is avery hard skill. Every
enterprise is looking for thosepeople and they're not able to
find them. And what we aretelling them is we will help you
(09:44):
augment your teams with softwareautomation. That's what we sell.
Fundamentally what we're sellingis automation that augments an
existing team. And the beauty ofthe right automation is that a
the enterprise they get up andrunning now. They don't need to
wait a while to hire Are peopleand then build a platform
because we sell them a platform.But to me the more important
thing, and this, I think, longterm is the right way to think
(10:06):
about any technology, the peoplein the organization that are in
our customer, who are notexperts at Kubernetes, by
working with our product, and byworking with our support
organization, they actuallyended up becoming experts. If
you can, in sort of, indirectlyor perhaps as a byproduct, get
(10:27):
people, you know, sort of, youknow, adept at Kubernetes, and
all the things that happenedaround it, right. That is, that
is generally good for our forour community, all right,
industry, right. So we sell agreat product, the enterprise is
happy, because they can now movemuch faster, right? They're TCO
is lower because they can dothis today, etc. They don't need
(10:48):
to wait to hire another 510 15people. But the existing IT
engineers and I'm using the wordit very loosely, we call them
DevOps, right depends on the onthe function they have. And then
broadly speaking it, they allget to learn this new
technology, which is going to bewith us for at least 10 years,
if not going to be all over timeafter learn this. And yeah, we
(11:10):
take pride in saying that ourcustomers, their engineers, you
know, months into ourengagement, Rafi, yes, the
enterprise is better off, butthe engineers are better off to
Eveline Oehrlich (11:20):
Yeah, we'll get to the future, hold that
thought on where this is going.Because that I want to dive into
that a little bit deeper. Butback to what you were just
saying. So we do now 22 was agreat year for Kubernetes. And
we've had in our organization,lots of questions for upskilling
in this topic. While it wasinitially viewed as something
(11:42):
only really large enterprisescould benefit from, we know that
it has improved in usability,most likely because as you said,
skills have gone up. But thereare still some challenges. Now,
there are some technicalchallenges with it. What would
you say if you think about yourclients, those you speak to
every day? What are some of thebiggest challenges leveraging
(12:05):
Kubernetes these days, and justfor right now just focus on the
technology in itself? Because Iknow there's a few. And then
we'll move further on once we'redone with that towards the you
already said that, which ismusic to my ears, the skills and
the skill development and thereduction of toil and all of
that? Well, we'll get to thatright now. Let's just focus and
hone in a little bit on thetechnical challenges you you
(12:28):
see.
Haseeb Budhani (12:29):
Yeah, absolutely. So many people ask,
you know, what does Rafay do,and I use the phrase Kubernetes
management, I didn't use thephrase Kubernetes for the
following reason. In my mind,the the biggest player in the
Kubernetes space is AWS, theyprovide an engine or product
(12:53):
called COVID PKS elasticKubernetes. Service, which is,
as far as I know, the most usedKubernetes offering right now.
So Amazon is the Kubernetescompany. So then what is rapid?
So once a customer decides I'mgoing to use maybe Amazon's
(13:13):
Kubernetes, maybe somethingelse, Azure, then they start a
journey, where they have to nowfigure out, okay, how do I
automate the provisioning ofthese clusters? There's
automation for that, and thenpeople who know different types
of technologies, right,TerraForm, etc, then I need to
understand how to upgrade thesethings. Okay, there's automation
(13:35):
for that, that you have to kindof figure out, then I need to
understand what are thecomponents that need to run on
this Kubernetes cluster, so thatmy applications can consume it.
They're all sort of raw out ofthe gate. So then you have to
learn these things. ThenSecurity says, Well, you really
need to make sure the rightpeople have access to the right
thing. So we have to think aboutwhat based access control and
the right level of identity andthe right level of access, tie
(13:57):
this back to the enterpriseSingle Sign On system. We should
really audit everything. who'sdoing what? Okay. All right,
let's go figure that out. Whilewe're on the point of
Kubernetes, is deployapplications. So case, we should
have pipelines of some sort,connecting back to maybe GitHub
or GitLab. Right? So we got tofigure that out. Hey, we have
(14:18):
certificates that we're pushinginto these clusters, for, you
know, TLS termination, we reallyshould think about some sort of
secrets management. Okay, well,let's go figure that out. And
chargebacks, service mesh,network policies and Kubernetes
policies, which is differentfrom their policies, and then
(14:39):
really, developers need access.So we should really think about
it the right developerexperience. And so each of these
things, and there's, dependingon the situation, there could be
many other things to do. Andthis is a challenge right now in
this industry, right? Nobody'swritten a book. I'm sure there's
a Kubernetes for Dummies, I'msure there is. But but it
doesn't talk about what it takesto not just get the toy up and
(15:02):
running. But to truly consumethis as an enterprise. By that
framework, what I just describedas a is a function of mine just
experience working withcustomers. But you know, there's
no Bible here. Maybe you shouldwrite one, I don't know. But
this is the challenge, right?And we're and the worry I have
right now is that whenenterprises jumped into this,
they don't know all thesethings. Now, could they know,
(15:23):
they just started, right? Andinitially, everything seems
easy. What's the big deal? Ijust go to the console, I build
a cluster, boom, boom, boom,it's working. See, I can do
Kotlin. But then start the realchallenge. Right, and now it
takes a year, who knows how longit's going to take depends on
the size of the company. Andthat's not okay. Right? This is
this is a detriment, right? Thisis going to get in the way of
(15:45):
progress. Right. So obviously,when when you see gaps startups
come about, and Rafi solve thatproblem, our job is a, we're
going to make for example, Eks,or Azure Kubernetes, which is
AKs, we're gonna make it out ofthe box enterprise ready, and
we're gonna help you manage nnumber of views across clouds if
you want. So we solve thatspecific gap. But but this is a
(16:07):
really hot technology promise.So our customers are pretty
sophisticated engineers, andthey're all very smart people.
This is what I actually reallyenjoy about this, this specific
job. I'm not trying to, I don'tneed to convince anybody of the
problem, right? We have openconversations, and they get to
the point of clarity, I get whatthe gaps are. But but sometimes
(16:28):
they don't know what the gapsare, unless they experience
that. But experiential learningtakes time. Right? You're going
to take a year, 18 months tofigure out what you don't know,
or what is missing. But yourcompany just launched 18 months,
this is the technology issue,right now in our industry. But
we're a small company,relatively speaking. So it's not
like I can I can solve this foreverybody. But this is what I
(16:48):
think about all day long. Bythis. Sometimes we as an
industry have the you know, wedo something that is bad, which
is reserved, you know,trivialize the the, you know,
the complexity. Everything iseasy. Look how easy it is look,
one command, and it's all done.Now, let's just have real
conversations about yes, indeed,this is hard. That's okay.
(17:10):
Because it's hard doesn't meanit's bad. It just means I need
to step back and think about howam I going to solve for the
complexity? I think it's betterto have that open conversation
versus try to kind of, you know,hide this or not hide as hide is
a strong word, but just, youknow, trivializing this or, or
trying to minimalize This is notokay. Right. And I truly believe
(17:31):
that this is hurting a lot ofenterprises or companies in
general, who are starting tojoin it because they are not
aware of how complex thattechnology is. And if they were,
they would make betterdecisions. And maybe some of
them will say, this is not forme, that's okay. But still, we
have to give them that clarityupfront, it's complex, here are
ways to solve it. That will overkind of, you know, at the macro
(17:53):
level that will help thisindustry more than a specific
vendor who's trying to get to asale. Don't think like that, I
think at the at the macro level,give people that education and
that clarity, and I think that'sgood for all.
Narrator (18:09):
SKILup IT Learning is the perfect online destination
to learn about DevOps anddigital transformation anytime,
anywhere. A digital learningplatform provides you with all
the resources you need toupskill and learn about these
important topics, includingexpert led course videos, access
to certification, prep courses,and a community of supportive
(18:31):
peers. A subscription basedmodel makes it easy for you to
gain the skills needed forsuccess in the modern workplace.
Visit DevOpsInstitute.com nowand explore our different
learning plans, get 30% off,SKILup IT Learning Plus now,
claim your offer and startlearning today with code Feb. 30
off, that's F E B three zero, of f.
Eveline Oehrlich (18:56):
And I think as technologists, I've seen this
over and over in conversationswith CIOs and VPs where they
have been promised by theirteams that oh, yeah, this shiny
new object can help usimmediately to increase flow,
velocity, speed, quality,whatever. And those executives
(19:16):
promise that onto the partner inbusiness, and six months into
the journey or 12 months, maybelonger even then they are all
disappointed because it hasn'tachieved the outcome they have
been thinking about right. And Ithink you hit as we say in
Germany, the nail right on itshead where we sit everywhere.
(19:38):
Yeah, we you need to berealistic about setting
expectations and making surethat we have these
conversations. Sometimes they'rehard. You did mention a few
things earlier, around thechallenges relative to people
and organizations, right you andin your examples, I could hear
some of that because we've stillsilos now Add sometimes a bad
(20:00):
word because we have experts incertain areas, right? Even
though we have dev SEC ops, thesecurity team is not yet as
integrated as it should be. Andso my curiosity is around people
and organizational challengesrelative to Kubernetes. In
bringing in such technology,what do you see your clients do?
And what, what's what works?Because many of our listeners,
(20:24):
I'm sure say, Yeah, okay, we'vedone this, but it fell flat on
its head, and it hasn't failed.Because it's Cuban Eddie's which
didn't work. It failed forcultural reasons.
Haseeb Budhani (20:37):
Yeah, there's, there's some influence we have.
And there's, of course, youknow, as a vendor, some
influences we don't have when weget to see a lot of stuff. What
one investment we've made inthis company, which is perhaps
unorthodox is that we built aservice delivery team, in the
company. And the point of thisorganization is, when a customer
(20:59):
says, people do talk to them, wesort of, you know, agree that
they should try the product, andthen we try the product. And so
this is pretty awesome, I shouldbuy it. And then they buy it.
And then what, right, justbecause you did a POC doesn't
mean it's it's plugged into yourinto your company, right? That
that's a process, right? A lotof people have to be educated to
bring developers and say, Hey,here's a, here's a platform, and
(21:22):
you want to think about ABC. SoSo we saw that happen again, and
again, where the where thecustomers would basically go
through this internal process,right, where they try to build a
framework, and they try toinvite developers and try to
sort of, you know, convince thedevelopers internally, hey,
don't do this yourself anymore.It's waste of your time, here's
(21:43):
a better platform, and we'regonna give you all the
automations. So we decided, youknow, what, we're gonna build a
program. We're just going to doit for you, for you, Mr.
Customer. So we build theprogram. So we kind of walk once
the deal is almost done, orwhatever, it's in contracts, we
say, hey, look, we have a team,they're gonna engage with you.
And they're gonna help you getto that finish line, whatever
(22:03):
that means to you. Right? Itcould be bring these 10 teams
over to eat gas, perhaps Right?Or whatever, or on prem
Kubernetes doesn't. So we didthis for the following reason,
because it is hard, right? Like,everything boils down to, this
is new. I know that we have beentalking about Kubernetes as a
community for I don't even know,whatever, whatever, no, six
years something like that sucks.Yeah, I meet customers all the
(22:26):
time, who say, Oh, I've beenworking on this for seven years?
And I'd actually tell themhonestly, seven years ago, I had
no idea. I never heard of it. Iheard about Kubernetes, five
years. All right. Okay. So it'snew, right? Even now, many,
many, many enterprises who areworking on Kubernetes are
probably in the first year ofusing. So how can they know?
(22:46):
Right? These are things that arenot obvious, right? How do you
build a practice? aroundKubernetes? And this is where we
see a lot of toil, right? Thisis where we see projects fail.
in companies where they buysomething, and then well,
nothing happens. Right. And it'snot that the product is good or
bad, right? You've seen enoughof our competitors kind of go
through this and we're learningfrom, you know, you know, we're
(23:08):
standing on the shoulders ofgiants, as they say, and we're
learning from others who havecome before us. And the biggest
mistake, I see the two mistakesI see that have been made in
this industry by other vendors.One is that they really focus on
Kubernetes and not Kubernetesmanagement. And this is what we
discussed earlier in the call,what is the distinction between
just a cluster and all the otherthings that need to happen. And
(23:30):
the second one was they did notinvest in helping their
customers. They relied on, youknow, systems integrators or
somebody else your problem, gofigure. And that's a hit or
miss. Don't do that be happy tohelp our customers get up and
running. Selfishly, obviously,because you want to make money
and you don't want to see churn.But but more importantly,
(23:50):
because our customers spend two,three months doing a POC, they
spent money, we want to makesure they get value out of the
money, and not just park oursolutions on our shelves or
whatever. Right? This issomething that has been really
important. It's been a massivegame changer for us as a
company. Because yeah, thatcustomers actually use a product
and they're happy. And happycustomers will people who are
(24:13):
right, because they're solving aproblem. But this this is I'm
telling you the story becauseall of this ties back to the
complexity, the the the unknownsin this space, what do I have to
do? Right? And if we've seenthis enough times are we have
enough customers, if we canessentially educate our
(24:34):
customers and we don't chargefor this, right? This is part of
our our engagement, right? We'regonna help you come up with a
framework that we've seen workand I don't know 10s and 10s of
other companies who want to dosomething else that's okay. But
here's what others have done.Here's how their standard
operating procedure or centeroperating model ends up being as
(24:56):
it relates to Kubernetesmanagement and you know, please
consider taking these actionsunder reality is most customer
said, This is awesome. This isgoing to make my life easy. I
don't need to invent a process,you are giving me a process. We
do this because of the problemyou described, which is, yeah,
complexity. So many things thatare not known people are
learning on the job, becausewell, how can they not happen to
(25:16):
know before? Right? We can'texpect them to know these
things. Right? Now. It'spossible. They've never done
this. It's unfair to expectthis, from engineers who tell
perhaps a year ago, we'reworking on your cloud, or even
VMware infrastructure in ourdata center, we can't expect
them to become PhDs inKubernetes, or that's completely
unfair. We have to help. Right?And we do. So I love
Eveline Oehrlich (25:43):
it. So there's tons of knowledge transfer
during those projects, I'massuming or do during the
activities. Now I've heard,we've heard platform
engineering. And we've seenquite a few organizations adopt
platform engineering, when wedid a survey on sre. We saw that
a very, let's say, notpredominant yet, but a very
(26:07):
popular way to shape and framean organization. And then the
other term is developerexperience. Those are two terms,
which every day we get aquestion on. Okay, how do we
start? What do we do? I wouldlike you to just kind of reflect
a little bit on the two terms,developer experience, of course,
(26:29):
and then the platformengineering, particularly there,
maybe drill down a little bitinto the pros and cons, what do
you see, relative to platformengineering?
Haseeb Budhani (26:40):
Yeah, so our customer, in pretty much every
account is a platform engineer,to be sent to. And we look for
them. Sometimes the name of theteam perhaps is not, not from
engineering. I mean, you know,they may call themselves
infrastructure engineering, orcloud operations or something,
but but they are the ones whoare responsible for delivering,
(27:04):
essentially, these platformties, concepts, and Kubernetes
is one of them, because theydon't want to have and maybe we
can take a step back and thinkabout why is this even
happening? Why does a batonexist, which, which is a, which
is a really interesting thing tothink about what has happened
beforehand, at least in thecontext of Kubernetes, it
(27:27):
applies elsewhere, as well, I'msure it's clearly applies to
cloud, what's happened is somedevelopment team, they decide,
hey, these container things arepretty awesome. They allow us to
move very fast. So you knowwhat, we're gonna do this, we're
just gonna do this containersaying and they go to it, or at
that time, maybe cloudengineering, and this entity
(27:48):
isn't anything like Kubernetes.And they probably say, No, this
is not something that we providetoday, they go, no problem, I'll
just do it myself. But thesedevelopers essentially go to the
cloud and their developers, butthey'll figure it out. They go
provision infrastructure,Kubernetes. And they now run
their applications. Okay, now asecond team shows up and says,
Well, this is really cool, weshould do the same. And then a
(28:09):
third one. Now, what's going tohappen is the skill set is going
to perhaps vary between theseand teams. But here's the most
important thing. You are nowmissing resources, by
definition, because threedifferent teams are doing
business, in a sense is not goodfor your business. Somebody
should do this, if indeedmultiple teams needed, it should
(28:30):
happen separately, is simple. Itshould happen centrally, why
would you have every team dothis, because then the processes
are, are not standardized. We noone will use one methodology and
one will use another and why letthem do the thing you pay them
for, which is not Kubernetes.It's the application, let them
write the app, you do theKubernetes for them. In comes
(28:53):
about from engineering. But nowthe platform engineering
organization is not just solvingfor, I just need an app
deployed, they have to solve forthe other enterprise problems
that we talked about earlier.Right, because they're part of
the IT function. And they haveto think in terms of risk and
compliance and security. Allthese other things have to be
thought about as well thatperhaps the developers will not
prioritized because they want towrite their application. This is
(29:15):
why platform engineering teamsexist. And these are the
customers that we sell to and wemake them successful. But here's
the thing, we have to reallyunderstand who is our customers
customer? So our customersplatform engineering, who is
their customer, the developer,if there are no developers who
want to use containers andKubernetes. Well, there's no
(29:37):
need for any of this stuff.Right? What is the point of
Kubernetes? If no, nobody's evergoing to run things, right? So
the developer has to actually bereally happy with the experience
that they get when they interactwith this new platform from
vendor x. Right. Let's not bringit off into this. So it behooves
us to understand what is thedeveloper need? Because if
(29:57):
they're not happy, well, end ofthe day, there is no sale here.
So we have to invest time inthis. And we have, what do they
need? What is their pipeline ofchoice? What is their experience
expectation around debugging ofapplications? What is the
experience expectation aroundself service, right, some things
they want to do themselves. Andsometimes they want to do a lot
(30:19):
of things themselves, and thecompany is okay with it.
Sometimes the company makes adecision, these things platform
or do these other thingsdevelopers can do. And look at
turns out, we have to be in themidst of that, if we are not,
we're very, very ignoring,actually the most important
constituent here, the developer,who is not our direct customer,
(30:40):
and that's okay. But he's ourcustomer, customer, she or they
are our customers customer. Andwe have to understand what it
takes for the platform team tomake their life better. And we
spend a lot we spend inordinateamount of time, try to make that
better and better and better. Sotoday, we be proud a lot of
capabilities to make developersreally happy in our platform.
(31:04):
And we continue to makeinvestments like as an example
var, spending a lot of timelooking at something like
backstage that I'm sure you'resure you've seen backstages? Or
is an overnight sensation in ourindustry to understand what does
it take to help teams Aaronbackstages service in house and
then be, you know, make it easyfor them to use the Kubernetes
(31:28):
or other plugins that they needto backstage and then allow them
to build their own dashboardswith backstage? Yeah, this is
like the two words you describemap like from engineering and
developer experience. These arethe two equally weighted, you
know, core tenets of ourplatform. If they're not, we're
never going to be a lastingcompany. You have soccer balls.
Eveline Oehrlich (31:51):
Right, beautiful. I have two more
questions. One is technical oneis fun. So let's do the
technical first, and we'll go tothe fun. You alluded a little
bit to the future of Kubernetes.So if you have to bring out that
crystal ball, what do you see inthe Kubernetes future?
Haseeb Budhani (32:11):
I think Kubernetes is going to be here
for a while. So so the future ofapplications, or any any modern
application, you know, in thenear term is going to have in my
opinion, you know, somefunctions component like a
lambda component in theapplication or some
(32:32):
microservices, there'll be somecontainer containers. And then
there'll be some managedservices, like they'll be using
rds, and AWS or whatever, right?So this was maybe these Kafka,
these are services that sort ofoutside of you. And my view of
the world is that the rightinfrastructure management
(32:53):
offering, I said, infrastructuremanagement, like Kubernetes is
going to help customersorchestrate all of these things.
You must, because you have tohelp truly help somebody, you
know, deploy and operate anapplication and then enable
developers to do this pilot. Soplatform and developers, we're
going to be the two incidents.Today, RAF is a Kubernetes
(33:15):
company. Five years from now, itshould not be recognized. Fibers
mineral raw fish should be amodern application
infrastructure company. Right.
Eveline Oehrlich (33:27):
Yeah, holistically, you're saying I
love that the holisticmanagement of infrastructure,
towards agility and whateverelse is necessary at the
organization's.
Haseeb Budhani (33:39):
I think that's where the mark is going for
sure. And you have to you haveto go there to work today.
Eveline Oehrlich (33:45):
Yep. Excellent. Now, the fun
question, what do you do? Yes.What do you find? If you don't
talk to customers, or you don'tthink about all these very geeky
things? What do you do for fun?
Haseeb Budhani (33:59):
Who has the time? We have two kids. You have
a 13 year old and a 10 yearoldso there's enough to this
that they have that we sort ofyou know, recently as you drive
them around, right, that's yourjob. Right. You're the
chauffeurs are here to bear. Youknow, we, you know, my son was
10 the a few months ago decidedhe's when I was a kid, I used to
(34:22):
play squash. He didn't like thelike badminton. As a family big
to play badminton at Stanford. Ilive in Menlo Park seven blocks
away from Stanford. Do that, youknow, these are the kinds of
things we do right. Everythingis outside of work. It's the
family. I mean, why do we do allthis work? Why do why do we work
this hard? It's because of thefamily so you know, all the rest
(34:45):
of the time goes.
Eveline Oehrlich (34:46):
Excellent. That sounds like me. I spend a
lot of time on the soccer fieldswith my two daughters. When
Yeah, lots of watching soccerand feeling sad about myself
that I'm so old and I couldn'tplay it anymore.
Haseeb Budhani (35:00):
Soccer last time I played I'm in my mid 40s
minor. My knees will give out.
Eveline Oehrlich (35:08):
I know. Now we're watching it on TV and the
last this last session or thelast championship was quite
exciting. Anyway, this has beengreat has said thank you so much
for joining me today on humansof DevOps podcast. I really
appreciate all of your insightsand I know our listeners do that
as well. Again, we've beentalking to Haseeb Budhani, co
(35:32):
founder and CEO of RafaySystems. Humans of DevOps
Podcast is produced by DevOpsInstitute. Our audio production
team includes Julia Papp andBrendan Lay. I'm the Humans of
DevOps Podcast executiveproducer evolutionarily, if you
would like to join us on apodcast, please contact us at
humans of DevOps podcast atDevOps institute.com. That's a
(35:55):
mouthful. I know. I'm EvelineOehrlich. Talk to you soon.
Narrator (36:01):
Thanks for listening to this episode of the Humans of
DevOps Podcast. Don't forget tojoin our global community to get
access to even more greatresources like this. Until next
time, remember, you are part ofsomething bigger than yourself.
You belong
Popular Podcasts
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
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.