1
00:00:00,000 --> 00:00:12,000
All right, welcome back, folks. We are here for another episode of Tanzu Talk. And with

2
00:00:12,000 --> 00:00:17,500
me today is Joe Ho from Great American Insurance Company. Say hello, Joe.

3
00:00:17,820 --> 00:00:18,100
Hello.

4
00:00:18,880 --> 00:00:23,700
So great to have you on. And this is kind of part of the series that we've been running,

5
00:00:23,700 --> 00:00:28,480
looking at folks who've transitioned from having more of an infrastructure background

6
00:00:28,480 --> 00:00:34,620
over to really having more of a platform focus and sort of the evolution that we've been seeing

7
00:00:34,620 --> 00:00:40,820
over the last, you know, five or so years where, you know, as folks are really trying to focus in

8
00:00:40,820 --> 00:00:48,360
on supporting developers, there's been a kind of change in the roles that people have. So with that,

9
00:00:48,360 --> 00:00:56,660
we'll go ahead. But my first question is usually around kind of walking through what sort of the

10
00:00:56,660 --> 00:01:02,340
day in the life before when you had more of an infrastructure background and feel free to kind

11
00:01:02,340 --> 00:01:08,080
of fill us in on, you know, what specifically your role was on the infrastructure side. I know when

12
00:01:08,080 --> 00:01:13,900
we spoke before, you mentioned, you know, kind of a really diverse background. We were talking

13
00:01:13,900 --> 00:01:21,220
networking. So I know that that came up. But yeah, kind of tell us about what where that came from,

14
00:01:21,220 --> 00:01:25,300
and how you kind of got into it. And then we'll get into what your day in the life was. But

15
00:01:25,300 --> 00:01:30,280
how did you how long did you do that for? And kind of what was that that role all about?

16
00:01:30,800 --> 00:01:40,100
Okay. Well, so I've been with Great American since, since I got out of college. I started in

17
00:01:40,100 --> 00:01:49,120
actually as an actuary, back when we only had green screens back in 1990. So I'm dating myself quite a bit

18
00:01:49,120 --> 00:02:00,280
there. But had an opportunity to put in the first server into the environment. And that was like,

19
00:02:00,280 --> 00:02:08,560
you know, we had a we had a product come in, that needed some storage, some shared storage. And we were

20
00:02:08,560 --> 00:02:16,420
like, how do we do this? Because it was back in the early days and cobbled together a server and

21
00:02:16,420 --> 00:02:25,200
some Arknet and installed Novell Network for this very, very focused delivery. So it was the first server

22
00:02:25,200 --> 00:02:35,440
in the in the company. And yeah, it was so, so it was really customer driven from day one, I guess.

23
00:02:36,400 --> 00:02:44,480
You know, it was like, find a need, and then fill that need. So throughout my career, that's kind of

24
00:02:44,480 --> 00:02:54,180
been my, my mantra, or, you know, or, you know, what, what is, uh, um, driven me to discover new,

25
00:02:54,180 --> 00:03:04,420
new technologies and explore new technologies. So, um, moved into, to networking around 2001. Um,

26
00:03:04,420 --> 00:03:12,280
I was given the opportunity to, uh, join an engineering group. Um, I had been working in networking and,

27
00:03:12,280 --> 00:03:16,680
uh, we had some changes in our organization and they were going to make an engineering team.

28
00:03:16,680 --> 00:03:23,580
So I decided to jump from, uh, from strip, from being focused on networking to move up there.

29
00:03:23,740 --> 00:03:29,040
And the first project I had was for monitoring. So we wanted to monitor our environment.

30
00:03:29,440 --> 00:03:36,800
Well, nobody wanted to monitor. Nobody. I was an intervener. I was, it was, it was very, um,

31
00:03:37,280 --> 00:03:47,340
frustrating. I was getting nowhere. Um, um, so we were a windows based environment and, um, it was,

32
00:03:47,340 --> 00:03:54,380
it was just a, an uphill struggle to get any movement because of political, you know, really it

33
00:03:54,380 --> 00:04:00,080
was political, uh, you know, that this other group was telling this other group how to, how to run

34
00:04:00,080 --> 00:04:09,880
things. So, um, um, I decided, well, I'm going to go down and join the infrastructure team to do the

35
00:04:09,880 --> 00:04:15,940
monitoring from the ground up. So I joined the server team on the Unix and Linux, well, Unix team

36
00:04:15,940 --> 00:04:22,820
at that point, cause Linux wasn't really a thing. Um, three people, me, including me. And, um, about

37
00:04:22,820 --> 00:04:34,840
two months later, the other two people left. So I was stuck with, um, some HP UX servers and, um,

38
00:04:35,180 --> 00:04:41,540
uh, a burgeoning Linux environment that I was trying to build out for monitoring and like figuring out

39
00:04:41,540 --> 00:04:49,080
how do I run this stuff? Um, so, um, starting from there is really where I kind of got into the,

40
00:04:49,080 --> 00:04:55,160
got, started thinking in a platform mindset, um, from a, from the monitoring perspective.

41
00:04:55,160 --> 00:05:02,620
Okay. Um, if that makes sense. Um, you know, like what are the, what are the, you know,

42
00:05:02,700 --> 00:05:11,700
core capabilities? How can we, how can we, um, provide, um, base levels of functionality across

43
00:05:11,700 --> 00:05:18,800
many different systems, whether it be, uh, a networking device or a, or a server.

44
00:05:19,080 --> 00:05:23,660
Um, because those were the main components at that point in time, maybe we had a small

45
00:05:23,660 --> 00:05:31,160
sand, but anyway, so just thinking like, you know, working with, uh, with consumers, but

46
00:05:31,160 --> 00:05:38,580
at their level, because the top down approach didn't really work to, to partner with them

47
00:05:38,580 --> 00:05:44,860
to say, what are the problems that you have? And how can we take some of that burden off

48
00:05:44,860 --> 00:05:50,560
of you and from a monitoring perspective? And we were doing it in, in Linux at that point

49
00:05:50,560 --> 00:05:58,020
in time, which was helping to build that platform out. Not too far after that, it was, it was,

50
00:05:58,020 --> 00:06:08,900
um, time for virtualization because, um, uh, I could not warrant purchasing hardware for all

51
00:06:08,900 --> 00:06:17,120
these little projects that I was having. So, um, um, it was time to, to make use of, uh, virtualization

52
00:06:17,120 --> 00:06:25,800
to get more out of less resources. And so, um, I brought in VMware and started to build a platform

53
00:06:25,800 --> 00:06:41,460
around that, how to, how to turn a, a set of, of activities for delivery into a consumable service

54
00:06:41,460 --> 00:06:48,740
that somebody, you know, with, with predefined inputs and outputs that is consistent every time.

55
00:06:48,740 --> 00:06:54,540
And, and can, can you give me an example of like, what's, what, what's, what were some of these

56
00:06:54,540 --> 00:06:59,980
activities? Like, well, I mean, I mean, it was not monitoring. No, well, we were focused on,

57
00:06:59,980 --> 00:07:08,140
so we, we were, we were building the, the monitoring was, uh, uh, uh, a use case for virtualization,

58
00:07:08,140 --> 00:07:14,660
but then once virtualization, well, you know, we, we had our first cluster, you know, we had our first

59
00:07:14,660 --> 00:07:20,760
non-prod VMware cluster and our, and our first prod VMware cluster. Then it was like, people started

60
00:07:20,760 --> 00:07:26,260
coming out of the woodwork saying, wow, you can deliver something that was very custom. And it took

61
00:07:26,260 --> 00:07:33,620
a long time to get servers onto the floor. And now you can turn that around very quickly. So we had,

62
00:07:33,620 --> 00:07:42,260
you know, we had lots of consumers of that product. And so I, I, my team worked to create

63
00:07:42,260 --> 00:07:48,740
the t-shirt sizes, if you will, or, you know, like, like you fill out this form and then you will have

64
00:07:48,740 --> 00:07:57,660
a VM and, in some, you know, piece of time you to, to basically, um, standardize the process versus

65
00:07:57,660 --> 00:08:05,480
everything being a customized tuned, everything was different, you know, um, you know, in the,

66
00:08:05,560 --> 00:08:09,160
in the old days, when you were building a physical server, it depends on who built the server

67
00:08:09,160 --> 00:08:15,060
as to how that server ended up being realized for all time.

68
00:08:15,180 --> 00:08:16,200
Lots of snowflakes.

69
00:08:16,600 --> 00:08:22,960
Lots of snowflakes. There you go. Um, so that, you know, seeing that from the monitoring side

70
00:08:22,960 --> 00:08:30,580
and realizing that we had to get better at consistency and standardization and seeing VMware

71
00:08:30,580 --> 00:08:37,540
as a platform on which we could do that with, um, at, you know, your gold images and, and whatnot,

72
00:08:38,180 --> 00:08:46,020
um, we, we built that out and it was, it, it was a beautiful thing. It was very rewarding and all

73
00:08:46,020 --> 00:08:54,660
along learning, you know, everybody that was involved in that learned a bunch of new skills at the same

74
00:08:54,660 --> 00:09:02,420
time, you know, learned about automation. We moved a data center, um, through automation from one data

75
00:09:02,420 --> 00:09:09,060
center to another using VMware and, you know, taking things down. So we had a big, a big project where

76
00:09:09,060 --> 00:09:18,660
we P to V to V to a bunch of things. We did vert, we did automation for that. So I guess, um, every,

77
00:09:18,660 --> 00:09:31,140
every step along my journey was filling in another puzzle piece and what would eventually be used in my

78
00:09:31,140 --> 00:09:39,780
current role as, as platform operations and building a platform on the, uh, Tonsu products previously

79
00:09:39,780 --> 00:09:44,740
pivotal. Okay. And, you know, just to kind of fill in a little bit of color, cause I'm curious,

80
00:09:44,740 --> 00:09:52,260
as you kind of built out that initial platform years ago on, on VMware for just, Hey, here's,

81
00:09:52,260 --> 00:09:59,260
instead of just a custom thing, every time someone wants a server, we can stamp out, uh, uh, you know,

82
00:09:59,260 --> 00:10:05,500
across a couple of base configurations, we can stamp out a, a VM for you. Who were the customers

83
00:10:06,300 --> 00:10:14,300
in, in that situation? Was, were it developed, was it development teams? Um, so most of, most of that

84
00:10:14,300 --> 00:10:22,780
time, um, we were delivering, this is, this is also where we started like an architecture, architecture,

85
00:10:22,780 --> 00:10:30,860
um, practice where, uh, customer would come in and that would be like a line of business.

86
00:10:30,860 --> 00:10:36,860
It wouldn't be really the business. Okay. It would still be somebody in it and they would say,

87
00:10:36,860 --> 00:10:45,500
Hey, we have this product that we just purchased. How do we get this in our environment? Um, sometimes we

88
00:10:45,500 --> 00:10:53,100
would be able to, um, um, get ahead of that and be part of the product evaluation because infrastructure

89
00:10:53,100 --> 00:10:59,820
is always, you know, in our environment, at least they're always the last to know that they're going

90
00:10:59,820 --> 00:11:08,540
to have, you know, something to, something to support and deploy and, and manage at the end of the day.

91
00:11:08,540 --> 00:11:14,300
And so the architectural side was trying to get ahead of that to say, Hey, get us in on the front

92
00:11:14,300 --> 00:11:21,180
end so we can understand before or while you're buying it, what the costs on the infrastructure

93
00:11:21,180 --> 00:11:25,900
are going to be and what the integrations are going to be, what operating systems, how big are they going

94
00:11:25,900 --> 00:11:32,860
to be? What do they need to talk to all these other kinds of, uh, um, non-functional requirements.

95
00:11:33,660 --> 00:11:42,380
So our customers were people that were trying to get their products deployed. And some of that

96
00:11:42,380 --> 00:11:49,820
was for development, you know, in-house development, but a lot of it was for, for COTS applications.

97
00:11:49,820 --> 00:11:58,620
Okay. And when we built, when we wanted to support development, those were pretty cookie cutter because

98
00:11:58,620 --> 00:12:05,660
we were a Tomcat shop or we are.net shop. So it'd be like, Hey, we need a Tomcat VM, you know,

99
00:12:05,660 --> 00:12:10,540
we're going to be running, uh, you know, we need a stack from dev all the way up to prod,

100
00:12:11,500 --> 00:12:18,460
make it so. And we'd build out, you know, 14 VMs for them and, and across those environments.

101
00:12:18,460 --> 00:12:24,300
Yeah. Okay. But the COTS actually dragged a bunch of specific requirements and...

102
00:12:24,300 --> 00:12:31,180
Exactly. Yep. Interesting. Okay. Yeah. That was one question, which I'm sure will come up

103
00:12:31,180 --> 00:12:36,620
again later in our conversation about, you know, who the, who the customer is. Um,

104
00:12:37,340 --> 00:12:44,300
and so, you know, if you think about, you know, it's, it's interesting, you, you described really this,

105
00:12:45,020 --> 00:12:52,140
this whole arc that's, you know, the, the industry has really gone through frankly, but I guess maybe

106
00:12:52,140 --> 00:13:00,940
thinking about the sort of most recent sort of time before, uh, before kind of engaging in sort of the

107
00:13:00,940 --> 00:13:08,780
new platform operations role that you've been in, what, what was your day kind of like at that point,

108
00:13:08,780 --> 00:13:16,380
you'd already sort of stood up the, um, the platform, the kind of VM based platform for years,

109
00:13:16,380 --> 00:13:20,780
and we're going to stamp that out that that's up and running. You already had that architecture

110
00:13:20,780 --> 00:13:24,860
practice to kind of get involved, but what was kind of the day to day like?

111
00:13:24,860 --> 00:13:32,780
It was a lot about, um, focusing on delivery of objects. Okay. It was very, um,

112
00:13:34,380 --> 00:13:42,700
or at least the, the, the, the, the people that were, that were, um, building the VMs are

113
00:13:42,700 --> 00:13:50,540
responsible for the virtualization environment or even hosting or networking. It's very object oriented. It's

114
00:13:50,540 --> 00:13:59,340
very, I am, uh, uh, um, I, I have to, I have to deploy this VM. I have to provision the storage.

115
00:13:59,340 --> 00:14:08,300
I have to create this F5 with no context as to what value it provides. Well, you know, what's the big

116
00:14:08,300 --> 00:14:16,140
picture of this thing? Um, you know, we, we, in the, from the architectural side, we try to have,

117
00:14:16,140 --> 00:14:20,700
you know, good documentation, uh, you know, an overall system diagram and stuff. But at the end

118
00:14:20,700 --> 00:14:27,340
of the day, that's really the, um, the way that it works is you're just given a unit of work and you

119
00:14:27,340 --> 00:14:36,140
just had to get that done. So, um, kind of, uh, probably less rewarding, um, because you don't see

120
00:14:36,140 --> 00:14:43,340
the full picture. And a lot of times it's, it would be like firefighting. Um, you know, that's the other

121
00:14:43,340 --> 00:14:51,100
part of it. You're either, you're either delivering things or you're firefighting for some unknown

122
00:14:51,100 --> 00:14:56,140
requirement that wasn't, you know, once you delivered it, all of a sudden they started to

123
00:14:56,140 --> 00:15:03,740
use it and the use case deviated from the original use case. And now all of a sudden, now it's, it

124
00:15:03,740 --> 00:15:11,020
doesn't fit. It's not fit for function. And that was always, I mean, it still is, I'm sure very frustrating.

125
00:15:11,020 --> 00:15:18,620
Yeah. Um, I can see how that, that happens. It's like, oh, I have this in place. Now I can use it

126
00:15:18,620 --> 00:15:23,420
for this other thing too. Exactly. Yeah. So we're like, I thought I was going to use it like this,

127
00:15:23,420 --> 00:15:28,540
but then the business requirements changed or somebody created different code. And now it's like,

128
00:15:28,540 --> 00:15:33,260
it's a, it's, it's, it's a completely different thing. And so we have.

129
00:15:33,980 --> 00:15:39,740
So this sort of is past or firefighting. Um, okay. Sorry. You were about to say.

130
00:15:40,220 --> 00:15:45,660
No. Um, I mean, we have, we have pretty stringent security requirements because we're in financial

131
00:15:45,660 --> 00:15:54,540
services and sometimes, you know, all of a sudden they'd want to consume something or connect to

132
00:15:54,540 --> 00:16:00,220
something in a non-standard way. And we're like, well, you can't get there from here, from our

133
00:16:00,220 --> 00:16:06,220
security constraints. And then that would cause all kinds of, you know, rework, which is, you know,

134
00:16:06,220 --> 00:16:12,700
something that I really, really hate to have to do. Um, I hate to have anybody else have to rework

135
00:16:12,700 --> 00:16:18,220
anything. Um, that's one of the good things about the platform is that it's built for rework.

136
00:16:18,940 --> 00:16:23,340
You know, it's built to automate any kind of rework that has to be done.

137
00:16:24,300 --> 00:16:29,180
You know, if you have to get, you just deploy, just deploy it again, just deploy another one that

138
00:16:29,180 --> 00:16:38,940
works right. And, you know, migrate users over. So the, the impedance, I guess, to having to rework

139
00:16:38,940 --> 00:16:46,940
is much less or not the impedance, but the, the effort to have to do rework is greatly lessened on

140
00:16:48,140 --> 00:16:53,340
the platform that we're running today versus what we do in the old environment.

141
00:16:54,060 --> 00:17:01,340
Okay. That's, that's super interesting as, you know, one of those almost like fundamental truths

142
00:17:02,780 --> 00:17:11,900
of in the, in the olden days, right. In the before times, we, you know, and it's not that it was

143
00:17:11,900 --> 00:17:16,620
actually true back then, but it was, it was the mindset and the assumptions, which is that we are

144
00:17:16,620 --> 00:17:22,380
going to build this system and we're going to build it. So it's resilient so that it doesn't have a

145
00:17:22,380 --> 00:17:30,220
failure and it's going to serve the purpose and it's that it's, you know, been designed for and, you

146
00:17:30,220 --> 00:17:36,460
know, and we're never going to have to rework it. And the reality is things will always fail. People

147
00:17:36,460 --> 00:17:41,500
are going to change what they're doing because, you know, the business isn't static. So why would the

148
00:17:41,500 --> 00:17:48,220
way people use their technology be static and things will have to get reworked? So kind of inverting

149
00:17:48,220 --> 00:17:54,140
that whole paradigm and saying, what if we just embraced the fact that things fail and they need

150
00:17:54,140 --> 00:18:00,780
to be reworked? Right. And so don't be afraid to blow it up because it can be recreated immediately.

151
00:18:00,780 --> 00:18:08,780
Yeah. Versus etched in stone. You know, they used to be children. I used to know every server name

152
00:18:08,780 --> 00:18:14,620
by heart because I was, I was intimately involved in what it did and why it's there and

153
00:18:14,620 --> 00:18:20,860
you know, everything about it. Yeah. Now it's just, Oh, that's just another app. Okay.

154
00:18:20,860 --> 00:18:24,700
Okay. We, Phyllis runs a little hot. We already know that about her, right?

155
00:18:24,700 --> 00:18:27,740
Yeah, exactly. And there's, there still is

156
00:18:28,620 --> 00:18:34,620
a lot of that because we, you know, we have, you know, 5,000 VMs that are still traditional

157
00:18:34,620 --> 00:18:43,980
because you can't run cots on the platform yet. Yeah. So now let's contrast that we're sort of

158
00:18:43,980 --> 00:18:51,420
starting to get there a little bit, but contrast more the day in the life for you, right? How you

159
00:18:51,420 --> 00:18:58,220
spend your time before kind of, Hey, we're stamping out tasks, sort of delivering objects, as you say,

160
00:18:58,860 --> 00:19:06,700
or we're firefighting, um, and potentially that leading to very painful rework. Now let's kind

161
00:19:06,700 --> 00:19:14,700
of fast forward. You're, you're kind of, um, on the, the operating a platform. Yeah. What does your day

162
00:19:14,700 --> 00:19:25,180
look like? Yeah. So, um, it's, it's very much focused on delivering capabilities, um, versus that,

163
00:19:25,180 --> 00:19:32,460
that can be used to automatically create objects. You know, it's not the individual object that matters.

164
00:19:32,460 --> 00:19:43,820
It's the capability to create objects. So, um, you know, we've got projects for, um, enabling, uh,

165
00:19:43,820 --> 00:19:50,460
new foundations. Uh, we're a pretty small shop. We've only got four foundations. Um, but we know that

166
00:19:50,460 --> 00:19:58,460
we're going to have more foundations. So, you know, how do we, how do we isolate those foundations away

167
00:19:58,460 --> 00:20:05,020
from the developers, um, that they don't have to know that they're foundations or, uh, we just enabled,

168
00:20:05,020 --> 00:20:15,900
um, the ability to use, uh, spring cloud data flow because that is a, that is a, uh, a capability that

169
00:20:15,900 --> 00:20:22,940
has business value. So, you know, how do we do that? How do we, how do we manage that? How do we scale

170
00:20:22,940 --> 00:20:28,860
it? What's it going to take? And so you're not really worried anymore about the individual instances

171
00:20:28,860 --> 00:20:34,140
of it. Although, you know, we try to understand architecturally what they're doing so that we

172
00:20:34,140 --> 00:20:40,620
understand the capacity, you know, the network capacity, the, the memory capacity of the platform that we need.

173
00:20:40,620 --> 00:20:47,900
But, um, at the end of the day, it's, it's much more rewarding to work with a business use case

174
00:20:49,020 --> 00:20:55,660
that says, Hey, we need to operate on this data to build this, to deliver this business value.

175
00:20:57,020 --> 00:21:03,980
And we think this is the right tool for it. How can the platform help us scale that?

176
00:21:05,020 --> 00:21:11,420
Um, so it's, it's, like I said, it's much more, much more focused on delivering capabilities than,

177
00:21:11,420 --> 00:21:19,020
than objects. Obviously you've got your regular, um, you know, you want to make sure the health of

178
00:21:19,020 --> 00:21:24,700
the platform is good. You look at that daily. Did my backups occur? You know, what are, is anything

179
00:21:24,700 --> 00:21:32,620
failing? Um, you, you get notified on those. Um, and then you've got developers that have, you know,

180
00:21:32,620 --> 00:21:43,020
that have fires, um, to them, but they're usually in dev and it's, and it's more of a more, more of a

181
00:21:43,020 --> 00:21:52,860
collaboration than it is a, an emergency because you're, you, you, we know what they're, we,

182
00:21:52,860 --> 00:21:58,700
we, we know the types of things that they're doing based on their consumption of the platform,

183
00:21:58,700 --> 00:22:07,580
um, ahead of time versus this wall between devs and devs and operations that exists in the more

184
00:22:07,580 --> 00:22:15,100
traditional, um, infrastructure roles. Yeah. That makes sense. Tell me, tell me a little bit more

185
00:22:15,100 --> 00:22:21,660
about that. So it's, you know, dealing with the, dealing with a fire in dev,

186
00:22:23,100 --> 00:22:29,340
how is it more collaborative and, and how is it that you already know kind of what's going on?

187
00:22:30,220 --> 00:22:32,940
How, how, how are you at that point? What's different?

188
00:22:32,940 --> 00:22:40,460
Well, we try to, we try to get before anything becomes targeted for the platform. We try to have

189
00:22:40,460 --> 00:22:45,740
a discussion with them. So we're, you know, we're going to create a new org and we're going to put

190
00:22:45,740 --> 00:22:51,820
some apps in it. Okay. We have, we try to have a discussion with them as it goes through an

191
00:22:51,820 --> 00:23:00,460
architectural gate in our, in our corporate, you know, process for, um, uh, project planning.

192
00:23:01,340 --> 00:23:08,380
So, um, we try to be at the table there to say, okay, what, you know, how many, how many of these

193
00:23:08,380 --> 00:23:12,540
things you're going to do? What kind of, how, how many apps are you going to have? How many APIs

194
00:23:12,540 --> 00:23:18,940
are you going to have? What's what services are you going to consume? So we have some like basic

195
00:23:18,940 --> 00:23:26,540
understanding of what the general consumption is. And then from a firefighting perspective,

196
00:23:26,540 --> 00:23:33,580
it's usually because we're still in the early, early phases. It's usually, um, about

197
00:23:34,620 --> 00:23:41,900
the developers understanding cloud native development and how, how to, how to,

198
00:23:43,180 --> 00:23:49,900
how to configure their projects so that they can plug into the concourse automation

199
00:23:50,620 --> 00:23:56,620
that we have and the, the ability to scale, you know, we have some, some issues with,

200
00:23:57,180 --> 00:24:05,020
you know, uh, you know, session IDs, uh, you know, not being stored in a, in a backend so that,

201
00:24:05,020 --> 00:24:09,420
you know, they, they can take available, uh, make available all the instances, this,

202
00:24:09,420 --> 00:24:15,260
the horizontal scaling that the platform provides. So it's a lot of education and it's good for both

203
00:24:15,260 --> 00:24:26,540
parties, I'll say. Um, because when you know a subject very deeply as, as our, our, our platform

204
00:24:26,540 --> 00:24:32,300
operations do and some developers, I mean, most developers do who have been on the platform,

205
00:24:32,860 --> 00:24:39,180
it's easy to assume that somebody new coming in also knows those intricacies.

206
00:24:39,180 --> 00:24:49,820
And so it's a, it's a, it's a, it's a, it's a, an opportunity to, um, educate, um, because I think

207
00:24:49,820 --> 00:24:56,620
that's a lot of what, uh, platform operations is and DevOps in general is it's educating each other

208
00:24:56,620 --> 00:25:03,020
because nobody knows everything. We all have more to learn. So that's, that's very rewarding. I think

209
00:25:03,020 --> 00:25:09,740
for, uh, at least that's the feedback I get from the other cloud ops, uh, the, the other, um, platform

210
00:25:09,740 --> 00:25:17,980
operations folks is they always get to learn something new and they also have the ability to

211
00:25:17,980 --> 00:25:26,540
provide that information. And we have documentation, um, for a lot of that stuff, but, um, and it's easy to,

212
00:25:26,540 --> 00:25:31,900
you know, go point at the documentation, but it's also, you know, fun to walk somebody through it

213
00:25:31,900 --> 00:25:39,020
because people learn in different ways. Yeah. That's a, that's an interesting kind of nugget there

214
00:25:39,020 --> 00:25:46,220
about, Hey, it's really this kind of constant process of educating and, you know, engaging with

215
00:25:46,220 --> 00:25:52,620
all these teams, keeping that beginner's mindset, right. And remembering that, you know, even though

216
00:25:52,620 --> 00:25:57,820
this other team has been operating on the platform for a while, so they get it, they know what they're

217
00:25:57,820 --> 00:26:03,500
supposed to do with say their session IDs. Um, this new team that's coming on isn't going to

218
00:26:03,500 --> 00:26:09,180
necessarily know all that. So how do you go about that? I'm curious also, like, you know, how do you,

219
00:26:09,180 --> 00:26:16,540
how do you kind of, uh, navigate egos in that scenario to like, not show up with, let's sort of walk you

220
00:26:16,540 --> 00:26:20,380
through everything in case like, they're like, well, we already know that you don't have to talk to us that

221
00:26:20,380 --> 00:26:29,180
way versus also not completely like showing up with nothing and letting them fail. Right. And then

222
00:26:29,180 --> 00:26:34,140
saying, Oh, it looks like you need to understand this principle in order to run on the platform.

223
00:26:34,140 --> 00:26:40,540
I don't know how you, I don't know that we consciously do it. It's just, there are some

224
00:26:40,540 --> 00:26:49,660
expectations that, um, new developers are doing pair programming with more experienced developers on the

225
00:26:49,660 --> 00:26:56,060
platform because it is different. Unfortunately, that always isn't the case.

226
00:26:57,260 --> 00:27:05,180
And, um, so we, you know, you have to, you, you, you have to gauge based on how they're asking the

227
00:27:05,180 --> 00:27:15,900
question, how informed they are and where, where to take it. Um, uh, whether to, you know, answer a

228
00:27:15,900 --> 00:27:23,420
question with a link to the center of excellence documentation or whether to, um, say, Hey, let's,

229
00:27:23,420 --> 00:27:27,580
let's, you know, can you share your screen? Let's go off on a team's chat and, you know,

230
00:27:28,300 --> 00:27:35,980
and let me see what you're, what you're seeing and, and, and that kind of thing. Um, it's, um,

231
00:27:36,620 --> 00:27:45,180
time is not, you know, time still our most limited resource, but we've realized over the past three

232
00:27:45,180 --> 00:27:55,020
years that time spent quality time spent with developers is probably more important than jumping

233
00:27:55,020 --> 00:28:01,660
from thing to thing really quickly as you do in operations often, you know, um, to be able to,

234
00:28:02,380 --> 00:28:09,420
um, because, because they'll, they'll get more out of it and they'll be, uh, you know, you're,

235
00:28:09,420 --> 00:28:13,620
you're putting time in now to save time much later on, I guess.

236
00:28:13,620 --> 00:28:22,260
Yeah. Okay. That's, that's another great nugget. I think, um, that, that sort of resonates with some

237
00:28:22,260 --> 00:28:27,940
other folks that we've talked to in terms of the balance of your time spent and spending more of

238
00:28:27,940 --> 00:28:32,260
that with developers. So, um, thank you.

239
00:28:32,260 --> 00:28:41,620
Right. The good thing about the platform is knock on wood is that it's available. It's not going to fall

240
00:28:41,620 --> 00:28:52,020
over. It's not brittle or fragile. So, uh, back to the, the core, the core function that we do on a

241
00:28:52,020 --> 00:28:57,460
day to day is delivering new capabilities. If, if we have to spend some more time with,

242
00:28:57,460 --> 00:29:05,300
with, with developers delivering product on the existing capabilities that those other capabilities

243
00:29:05,300 --> 00:29:12,260
can usually slip, you know, a week or so, because it's just an understanding that we're going to

244
00:29:12,260 --> 00:29:17,460
constant, we we've demonstrated that we'll, we'll be constantly delivering capabilities. And so we're

245
00:29:17,460 --> 00:29:24,420
not tied to arbitrary hard dates. Yeah. Because, because, you know, we're delivering more and more

246
00:29:24,420 --> 00:29:32,180
apps at a much rapid, much more rapid pace than has ever been achieved at the company. I mean, we've,

247
00:29:32,180 --> 00:29:38,820
we're, we're, we're, it's not huge. We're up to 300, you know, different apps, not app instances,

248
00:29:38,820 --> 00:29:44,500
but apps. And that's, that's in three years. I mean, that's, there's nothing, nothing that we've ever

249
00:29:44,500 --> 00:29:53,860
done that's had that kind of scale. So there's the evolving culture is accepting that.

250
00:29:54,420 --> 00:30:02,260
Yeah. Well, and, and to your point before, where, instead of being focused on those delivering those

251
00:30:02,260 --> 00:30:08,420
objects, which is kind of like, it's a, I'm going to mess up saying, assembly line, well, it's the

252
00:30:08,420 --> 00:30:13,620
Sisyphean task. Oh yes. It never goes away. So you roll that bolt up the hill and you're just going

253
00:30:13,620 --> 00:30:16,980
to have to do it again tomorrow. That's right. I got another OS upgrade. They're going to need

254
00:30:16,980 --> 00:30:23,140
another VM in, you know, X number of years. Yeah. So if you've automated that, then suddenly it's like,

255
00:30:23,140 --> 00:30:29,460
oh, I have more time. Right. And, you know, getting started is probably where the most work

256
00:30:29,460 --> 00:30:35,220
is. And you, once you've covered, like kind of using just 80, 20 rules, once you've covered 80% of

257
00:30:36,020 --> 00:30:42,580
these are the objects that we've had to, we've had to deliver again and again and again. Now that's

258
00:30:42,580 --> 00:30:49,220
automated. So we still have like the ever evolving remaining 20%, which is kind of your corner cases.

259
00:30:49,700 --> 00:30:55,700
But that also means that we have more time to invest with these developers to help them

260
00:30:57,140 --> 00:31:02,180
do, you know, be able to deploy onto the platform and be successful. And that is actually a better

261
00:31:02,740 --> 00:31:08,580
return on our time than making sure that we're constantly like getting more and more of those

262
00:31:08,580 --> 00:31:15,700
corner cases delivered, which to your point is like, you're, it's not that you're past the point of

263
00:31:15,700 --> 00:31:21,860
diminishing returns, but your, your overall return is just, okay, you've shifted the balance. And so much of

264
00:31:21,860 --> 00:31:31,220
the, the base object delivery and firefighting has now been kind of automated away. So without a doubt,

265
00:31:31,220 --> 00:31:37,460
you said that very eloquently. So I want to shift to kind of a couple more questions.

266
00:31:37,460 --> 00:31:46,420
Um, and you know, thinking about what are some of the things from your days on the infrastructure side,

267
00:31:47,220 --> 00:31:51,940
knowledge, skills, relationships that you've kind of brought with you and that,

268
00:31:52,500 --> 00:31:58,340
that continue to serve you well, right? Like those are, uh, yeah, the, the, as Marie Kondo would say,

269
00:31:58,340 --> 00:32:03,300
the things you want to bring into your future. Yeah. Um, what are, what are some of those things?

270
00:32:03,300 --> 00:32:12,260
Um, so, well, they're, they're actually, I don't know that there's anything that I, that I have done in

271
00:32:12,260 --> 00:32:18,260
the past that isn't useful to me in the future, in the present and on into the future. Um,

272
00:32:19,540 --> 00:32:27,780
as I, I said, in our, uh, previous conversation interview about networking, a firm handle on how

273
00:32:27,780 --> 00:32:36,740
networking works that not from a, you know, um, router, you know, switch perspective, but from a

274
00:32:36,740 --> 00:32:43,860
understanding how DNS works, understanding how HTTP works, understanding how, you know, connections

275
00:32:43,860 --> 00:32:53,380
work and, you know, where firewalls can block you. Um, and what that looks like from a client side or from a server side.

276
00:32:53,380 --> 00:33:05,300
Um, that's, that's something that, um, I, I consider very, very foundational, um, the virtualization

277
00:33:05,300 --> 00:33:16,420
aspect of it, um, probably not nearly as important whatsoever. Um, I mean, we, we, we, we never touch

278
00:33:16,420 --> 00:33:26,500
VMS, uh, which is, you know, pretty amazing, um, to tell you the truth because we, we have like

279
00:33:26,500 --> 00:33:34,260
a third of all the VMs in the environment, in our environment, or maybe a quarter. And yet we never,

280
00:33:34,260 --> 00:33:40,020
ever have to deal with them, uh, which is, you know, a thing of beauty. It says a lot for what,

281
00:33:40,020 --> 00:33:50,020
what, what Bosch does, um, understanding, you know, uh, web servers, you know, that's something that, uh,

282
00:33:50,020 --> 00:33:59,220
and, and monitoring for that, for, from that, uh, respect, um, what, what are important things to monitor?

283
00:33:59,220 --> 00:34:07,860
What are key performance indicators? What are the, um, what are the things you need to poll versus what

284
00:34:07,860 --> 00:34:19,460
are the things you need to, um, just watch for events? Um, those, uh, have really helped in, um,

285
00:34:19,460 --> 00:34:29,860
in the first year of the platform to, to, um, on both the app side and the platform side to really

286
00:34:30,660 --> 00:34:39,940
tone down the, the, any kind of external heavy lifting that has to be done to monitor the platform.

287
00:34:39,940 --> 00:34:46,900
No, the platform provides metrics. Let's just look at those and let, and only, only put in extra

288
00:34:46,900 --> 00:34:52,500
monitoring where, or write a script to check things out where it's really, really required, like where we

289
00:34:52,500 --> 00:35:00,420
can't get this information. Um, so the, the, the fact that, you know, built the monitoring platform

290
00:35:00,420 --> 00:35:11,620
back in, you know, 2000 and, uh, starting in 2001, that really helped, um, educate, uh, the other

291
00:35:11,620 --> 00:35:17,780
platform teams and the, and the developers. Like, how do you guys know that my app is up? Oh, well,

292
00:35:17,780 --> 00:35:25,700
we, we get events when it crashes, you know, there's logs that we stream and watch for events. So that's

293
00:35:25,700 --> 00:35:30,340
been very important because in the past we had to ping those things. You know, you had

294
00:35:30,340 --> 00:35:35,860
some synthetic transaction always constantly running against your app to make sure that it was up.

295
00:35:37,300 --> 00:35:43,700
Okay. Now, um, on the flip side, what are some of the new things that you've had to learn

296
00:35:44,580 --> 00:35:52,020
in your, your platform role, um, that, you know, you've kind of just added to your, your arsenal of

297
00:35:52,020 --> 00:36:02,740
skills and knowledge? Um, so, um, Bosch, uh, you know, Bosch is huge and we use it both in the platform and

298
00:36:02,740 --> 00:36:11,300
outside of the platform. Um, um, the way that containers are secured, the, the way that containers

299
00:36:11,300 --> 00:36:18,100
get, get out. So there's that networking component that I talked about as, as, um,

300
00:36:18,100 --> 00:36:26,340
needing to understand or providing value to understand, but how, how the containers actually leverage,

301
00:36:27,140 --> 00:36:38,580
um, that in Envoy and the go routers is, uh, you know, when, when you, when you're troubleshooting

302
00:36:38,580 --> 00:36:45,940
something, you need to know how that stuff works, where, where layers are, where the ASGs sit in the

303
00:36:45,940 --> 00:36:55,860
environment. And sometimes you forget those things and then you're rudely, rudely, um, uh, awakened

304
00:36:57,140 --> 00:37:02,500
by, uh, the fact that, oh yeah, we have ASGs and we did, that is a new connection and we didn't

305
00:37:02,500 --> 00:37:07,860
account for that. And ASGs don't log when they fail and that kind of thing. But once again, you're

306
00:37:07,860 --> 00:37:20,740
finding that in dev versus in prod, which is good. Yeah. Um, learning a lot about, um,

307
00:37:21,380 --> 00:37:29,620
how modern applications are structured and built, um, you know, from the developer standpoint, um,

308
00:37:31,620 --> 00:37:41,220
you know, it's not, we work in scripting a lot in operations and infrastructure and it's top down.

309
00:37:41,220 --> 00:37:47,140
I have, you know, pre-described input inputs and outputs and modern apps aren't built like that

310
00:37:47,140 --> 00:37:52,020
really a lot of times at all. They're, you know, they're event driven and they're

311
00:37:53,220 --> 00:37:58,820
controllers. So learning about spring, um, not that I could write any of it,

312
00:37:58,820 --> 00:38:06,500
but it's very interesting. Learning Python is, uh, you know, that's, that's something that's very

313
00:38:06,500 --> 00:38:16,500
useful. I hadn't used that before. So. Okay. Any, uh, how are some of the ways that, that you've

314
00:38:16,500 --> 00:38:24,420
learned this or are there resources that you found valuable, um, you know, experiments or experiences

315
00:38:24,420 --> 00:38:31,540
that really pushed up your knowledge? Well, I mean, there are so many examples in the community of

316
00:38:32,100 --> 00:38:39,860
something that you need to accomplish. Um, I mean, I haven't, would none of us, none of us have gone to

317
00:38:39,860 --> 00:38:48,900
training for, you know, the platform. Um, the documentation for the most part is extremely good. Um,

318
00:38:49,700 --> 00:38:57,380
you have to read it. Um, and you have to grok it and we have a sandbox environment and you have to,

319
00:38:57,380 --> 00:39:06,020
you know, you have to feel it viscerally. You have to experience how it's done. Um, so, um, once again,

320
00:39:06,020 --> 00:39:10,900
that's, that's one of the benefits of the platform is it allows you to fail fast. It allows you to try

321
00:39:10,900 --> 00:39:15,780
things and see how it works and you're not breaking anything and you're not causing anybody else any

322
00:39:15,780 --> 00:39:22,820
work in cleaning up after you, which is, you know, what is traditional in the old environment, uh,

323
00:39:22,820 --> 00:39:28,340
the, the, um, the non-platform environment, I shouldn't say the old environment. Um,

324
00:39:29,460 --> 00:39:38,980
it's so, um, it's, it's really self-discovery, you know, doing, doing the work yourself, finding,

325
00:39:38,980 --> 00:39:50,100
you know, being pointed at some resources and then giving yourself time to, to, to experience it

326
00:39:50,100 --> 00:39:58,500
viscerally, you know, you know, build it, run it, and then take it from there and giving, uh, you know,

327
00:39:58,500 --> 00:40:04,260
having the, having the organization allow you that time to learn how to do that stuff.

328
00:40:04,260 --> 00:40:10,180
Is there anything, if you were to go back and do it all again, is there anything you would do

329
00:40:10,180 --> 00:40:13,540
differently to make this transition?

330
00:40:13,540 --> 00:40:19,780
I think so. We started, we started conversations with

331
00:40:20,900 --> 00:40:27,620
Pivotal probably in maybe 2014, I'm thinking.

332
00:40:28,980 --> 00:40:37,300
And the, the reason we didn't do some adoption then, we always walked away saying that's really

333
00:40:37,300 --> 00:40:45,140
great. We don't have any use cases. And that was short sighted. Nobody wanted to be the guinea pig

334
00:40:45,140 --> 00:40:51,860
that would jump into, you know, that would say, I will make the financial, you know, the financial

335
00:40:52,100 --> 00:40:59,060
commitment to make, to, to try this thing out. No, no, no one of our organizations, no one of our

336
00:40:59,060 --> 00:41:05,860
business units, no one of our, uh, application groups. And I think that was, that was short sighted.

337
00:41:05,860 --> 00:41:17,300
I think, um, being an early adopter helps and providing employees with some non-project time

338
00:41:17,300 --> 00:41:25,860
to explore those things is, um, is very much more important. And we're doing that now as an

339
00:41:25,860 --> 00:41:32,420
organization, we see the value in exploratory work that may, that may go nowhere, but at least you

340
00:41:32,420 --> 00:41:38,820
learn something and you learned whether or not it was going to be useful. So that was, that was,

341
00:41:38,820 --> 00:41:46,100
I would probably adopt sooner. Okay. Long-winded answer to a very short question. A very short

342
00:41:46,100 --> 00:41:52,660
answer. We needed all the color and texture. Okay. Okay. And then kind of, you know, I guess a related

343
00:41:52,660 --> 00:42:00,820
question is what advice would you have for folks who are, you know, currently in more infrastructure

344
00:42:00,820 --> 00:42:09,460
centric roles, maybe they're a VI admin or vSphere pro, um, or, or maybe they're more on the networking

345
00:42:09,460 --> 00:42:16,260
side. Networking, hosting. They all have, I mean, you're doing it, all of it. What, what kind of advice

346
00:42:16,260 --> 00:42:19,860
would you give to them if they're kind of, you know, they're interested in all, you know, they,

347
00:42:19,860 --> 00:42:24,500
they, they hear the DevOps buzz, they, they're curious about it. What, what should they be doing?

348
00:42:24,500 --> 00:42:35,540
There, there is so much free training available out there, um, from VMware and the hands-on labs,

349
00:42:35,540 --> 00:42:49,460
um, from the Kubernetes, uh, uh, uh, environment. Um, there, there is no, uh, lack of, of information

350
00:42:49,460 --> 00:42:57,380
to improve your skills and to understand how, how the new world, the new internet is working. Um,

351
00:42:57,380 --> 00:43:04,740
so I would, I would be, uh, telling people to, to, I would, I would be recommending, not telling people

352
00:43:04,740 --> 00:43:13,060
to, you know, ask your management, Hey, can I have four hours a week, you know, to study this thing,

353
00:43:13,060 --> 00:43:22,020
um, um, or do it on, you know, in your free time, because, um, that's, that's how hands-on is how you're

354
00:43:22,020 --> 00:43:28,180
going to learn this stuff. You can learn a lot in books, but you, until you've, you know, press the

355
00:43:28,180 --> 00:43:37,380
button and deployed something, um, it's, it's eye-opening the power that, that you can grant, you can give to

356
00:43:37,380 --> 00:43:42,340
your customers and eventually, you know, the, the capabilities that you provide to whatever

357
00:43:42,340 --> 00:43:52,820
businesses you're in. Yeah. Okay. That's, uh, that is a great starting point. Um, and so with that, uh,

358
00:43:52,820 --> 00:43:58,740
I, I just want to thank you for your time and insight. It's been really interesting to, to kind of hear your,

359
00:43:59,300 --> 00:44:06,340
your journey. And I hope this is valuable for lots of folks out there who are dipping their toes into the,

360
00:44:06,340 --> 00:44:11,860
to the platform side and what they're, they want to do next and in their career and also what they're

361
00:44:11,860 --> 00:44:25,780
delivering to their businesses. So thank you. Thank you.