The Evolving Role of Technical Evangelists in IT with Broadcom’s Jeremy Rossbach

Hey everyone. Welcome back here to techron tv. Our next guest is Jeremy Rossbach.

Jeremy is Chief Technology technical evangelist at Broadcom. Jeremy, welcome to Tech Trunk tv. Man.

It's good to see you, Alan. Thanks for having me back. Pleasure.

So, well, Jeremy, let's start about you. Right. You know, let's be honest, technical evangelists have have had a bit of a rough ride over the last two, three years, right?

Mm-hmm. There was a time there, Jeremy, where like the pendulum swung where companies were investing more in evangelist, more in people who talk the talk walk, the walk of the developer of the ops person of the IT person than the traditional bad carrying salesman who took you out for martinis. Right.

That, that was out of style. Mm-hmm. Over the last couple years, we've, I've seen more companies investing in salespeople mm-hmm.

At the, at the, at the, uh, loss of the evangelist crowd. Mm-hmm. What do you see?

You know, it's just a title. Uh, to be honest, when they, uh, wanted me to be more customer facing the face of the product, I just started Googling terms that would fit that. And to be honest, our main competitor Thousand Eyes was hiring tons of chief technical evangelists.

So, uh, that's where I got it. But to be honest, listen, in today's IT age, you want somebody that you can trust and Yep. Is not just blowing fluff, fluff at you, you know?

Um, I'm that person because I spent the first half of my career running data centers, being the guy that woke up at 3:00 AM because a router went down racking and stacking one new servers until my hands were bleeding, just because I had to have all that cable management looking perfect. Mm-hmm. No Spaghettis for you.

Yeah. So, you know, I I, I know what it's like to drop a router in an entire branch office in the middle of the night because I didn't want to go into the office and do the changes. I did it from home and screwed up.

So, yep. I get it. And, you know, and I believe I'm that that person.

'cause I can tell those war stories. I've been in the trenches and, you know, it pays off because like, if my company sends emails, say directly from Broadcom to our customers, getting them interested in something, the open rates are, are pretty normal for the industry, but they've been starting to send emails to our customers directly from me. And then when someone sees that has heard from me before or seen me talk somewhere online or, or at a conference, we get 50 more, 50% more open rates and clicks in those emails because they trust the sender.

And I just want to be that tr you know, maybe it should be advisor, trusted advisor. Yeah. Whatever.

It's, I I get it. Yeah. You know, look, the big thing when you talk to the PR agencies out there, and the marketing pros out there now is, you know, influencer marketing.

Mm-hmm. And, and what you just described actually is influenza marketing, right? Mm-hmm.

Mm-hmm. People say, okay, he, I, I know Jeremy, I, he's not a sales guy. Yeah, right.

We have a real strong anti-marketing sales Oh, of course. BIAS course and in it. Mm-hmm.

And, um, and influencer marketing is, is the way to go. Yeah. Jeremy, for young people out there who say, you know, that's what I'd like to be, that's what I'd like to do.

Mm-hmm. You talked a little bit about your journey, 15 years in data centers and so forth, but what advice would you give to people out there who maybe want to kind of follow in your footsteps? You know, um, the reason why I love to do it is two reasons.

One, it's very creative, and I am a very creative person. I was a music major in college. I drew, I've drawn all my life and made money at sketches and stuff.

So this gives me ability to be creative in a very lucrative field like it, um, where I can create videos, right. Uh, speak in front of an audience. And then the fact that, um, I feel like a trusted advisor.

I feel like people trusted me. It's, it's, it's big. You know, I wanna do an amazing job at what I do, and I want it to have an impact.

So if as an influencer, you know, um, I'm, you know, I'm not just trying to, uh, get people to click and buy something. I'm getting people to think differently about their IT strategies. Be aware down the road what's gonna happen.

And to be honest, consider the other fluff that's coming at you before you, you know, pay for something. You know, are you future proofing down the road with the solution or are you just putting a bandaid and duct tape on something? Now I want to be that person that's trusted, that helps someone, uh, down the road and now with an IT issue, and then see that impact along the way and then build a relationship.

Very cool. Yeah. Thanks for that.

Yeah. Alright. Let's, yeah, we kind of get your role at Broadcom.

Mm-hmm. But, you know, look, today, Broadcom, you know, they still king in the data center space mm-hmm. Whether it's with Broadcom, you know, labeled machine or Broadcom chips within other companies machines.

They also have a huge, you know, they've invested heavily in software over the less, let's say five, seven years acquiring some of the leading names out there, including VMware and ca and, and others. Where do you reside within this whole Broadcom? Yeah, we were, uh, one of the first big acquisitions ca technology.

So I'm former ca Technologies computer associates. Uh, Broadcom bought us in 2018, of course, scared the hell out of all of us. Um, and, uh, that's where I come from.

So I had been pre-sales, I was running data centers for ca at the time and before we started consolidating them, and then went into pre-sales as like that technical sales salesperson doing POCs and demos and, um, and, uh, been still there ever since I moved from server virtualization management software to process automation to now network, uh, observability. And, uh, it's an exciting time to be in network observability because network monitoring is old and three th 30, 40 years old and gets kind of crusty. But with the fact that the networks are so complex now with cloud and AI and SD this and SD that and on and off-prem, it's a really fun time to be in a, in, in this area where we can do a lot of help, you know?

Um, and, uh, you know, listen, real quick, I just wanna say, and I say this all the time, you know, there's always so much fun out there about acquisitions and m and a, I just want folks to remember, and it's a reason why I'm still here, is that even though they're always disruptive, um, there is decades and decades of people behind the software that are still here that are proud of that software that work every day to make it better for our customers. And even though there's a lot of news and a lot of complaints, I get it. Uh, but the, the, the people behind the software are still proud of what they do and what they're doing for our customers today.

You know, maybe I'm jaded 'cause I've been in tech so long, but m and a is a way of life in tech. Mm-hmm. Little companies get bought by medium-sized companies, medium-sized companies get bought by bigger companies.

And the fact of the matter is, it's, it's almost a, like an innovation farm system. Mm-hmm. That's how I look at it.

Mm-hmm. A lot of companies get to a certain size, look, they've got great pipes into customers. They, you know, they service 490 of the Fortune 500, but their innovation engine is not as dynamic as you get at these little companies out here in garages and accelerators and what have They move faster than the big ships like Broadcom.

Yeah. Not just Broadcom, all of them. Mm-hmm.

And so they look for innovation. Mm-hmm. They look for, and then it, it's not just usually, you know, the size of a company, Broadcom buying that smaller little innovation engine.

Mm-hmm. Generally there's a medium sized company in between that takes those innovations and builds them in more as a product or, or as a feature into a product. Mm-hmm.

And then a Broadcom takes those products into a platform. Mm-hmm. Right?

So it's innovation that becomes a feature into a product that becomes part of a platform. Mm-hmm. That's been my observation on this over the years.

And I've, I've also been involved in a lot of startups, venture startups. com, we actually, I helped a company, we eventually iPod, but we did 30 acquisitions in 36 months. Oh my Gosh.

Almost one a month. Wow. That's Crazy.

And yeah, you're right. Sometimes you break a few eggs making Omelets. Yeah.

Yeah. Yep, Yep, yep. But overall, you know, it worked.

Mm-hmm. Anyway, Jeremy, let's pivot. Yes.

Let's talk about the maturity model framework for successful AI ready networks. So, uh, thank, get Rid of the marketing gobbly Go and Exactly. And tell me what we're talking about.

Yes, Alan. Thank you so much. So listen, you know, um, a lot of the conferences I go to now, or that I speak at have an AI theme, uh, these networking conferences.

Um, and it's great, you know, uh, AI's a great buzzword. Everyone's launching onto it. But, uh, I just want people to slow down a bit, the network operations teams to slow down a bit, think about what they have to do to get ready as they're building these AI ready networks.

What is an AI ready network? It's either a network that is so resilient that can handle very large AI workloads, um, maybe near perfect networks. You know, AI workloads today, demand minimal latency, minimal packet loss, huge capacity near perfection, um, or obviously the results of your AI engine made different than what you expect.

And then there's the other ai, uh, ready networks that are using AI to help our network operations teams troubleshoot. So a, a chat bot or a knock bot or something like that, what I'm talking about today is really getting ready. Our networks resilient enough to handle any AI initiatives.

We give it, for example, Meta's AI workloads. We, we can only imagine are a size that we don't even probably have a name for yet. Terabyte or gigabyte, you know, XO Byte, their workloads spend 54% of their time on their network.

So if their network isn't resilient, if it's dropping packets here and there, you know, the game, they, they can't stand for that. AI workloads won't stand for that. So what I'm talking about here today, and I know it's a short segment, is if you have a resilient network observability practice that is feeding and analyzing data that's being collected, and then feeding those answers to your AI engine, imagine how fast, how accurate those results are gonna be when you come in in the morning and say, Hey, my knock bot, where should I focus my attention today?

And it knows that Seattle's having packet loss above 10%, and three tickets have come in from Seattle, and here's the IP and name of that router that's drop in packets. I mean, none of that happens without really good data collected. So this maturity model that I'm talking about is, uh, uh, developed by us and my team and backed by industry analysts.

And, you know, it goes from like one to five. So at one in our, our network observability ma mature maturity, we're all where we used to be, which is probably not where we are today. Very hyper reactive, tons and tons of little spot tools doing this part of the network and that part, part of the network.

Nothing talking, only monitoring traditional SNMP and it's swivel chair monitoring. And it's, it's where we used to be. Hopefully no one is, uh, nowadays.

And then you move to from hyper reactive to more reactive, um, where we have a hundred percent land coverage. We're looking at all the metrics that we need to look at, like alarms and faults and logs and packet loss and capacity and voice latency and jitter, and you name it, the tons of data that's on a network that's very valuable. We're collecting that now.

Um, are we doing anything with it yet? No, it's still a lot of noise in our dashboards or, or our consoles. Um, we're also starting to, uh, consolidate some tool sets as well.

And then in stage three, we talk about going from manual to traditional and then now to more of a modern approach, a more proactive approach. So all the data I just said that we're collecting, we're now doing something with it. We have a solution in place that has some analysis and some, uh, capabilities that can do predict predictive and thresholding and baselining.

And, and you get it, you know, 30, 60, 90 days out, this is what it's gonna look like. Um, we're also now expanding our monitoring and observability from traditional to more software defined. So we're looking inside of SD wan, uh, or orchestrators and, um, and other software defined platforms, SD lan, et cetera, like an a CI, uh, or VIT for the wan.

Um, so we can understand more about the overlay and the underlay, uh, correlation. So our virtual network running on our physical infrastructure, how does that correlate, you know, if there's a problem in that SD WAN n part of the network, is it the, the physical hardware affecting it, or is the virtualized, uh, weigh in on top of that is at the, um, affected root cause? And then finally, now that we're, you know, we have a solution that, um, is using analytics, is using proactive predictive behavior, is taking all the data and actually normalizing it and correlating it and surfacing it as one, uh, item to look at across multi-vendor technology.

Like focus here in this part of the network, not caring about if it's Cisco or if it's Juniper, et cetera. But correlating that information. We then move forward, uh, to a very predictive type of environment by now moving, um, from LAN and software defined inside the four walls to now moving to stage four, doing all that outside of the four walls of the data center.

So these unmanaged networks like ISP and Cloud, because that is the backbone of most of our enterprise networks today, is public networks. And none of us are gods in a public network. We're not the network admins there unless you're for with a Verizon or a Century Link.

So how do you gain that visibility? How do you do everything I just said to do inside the four walls of the data center where you are a God in these public networks that are basically delivering your customer's network experiences. So being able to gain visibility into those ISP and cloud provider networks, get enough information like packet loss, latency, jitter, et cetera.

What's my voice traffic look like? What's my data traffic look like to then not fix it because I said, you're not a network God, but to then find meantime to innocence. So to be able to spend minutes understanding by looking at the entire end-to-end network path of your customers and seeing that this packet loss is occurring in a device in CenturyLink or Comcast or whatever, uh, having packet loss with this IP address, I now know this is the root cause I go to that ISPI give them the information and then magically it gets fixed.

Um, again, this is the innocence for NetOps teams is, is available today within minutes instead of hours, uh, if not longer. And then finally stage five, again, moving towards getting our network's AI workload ready is automation. So now we all know automation has hasn't caught on.

Uh, for one, I wouldn't want a piece of software bouncing one of my routers, bouncing one of my switches. 'cause it thinks it should in the middle of the day. But, you know, can we do those mundane lowline tasks of escalating a help desk ticket, closing a ticket, enrichens enriching it with more data that has been collected from your network observability solution, uh, those type of rollback a configuration, you know, I've, I have many times fat fingered a config and drop the site.

Um, to be able to roll that back and get it up right away is automation. Um, these are the things that we can do, uh, to take care of those low hanging fruit use cases until we're all ready to really fully implement automation. And I know we'll get there.

I mean, we're already talking about ai. We still haven't even, uh, fully implemented automation. But, um, yeah.

So these stages of, you know, moving from hyperreactive all the way to automated network operations and network observability is what is the foundation that's needed for these AI complimentary solutions that are coming out today that are gonna do amazing things. I've seen, uh, at the last network automation summit I went to in, in November in, in Colorado, watching someone use their own chat bot to, to do exactly, you know, as, as we're asking it questions and, and it's telling us about the topology or, or an affected root cause area of the network that is not done with that AI complimentary solution that is based on the data you are collecting, you're feeding that data from your network observability solution. You're the one making it smart, you know, know garbage in, garbage out.

Like I always say, if I only have eight out of the 10 ingredients for my kids' buffalo chicken dip that she loves every time she comes from from college, it's not gonna taste the same if I had 10 outta the 10 ingredients. Right? So if we're feeding all the ingredients from a very robust and mature network observability solution, you can expect good AI outcomes.

Absolutely. Jeremy, I didn't wanna stop you when you were on a roll there, but you know, so much of it kind of hit home with me. Um, you know, the whole automation thing, I, back in 2001 I found co-founded a company.

We were trying to go from IDS to IPS. Simple, right? Mm-hmm.

You see an obvious code red worm. Mm-hmm. Or, or, you know, whatever was hot that day, block it.

Yeah. No, no, no. People were real hesitant.

Then we rolled out a vulnerability management to tool that had the ability not only to find vulnerabilities, but to put patches on. Mm-hmm. No, not, not on my watch buddy.

Mm-hmm. Mm-hmm. You know?

'cause if it does something wrong, you're the one to blame. Exactly. Yep.

Yep. But to me, the lesson I've learned through all of this, and I think it's it's still true today, and it's gonna to be true, especially with ai, is it's a, it's a question of trust. I, I need to do it, do some of these low level obvious things and let me build my trust.

Mm-hmm. Yep. Let me, let me get a good feeling about it.

I'm confident and then I'll let it do more. So it, that, that's just, I think that's just the way it is. I agree.

I agree. Because one router change can bring down the entire east coast of a network. I mean, It doesn't have to bring down the entire East coast if it brings down my CEO's laptop.

Very true. I'm in trouble. That's what I used to hear, what I used to hear back in the day.

But Jeremy, let me ask you, is there a place people can go look at and see maybe a diagram or maybe some, uh, you know, information on this maturity model framework? Yes. Um, so, uh, we'll put that link in the show notes, I'm sure.

But yeah, we have a, uh, great landing page about optimizing network operations, uh, with the help of AI has a lot of great white papers, including the white paper. I'm referencing that myself. Uh, and my team wrote, it's also backed by Seamus McGillicutty at EMA, uh, where he says that, uh, many customers should be measuring themselves against our maturity model to find success.

So very happy for that little, um, talk track there. But I wanna let a, a couple more plugs. I'm gonna be in Dallas next week, uh, for own the AI Networking Summit sponsored by og.

I'm speaking about this very same topic. I'm also MCing the event. So, uh, I'm very cool.

Yeah. Sitting there. I believe my friend Charlene O'Hanlon, who used to work here with us is, is, is uh, is moderating a panel there as well.

Wonderful. Wonderful. Um, yeah.

Well, I'm sure I'll be, uh, watching it 'cause I have a lot of free time. Not only am I MCing, but I'm speaking, but I'm gonna be talking to a lot of folks there as well. So if you're at the event, please look me up.

Also, uh, in two weeks, uh, Broadcom is hosting a, uh, network Observability Virtual Summit. Um, and I am interviewing, uh, a bunch of our customers, one from BT Ireland, and ISP Atos is a, uh, managed service provider. Um, and then Deutsche Bond is a German railroad company and we talked to them about this very same topic.

What does their maturity look like and how do they expect to get to AI ready networks? I'm gonna put that link for the registration page in this, uh, show notes as well for you. Alan, Appreciate it.

Jeremy. Hey man, it's been a pleasure having ya on. Come back and visit us again.

Will do. Keep us posted and it's always great to catch up. Great.

Thank you Alan. Jeremy Beck, chief Technical Evangelist, Broadcom here on text on tv. We're gonna take a break.

We'll be back in a bit.