00:00 [music]
00:04 >> All right, Will. It's good to see you. I'm excited to catch up on the network
00:08 episode 4. We uh we took a a little break, but it's nice
00:12 to be back in the studio. I know you've been traveling all over. I've been glued
00:17 to a computer, so let's uh let's catch up a little bit. I think you you've had
00:21 back-to-back travel weeks, right? Paris and then mercifully just a drive into
00:25 Austin, but uh what did you did you learn? Yeah, well, um so I was in Paris
00:30 with Cisco and um I spent time with uh the customer experience team.
00:35 And for those of you that don't know, Liz Santoni took over this group about a
00:39 year and a half ago. And there there were some challenges. Her predecessor
00:44 had focused on CX, but it was overly complicated. And, you know, the
00:48 objective with customer experience is to uh improve
00:53 uh the overall, you know, experience in dealing with Cisco from from a licensing
00:58 perspective and then also from a partner enablement perspective. So, um Cisco
01:02 recently launched Cisco IQ. And this is an agentic-oriented framework that um
01:07 that intends to address both of those things that I just mentioned. A lot of
01:10 the information that the team shared with me was under non-disclosure, and it
01:14 was a group of 10 or 12 industry analysts. And we had met a year ago when
01:19 the team put the stake in the ground. And what I can say is that it's amazing
01:22 to see what has transpired over the last year. Um they have several hundred
01:27 developers that are spread out all around the world that have been uh
01:30 working on on Cisco IQ. And more will be announced at Cisco Live, and I know you
01:35 and I will be attending that event together.
01:38 That was great. And then right after that,
01:41 I drove a short 30 miles from my home in Bastrop to Austin. I'm still here, and I
01:46 spent time with uh with Palo Alto Networks. And the big news this week was
01:51 their close their official close and integration of CyberArk and a new brand
01:56 uh Era. Um and that that new brand is intended
02:01 to nest within Cortex and some of the other brands as well. And I had, you
02:07 know, an opportunity to spend time with the executive leadership team and in the
02:10 cash and they're pretty bullish. Identity is very foundational, Shawn, as
02:14 you know, to securing agentic workloads and and applications. And observability
02:20 is as well. That's another observation that I made after attending RSA
02:23 Conference. So, yeah. So, um I've been quite busy and then I'm heading to rainy
02:28 Sydney with my wife on Friday for a vacation. I think we're going to get one
02:31 day of sun at the end of next week, but I don't care. I've never been. I'm super
02:36 excited about being down under. But what about you? What have you been up to?
02:39 Hey, you're going to get Concierge key yet, Will? I'm trying hard, buddy. Well,
02:44 let's I mean, we can talk a little more about Cisco, a little more about
02:47 security, and a little bit more about agentic AI. I think all excellent
02:52 topics. Um I've been working on a bunch of different research projects around
02:57 quantum-safe networking, specifically post-quantum cryptography and then
03:02 quantum key distribution. So, I would say, to give a little bit of context,
03:06 that it's important to note one, that recently Google made an announcement
03:12 that they were moving up their migration to post-quantum cryptography. It was
03:16 initially slated for 2031. They pulled it forward to 2029.
03:22 IBM Think was recently and they, you know, IBM is one of the leaders in all
03:27 things quantum. For sure, yeah. But uh their CEO
03:31 made the comment during his keynote that he thinks quantum advantage could
03:36 potentially be achieved this year. So, I I think
03:39 the point here is that maybe quantum computing in a non-science experiment
03:45 context is is maybe closer than we think it is. And I think there's a lot of
03:49 signals we're seeing from the market that that would with that.
03:53 You know, I referenced Cisco. They have a a new universal quantum switch that
03:59 they announced. That's right. To be clear, this is not a commercial product,
04:03 but this is, I think, showing you where the, you know, where the puck's going a
04:08 little bit. They've got a a research paper that I can maybe cite in the show
04:13 notes once this is published, but it kind of looks at what a future quantum
04:17 data center might And so, you would have different types
04:20 of quantum computers operating with different modalities, and they would all
04:24 be connected in concert through, among other things, a quantum switch like
04:28 this. Mhm. You referenced Cisco Live. I expect
04:32 we'll probably hear some more, but I mean, what's what's your kind of what's
04:36 your read on this? Do you I think quantum is one of those things for a
04:39 long time it's felt like very important, but very distant. Yeah, yeah.
04:43 >> Now it feels a lot more tangible. Yeah, no, I agree with you, Shawn. And, you
04:47 know, by by all accounts and all measures,
04:51 um that that, you know, that 2029, you know, is is is likely because
04:57 recently there have been several infrastructure providers, including IBM,
05:01 that have attested to the fact that, you know, it could be here by 2029. I I
05:05 mentioned Palo Alto Networks the time spent this week with the team, but I've
05:08 also spent time with their PQC team as well.
05:11 You know, and they've done a lot of surveys, they've done a lot of research,
05:14 and have determined that, you know, most enterprises are not ready for this. And
05:18 the concern is around harvest now, decrypt later. And so, bad actors can
05:23 steal that information, keep it, and then once we reach, you know, this this
05:27 this central moment, they can basically defeat all encryption.
05:32 And then that becomes interesting. There arguments that, hey, you know, will will
05:36 will data stolen three or four years from now be useful? It really depends,
05:40 you know, if it's tied to critical infrastructure and that sort of thing.
05:42 What what I love about what Cisco is doing with its proof of concept with
05:45 this universal quantum switch is that it's it's birthed out of their outshift
05:49 organization and if if our viewers aren't familiar with outshift, this is
05:53 Cisco's incubation engine and they've done a lot of great practical research
05:57 that's actually been launched. Um agency was was an attentive framework and
06:02 security architecture that um that came out of outshift. Uh it was moved to the
06:07 Linux Foundation last year. I wrote a Forbes article on it. You wrote a great
06:10 research note on, you know, capturing the essence of this switch. What I think
06:14 is really compelling about where Cisco's headed with this is that it uses
06:17 standard fiber, so standard interconnect number one, and it operates at room
06:21 temperature. And so those are huge considerations, right? So yeah, I agree
06:25 with you. It's it's going to be here, you know, sooner than than expected from
06:29 from my perspective. And organizations really need to figure out number one,
06:32 what is their exposure, right? So they need to do an inventory of their
06:35 cryptographic assets, number one. Number two, they need to put measures in place
06:39 to to safeguard it. And, you know, this week you and I caught the announcement
06:44 from from AT&T working with Cisco on a quantum resilient SD-WAN service. And
06:50 AT&T is the first service provider in North America to launch something like
06:54 this. And I I believe we'll see other offerings, you know, you know, you know,
06:58 come down the pipe because again, there's a lot of attention being paid to
07:02 PQC and and the need to be prepared sooner versus later. That's right. I was
07:07 I was really interested to see this from AT&T and and I'll I'll provide a broader
07:12 frame of that is if I kind of scan operators that are engaged in PQC
07:18 quantum key distribution short list. Singtel has done some interesting stuff
07:22 with a hybrid architecture for that. I think one thing to keep in mind is
07:26 Singtel operating that in in Singapore, the space constraint plays to your
07:31 advantage when you're talking quantum. Yeah.
07:35 BT similarly has done some interesting work with a quantum safe metro a
07:40 around central London in partnership with some of their financial customers.
07:45 So, obviously high-value data that that needs to be protected against emerging
07:50 threats. But, AT&T, this is being sold through AT&T Business. It's like an
07:55 augmentation of their SD-WAN offering or their secure service edge offering. And
08:00 as you said, Cisco 8000 series secure router. But, this this is something that
08:05 customers can get purchased later this year. So, again, this is not a science
08:09 project. And I Right. So, this interestingly plays to Cisco's larger
08:14 posture towards quantum networking. Obviously,
08:18 a useful quantum system to do things like solve for drug discovery or weather
08:24 modeling, it needs to produce a sufficient number of qubits.
08:28 Yeah. Uh a lot of the leading systems right now don't really do that. So, you
08:33 kind of >> fragile, right? Yeah.
08:35 >> Right. And you get into a situation where it's like, well, you know, we need
08:38 to pursue two road maps at once. We need to make these individual computers much
08:43 larger, but at the same time we also need to think about just the force
08:46 multiplier you get by networking them together. Cuz that's what we're seeing
08:50 right now with AI is like, how can I get all of these disparate bits of
08:53 infrastructure organized and orchestrated to act like one big
08:57 computer in service of X. Same logic's going to apply to quantum. So, I I feel
09:02 like Cisco's really way ahead here. You know, and and you know, and I I
09:06 spend a lot of time with Cisco. And I'm also impressed with what they're doing
09:11 with their custom silicon. I mean, beyond silicon one, the Alif team has
09:15 developed you know, quantum entanglement silicon.
09:18 It's hard like you know, I profess to know about like probably 10 or 15%.
09:23 Right? You know, if anyone tells you that they're an expert in quantum,
09:27 they're they're lying because it's it's changing. It's entirely complex. The
09:31 entanglement process, it's it's almost like teleportation. And it's hard for me
09:36 to articulate and explain, you know, just in a few minutes here. And you
09:39 know, we might even want to dedicate a podcast to, you know, going really deep
09:43 into quantum and nerding out on this, man, cuz like there's a there's a lot
09:47 there. And when you combine the potential of quantum with with a gentle
09:52 AI, I mean, talk about a force multiplier potential.
09:55 >> I mean, we talked so much about the parallel processing you get from GPUs
10:00 with quantum. It's not one or zero, it's one and zero. So, it's an exponential
10:04 increase in in the ability to to do computation. So, I mean, super exciting.
10:08 Maybe we'll pencil that in. I'll I'll collect my thoughts on superposition and
10:12 coherence. We'll get that going. Let's talk a little bit about
10:16 a gentle AI. I think by the time people are watching
10:20 this, I'll have a a report published on RCR Tech that really looks into scaling
10:26 AI ops. And I think kind of the thing that emerged in the research that I did
10:30 for that report, the conversations that I had with operators and with vendors,
10:34 notably Red Hat gave me a lot of good there, is that operators have gotten
10:40 really good at data analytics. They've gotten really good at drawing insight
10:44 from data that has historically not been put to highest and best use. So, now
10:50 with AI ops, it's scaling insight in an operational way. It's like, okay, well,
10:55 I know this based off of the data. So, now I need to do this. And with, you
11:00 know, an agentic system done appropriately, you can do that in
11:05 closed-loop fashion. I think where we are today is still human in the loop.
11:09 That has to do with trust and anything else. And I think um you know, Rakuten
11:14 gave me some some interesting uh commentary for the report where it was
11:19 this idea of how do you quantify trust? And
11:24 uh I don't have the the numbers to hand,
11:27 but it was essentially if I set up an agentic system and I keep a human in the
11:32 loop for 3 months and I achieve 90% plus outcome that I I wanted that was based
11:39 off of my intent, then that's where you go. You can take your hands off the
11:43 wheel. Yeah. But I think with so many things, telecom, and you know, Rakuten,
11:49 I wanted to work with them on this report because I think they have a great
11:52 point of view on it given that it's a new network, so there's not a, you know,
11:55 50 years of of technical debt and legacy.
11:58 >> Yeah, yeah. Green field is a lot easier, right, than brown field. It's not
12:02 exactly a one-to-one analog for every operator, but I do think it is
12:06 instructive. But I, you know, I guess the the big takeaway for me in that
12:10 whole thing was, okay, we're going to do intelligent AI. So, that is a technology
12:15 stack, but it's also an operating model. So, I I would tend to think of it as an
12:20 execution stack where AI is an active participant in decision-making, which,
12:26 you know, you can buy all the technology in the world, but if you don't really
12:29 kind shift the organizational mindset, you're not going to achieve what you
12:33 wanted to. Right. I mean, you've got to have AI change management, right? You
12:37 know, in the mix, and I've recently spent a lot of time with with NTT. I may
12:42 have mentioned this on a prior podcast with you, but I was at their their
12:46 upgrade event, and they talked about AI change management and how important that
12:50 is because you can operationalize this all day long, but if you don't have
12:54 buy-in, if you don't have people that are actually willing to step up and
12:57 experiment and be confident. And honestly, like with any technology, you
13:01 know, sea change like we're experiencing with AI, you're going to have folks that
13:04 are resistant to it, and you're going to have fallout. But I think it's really
13:07 important. I think it's interesting what what Rakuten has been doing here. You
13:11 know, with Symphony, you know, they they've been sort of, you know, moving
13:14 sort of outside of their uh uh chops, and you know, they they've had
13:18 some hits, they've had some misses, but but I really like what they're doing.
13:21 And this isn't exactly a parallel, but I think it's relevant to mention. You
13:25 know, I've I've also spent a lot of time with HPE and Juniper over the years and
13:29 and there was a big announcement just just a few days ago around um
13:33 self-driving networks. And this is something that that Juniper began its
13:37 mission about 10 years ago. And then now as part of HPE, you know, they have the
13:42 engineering velocity and they have the scale to take it to market. And it's
13:46 really interesting what they're doing and it and it addresses on the
13:49 enterprise side of things what what you mentioned on the public mobile network,
13:54 you know, side of things. Um it's going to drive improved assurance. It's going
13:57 to drive eventually automated remediation. I mean, and that's the
14:01 whole, you know, value from my perspective with the Gentex frameworks
14:04 is it's driving automation at scale, right? And so I I love what what HPE is
14:10 doing on the the enterprise side. I love what Rakuten is doing on the service
14:13 provider side. And this stuff's going to come together as well. I mean, HPE
14:17 acquired Affirmed Net, so they have a private 5G. And so it's going to be
14:20 really interesting to see how this, you know, from my perspective all nets out.
14:23 But but but the good thing is it's going to hopefully accelerate, you know, the
14:27 adoption of AI and, you know, and the transformation that that potentially is
14:31 going to occur is is phenomenal from my perspective.
14:34 >> Yeah, and I mean, anyone watching this has heard this over and over again, but
14:38 it is such early days with AI. And I mean, I'm not making any a comment on
14:43 the financial markets here, but I mean, it's early days. Enterprise adoption is
14:48 is double digits, low double digits. And so there's there's so much work to be
14:52 done. And it's interesting, you know, I without giving away too much
14:56 information, my wife's position is with a large
14:59 company and she does a lot of data architecture, data management, system
15:04 integration. So AI is very much central to her remit and it's interesting to see
15:09 first hand this problem that we all know exists kind of play out, which is why do
15:14 you how do you pick the right experiment to run? How do you how do you pick
15:18 something that's probably going to work then scale it out across a bunch of
15:21 complex data system software systems. It's a tough nut to crack, but the
15:27 progress is astounding. To to that to that latter point, Shawn, as well, Dell
15:32 has this whole notion of customer zero, right? And so, they're very focused on
15:36 their AI factory and all the infrastructure from a compute and
15:39 storage networking to a lesser degree perspective. And what what they've done
15:44 is they've they've they've looked at, you know, practical applications within
15:49 their organization. Think about supply chain, think about customer support,
15:53 think about sales. And and they've they've kind of, you know, aggregated
15:57 all of those learnings and they're sharing them with their customers,
16:00 right? Um so that, you know, and that's the biggest challenge, to your point, is
16:05 sifting through what's really going to move the needle and what's not. So, I
16:08 like that. I I like that approach from from Dell Technologies and um it's a
16:11 great blueprint, you know, for for other infrastructure providers to to follow
16:16 because it just creates stickiness as a trusted advisor at the end of the day.
16:20 100% and I I've heard Dell's COO, Jeff Clarke, talk at good length about how
16:25 Dell internally reoriented the organization to take advantage of AI.
16:29 And as you said, it's a green blueprint that any enterprise could follow, any
16:33 large enterprise. And um that's a good reminder. I leave Sunday for fabulous
16:37 Las Vegas to go to Dell Technologies World. Always an excellent show, so I'll
16:41 I'll I'll have lots of lots of new new information and new insights to share
16:45 with you after that. And then, you know, it'll just be a a nice few days at home
16:49 and then we'll see each other in Vegas for Cisco Live. Yeah, well, I'll miss
16:53 you at Dell Tech World. I'll be down under, as I mentioned earlier, but um
16:58 have fun. Enjoy Vegas to the extent that you can.
17:02 I mean, you and I live there in the month of June, so
17:04 I think I think this year it'll be just 3 weeks, not not all 4 weeks, but uh but
17:10 I'm I am looking forward to Cisco Live. It'll be fun for you and I to hang out.
17:13 And I think you and I are going to try to record a podcast
17:16 at Cisco Live, right? Yeah, that's right. That's a a good reminder to me
17:21 that I need to bring the necessary equipment to do that. Please do because
17:25 I I don't have the necessary equipment. I draft in your greatness when it comes
17:28 to that, buddy. All right. Well, well, it's good to
17:32 catch up with you, man. Safe travels and always a pleasure to go beyond the
17:35 network. Yeah, you too, buddy.
No matching passages. Try another word.