MWC is one of the globe’s hottest events for tech industry announcements. In the latest of our MWC interview series, we caught up with Shekar Ayyar, Chairman and CEO of Arrcus, to discuss some of the company’s latest ground-breaking innovations.
Arrcus pioneers a networking paradigm shift, shifting from hardware-centric to a software-driven approach, offering unparalleled flexibility for deploying advanced workloads like AI, 5G, and multi-cloud environments. During the conversation, Shekar emphasizes Arrcus's strategic partnerships with industry leaders, including Broadcom, Intel, Nvidia, and VMware, enabling the company to leverage cutting-edge technologies while maintaining a unified, hardware-agnostic operating environment.
Want to hear more from Shekar? Tune in now!
Steve Saunders:
So how are you?
Shekar Ayyar:
Very good, thank you. Great to be here at the show again.
Steve Saunders:
Yeah, it's a big show and you've got some announcements what's going on?
Shekar Ayyar:
We do have some pretty exciting announcements. One which we are doing with Broadcom around support for their latest platforms in networking. And then we are talking about how we have helped SoftBank go into production trials with their 5G mobile user plane. And then we've also got announcements with Red Hat from IBM as well as VMware, which is now part of Broadcom.
Steve Saunders:
So you're really developing your ecosystem of partnerships, which I think that's a key part of your strategy, right? It is absolutely as big as any ecosystem of giant companies and obviously that's an impressive list. Tell me a little bit more about what you're doing with Broadcom. What do you bring to the table and what are they bringing?
Shekar Ayyar:
Sure, absolutely. I mean at a technical level it's extremely simple. Broadcom is effectively the hardware and processor company and we are the software and operating system company. So in order for any device to operate well, you need these two components in the infrastructure. Now, what is more interesting though is particularly with the advent of AI and then preceding AI, edge networking and multi-cloud and 5G, there is more and more focus on flexibility at that operating system level that enables different types of workloads on top, and different types of processor environments underneath. The analogy to this is what had been done with VMware where we had abstracted out the compute interface. What we are doing at Arcus is in fact abstracting out that network interface so that customers can get the best processing capability or the just right processing capability that they need, but with one common operating system environment.
Steve Saunders:
And so you can scale it up for AI, for example, or you can reduce it or scale it down geographically up to the edge of the network.
Shekar Ayyar:
Exactly.
Steve Saunders:
So one common environment managed continuously throughout your network operations.
Shekar Ayyar:
Exactly.
Steve Saunders:
And who's really interested in this? Is it enterprises? It sounds like it's something that a large enterprise might be interested in or a data center provider might be interested in, or is it primarily the CSPs and the carriers?
Shekar Ayyar:
I mean, it's really all of the above. So it's a fairly horizontal play. So we've got enterprise customers, we've got cloud and interconnect companies, and then we've got telecom communication service providers, etc. as the domain. Now interestingly, the use cases can be a bit different. I mean, an enterprise can come in and say either we have a data center, we need top of rack switches, spine/leaf, or we are moving to the cloud and we need something that actually allows it to be our network even as we are moving to the cloud…we service both use cases. Then when we go to cloud and interconnect companies, they're essentially saying, look, we are building these gigantic data centers for our customers, and then we need interconnection technology that connects between these data centers, takes them to the edges, et cetera. And then when we talk to communication service providers, they have a network that has historically been siloed and saying, look, cell site routers, this is the part number I get from large companies. I mean Cisco, Juniper, Ericson, Nokia, or I'm doing a core router, or I'm doing a PE router. And that has been their mode of operation. They're now saying, look, I'm now struggling with applications like AI that come and start testing whether in fact I'm flexible at that network level and how can I get there? And that's where we help them.
Steve Saunders:
So you are giving them the flexibility to onboard any service?
Shekar Ayyar:
Exactly.
Steve Saunders:
And AI has obviously come out of left field a little bit with the level of excitement about it, but it is real now and it does demand very high levels of processing, and you can accommodate that in a scalable way?
Shekar Ayyar:
Absolutely. So think about us as taking the best that the hardware can offer. So for example, the latest silicon announcement that we have been talking about at the show with Broadcom gives you capabilities that has historically not been provided for 800 Gig ethernet technology. As an example. We've also made announcements with Nvidia in support of their Bluefield DPU. We've done similarly with Intel for their SmartNIC. And so if you look at the combination of these things, it's very powerful and it actually, like you described earlier…
Steve Saunders:
it's all best in class?
Shekar Ayyar:
Absolutely. And it can scale up or scale out as the customer might need. And the idea that you can now have one operating environment that can go all the way from the most sophisticated routing applications to the most simple switching applications all with one layer of software is extremely powerful.
Steve Saunders:
This is pretty revolutionary stuff because if you look at the history of the telecommunications industry from the first telephone, it was like, buy my product. Right? And in some ways, 120 years later, we're still in a product-centric mode. If I go to Nokia, they want to, they'll say, what are you trying to do? Well, here's our product for that. But you are discarding that. You've thrown that away.
Shekar Ayyar:
Absolutely. I mean, my analogy to that is the compute industry essentially got that ‘aha’ moment probably in the late nineties where they said, look, we can't actually be on single siloed environments. I mean at that time it used to be HP and Dell and IBM, and they then kind of came around to VMware style virtualization and then the cloud AWS, and now of course GCP, Azure, OCI, and nobody actually buys a server if it's not virtualized. And many people are all hybrid in their thinking about on and off-prem compute. But network is still about 30 years behind. Everybody still says, oh, I need this box. It's a part number from Cisco or Juniper or stuff.
Steve Saunders:
It's a very expensive way to build a network.
Shekar Ayyar:
Extremely expensive, very inflexible.
Steve Saunders:
Great for the equipment manufacturers. But yeah, we are moving beyond that.
Shekar Ayyar:
Exactly, exactly.
Now, interestingly, even they, and as you know, Cisco with Silicon one for example, is an idea where they are now saying, look, we can't actually be siloed systems. We need to start working in that processor space. And so the idea that something like Silicon One potentially with something like Arrcus on top could service the industry. I mean, these are just revolutionary concepts for the network industry, which has historically been so conservative.
Steve Saunders:
A big picture question, how do you feel that the tier one carriers are getting on with reinventing themselves for 21st century communications as it were? Because they're in an interesting position now, aren't they? Because I think they feel that they risk being disintermediated or going out of business in the next five years because of the level of change. At the same time financially, they make quite a lot of money at the moment still. So that's the worst combination. Do you think they're being aggressive enough with looking at their own business models?
Shekar Ayyar:
They are not, unfortunately. So I think slow is the word I would use in general to describe the carrier universe. And the money question is interesting because honestly, I think the money pool is also drying up. I mean, it used to be a kind of cash cow business when you had to provide data and voice to consumers like you and me and then charge us a boatload of money; so that no longer is the case. And so increasingly it is imperative for carriers to find ways that they are going to make money in the future. And it's not actually quite obvious how that's going to happen. Now, there are opportunities for the carriers to do that. So, for example, I think they should be taking their network, making it more flexible, making it more dynamic, getting into the business of supporting interconnection, supporting AI, and not just focused on, I have a 5G network and I can get you better connectivity. I mean, that is the old mantra. Now the problem lies.. So in technology, actually they have a path forward because there are companies like us that can partner up with frankly every carrier in the world and show them how to build out this flexible network. That part we have done. I think the more challenging environment for them is really on the sales and marketing and go-to-market side where their genes have not yet been cultivated to go in and make that transition from old sales to new selling.
Steve Saunders:
Absolutely. Yeah. It makes the whole thing very exciting. But of course you are in a good position here to provide the platform. If they build it, the services and the applications will come.
Shekar Ayyar:
Yeah, it's much faster because the carrier world used to be about “build it”, and six years later we can have a service deployed. We're telling them, build it, and you can deploy the service today. You don't even have to wait until tomorrow. I mean, for example, right now we have, and we will be announcing the soon we have large network operators and enterprise companies, et cetera, that are building multi-cloud network fabrics using Arrcus that can then be used to deploy services both to their internal customers as well as external customers. That sort of an engagement has been done over a matter of six months. And so it wasn't six years. So in six months, somebody is going to be ready to go and launch a service and actually provide services to their customers, soup to nuts and start to finish.
Steve Saunders:
and revenue generator.
Shekar Ayyar:
and revenue generator. So this is the power of what we do.
Steve Saunders:
Fantastic. It's great to meet you here at, Barcelona. It sounds like things are going really well.
Shekar Ayyar:
Like wise. Wonderful to see you. Thank you so much.