Arrcus is taking a game-changing new hardware-agnostic, open-source approach to the challenge of cloud infrastructure. But in an overheated and over-crowded space, where IPOs are still on the dark side of the horizon, is that enough? Company founder and CTO Keyur Patel thinks so.
While market incumbents like Cisco typically have “rigid point solutions,” Patel told Silverlinings Arrcus has built a flexible software architecture that uses merchant silicon and provides an abstraction layer so companies can focus on services rather than chips.
“Our software does to silicon what VMware did to compute,” he said. “We can actually run the same software in the cloud and use cloud constructs to build a router or on-prem use on-prem servers to build a router out of it. So, what it gives us is breadth of deployment points that other folks don’t have.”
Founded in 2016, Arrcus has raised roughly $77 million to date to tackle the routing, switching and multi-cloud networking markets. According to Patel, the company was created to capitalize on the confluence of several market trends, including a shift from Layer 2 to Layer 3 connectivity in data centers, need to handle more interconnections, growing desire among operators to monetize networks with services and the rise of merchant silicon in the router market.
Though merchant silicon was fairly common in the switching market at the time, routing until then had remained a “captive market” with vendors like Cisco each supplying their own silicon, Patel explained. But once Broadcom came in to mainstream routing silicon, Arrcus’ founders realized they had an opportunity to holistically address both routing and switching with hardware agnostic software, he said.
While Arrcus’ approach was novel at the time, Patel acknowledged several other companies are now pivoting toward a similar system. He pointed to Cisco as the best positioned on this front given its portfolio already spans routing and switching. But Patel added Juniper and Arista could make a play as well.
Cisco's approach
Indeed, it seems Cisco has already stepped up to the plate.
A company representative told Silverlinings it offers a "have it your way" model with its routing and switching solutions, allowing customers the option to buy complete systems with integrated hardware and software or take a disaggregated approach with just hardware or just software.
The representative added, "We can abstract our software onto non-specific hardware platforms," with Cisco's IOS XR routing and Cisco's NX OS switching products examples of its disaggregated offerings. Its Silicon One chipsets can also be run using a disaggregated model, which is something Meta and others are already doing.
DriveNets is also on the pitch in a significant way. Like Arrcus, DriveNets is pursuing a hardware-agnostic software approach to routing and switching. And in addition to bagging $587 million in funding to date, DriveNets has already managed to secure a position as a key vendor in AT&T’s network.
When asked how Arrcus will be able to avoid being crushed by a market giant like Cisco or competitors like DriveNets, Patel pointed to the size of the market and a trend toward multi-vendors solutions.
Besides, incumbents like Cisco, Juniper and Arista who emulate it will “suffer from the fact that they have legacy products to handle and manage” which don’t allow them to build newer generation architectures, he argued.
He also dismissed potential threats from other new startups in the routing and switching space, arguing most of these are focusing on point solutions to start will and will have a hard time going back and revisiting their architecture.
Patel concluded: “As long as deployment architectures are common – which is happening now – we can deploy everywhere. So, the amount of market or the blast radius we have for the market we’re targeting is way, way more than the competition that is out there.”
Arrcus' constant vigilance
The CTO said he doesn’t see any other external hurdles that might hinder Arrcus’ growth. But he added it must be careful not to become complacent. Otherwise, it could end up being its own worst enemy.
“Building a great software product is like having a human body and managing that. The body is in a perfect condition when you are in your teens and early 20s,” he said. But if you don’t maintain it as it grows older, it can become more unstable and there can be bugs. It could even get to the point that you throw your back out just by sneezing – not that that’s a personal example or anything.
So, more than anything else, Arrcus will focus on keeping its software in tip-top shape.
Arrcus' long term prospects are unclear, particularly given the presence of DriveNets in the same market. Patel said the company’s five-year plan is to grow.
One successful exit option for Arrcus would be an acquisition by a competitor like Cisco. Whether such a deal will come to fruition is the big question.
January 27, 2023 8:45 pm ET: This story has been updated to reflect the correct amount of funding Arrcus has raised to date.