White Box

Cisco Takes Bold Software Step to Counter White Box Threat – Report

Two years ago, at Mobile World Congress (MWC), former Cisco CEO John Chambers predicted that his company's biggest competitor in future would not come from the likes of Ericsson, Huawei and Juniper, with their proprietary technologies, but from so-called "white box" solutions. (See Think Outside the White Box and CEO Chat With John Chambers, Cisco.)

This week, the network equipment giant was reported to have taken its boldest step yet to address the white box threat.

According to an as yet unconfirmed report from The Information (subscription required), Cisco Systems Inc. (Nasdaq: CSCO) is planning to start selling the software that goes into its networking devices on a standalone basis. Citing two sources familiar with Cisco's strategy, the report says that customers of the operating system -- branded Lindt [Ed: Does it taste like Swiss chocolate?] -- will be able to use it in conjunction with hardware from other companies, instead of relying on Cisco's purpose-built chips.

While there have been no official statements (Light Reading approached Cisco but has yet to hear back), the move would clearly look like a response to the business challenge posed by Arista Networks Inc. and its ilk -- rivals that have already started selling software on a standalone basis.

"Cisco's switching share has declined from around 70% in 2010 to the mid-50s today as Arista and white box alternatives have won share in the data center," says James Crawshaw, a senior analyst with the Heavy Reading market research group.

So why are white boxes such a big threat?

In case you'd forgotten, a white box is a cheap(er), off-the-shelf server or switch running open-source software. From a cost and flexibility perspective, their attractions are becoming increasingly apparent to companies building enterprise networks and data centers. But in a conversation with Light Reading's Steve Saunders during MWC 2015, Chambers argued that white boxes would eventually become viable across a broad sweep of wide area networks. (See Think Outside the White Box.)

For the makers of proprietary equipment, the ramifications of such a development are scary. Unless they somehow adapt themselves to a white box world, many traditional vendors and service providers could ultimately disappear.

Before The Information's story surfaced this week, Cisco had already taken a series of major steps aimed at preparing for this software-driven future. Way back in November 2014, it was revealed by Light Reading that Cisco had reorganized its then 25,000-person engineering team into separate hardware and software groups in response to customer demands. More recently, job cuts have reportedly been linked to Cisco's need for "different skill sets in a software-defined future." (See Troubled Cisco Looks to 'Bust Silos' and Cisco Set to Cut 14,000 Jobs – Report.)

But Lindt would go much further, and could have an unwelcome impact on Cisco's earnings. By selling switches based on its proprietary silicon, says Crawshaw, the company can maximize profits, lock in customers and guarantee better switching performance, which is what has mostly happened up to now. With something like Lindt, Cisco would basically be giving up the profits it makes on its proprietary silicon.

For more NFV-related coverage and insights, check out our dedicated NFV content channel here on Light Reading.

Pressure from customers has been mounting, though. While they might value some of the performance benefits of the proprietary silicon, web-scale customers want the control and "programmability" that comes with white box technology, explains Crawshaw. They also want to avoid being locked in to one vendor's gear.

And it's not only the web-scale players that seem to have been demanding change. "Cisco has finally woken up to the fact that it is not about shifting tin but software, and how you consume that is key," said Simon Niland, a managed networks specialist for Australia's Telstra Corp. Ltd. (ASX: TLS; NZK: TLS), during the MPLS, SDN and NFV World Congress in Paris last week. "You want to make decisions month by month and so it is about not tying yourself to one solution." (See Tata, Colt Take Vendors to Task on SD-WAN Interoperability.)

Quantifying the earnings impact of a white-box-type move is not easy. "The profit margin on software is in fact much higher than hardware because of the costs of producing the semiconductors," says Crawshaw. "However, the absolute amount of profit on the pure software solution will be lower than the current model of hardware plus software -- how much lower remains to be seen."

Exactly how far reaching the latest software move will be is still unclear, and probably won't become clearer until Cisco itself has something to say about the plans (assuming, of course, there is some basis to The Information's report). But Cisco is already well on its way to becoming a "software-first" or "software-only" player in parts of its business, says Gabriel Brown, a principal analyst with Heavy Reading.

"In the mobile carrier network, its EPC [evolved packet core] and SON [self-organizing network] products are all available as software-only solutions, or can be pre-integrated with UCS [unified computing system] hardware," he says. "It is a customer-driven choice."

"I don't think Cisco will walk away from customers that want pre-integrated hardware and software, but its strategic direction in the mobile core is already heavily software-oriented," adds Brown.

Indeed, Cisco has made it quite clear that growth will come from opportunities in the software and services markets. In that regard, it is little different from many other traditional hardware players under pressure to evolve. But the white box forces at work cast huge uncertainty over its future status and shape.

"I am pleased with our progress on business transformation to software and recurring revenues," said Kelly Kramer, Cisco's chief financial officer, in mid-February, when the company reported a year-on-year sales decline of 2%, to about $11.8 billion, for its November-to-January quarter. As software becomes paramount, balancing customer needs with investor expectations is likely to get much harder.

— Iain Morris, Circle me on Google+ Follow me on TwitterVisit my LinkedIn profile, News Editor, Light Reading

mendyk 3/29/2017 | 9:10:20 AM
Extremes I wonder how many operators really want to make technology decisions on a "month by month" basis. Lock-in is an obvious bogeyman, but "month by month" sounds like unnecessary instability.
Roz Rose 3/29/2017 | 10:13:00 AM
Re: Extremes On that same note, my sense is that the number of CSPs who want to procure HW and SW separately and do their own integration is limited. The Webscale guys are all about DIY. In my research, they (CSPs) profess affection for "best-of-breed", "choice" and "disaggregation", but it's not clear to me that many have the skill sets (yet) to consume equipment in this manner.
iainmorris 3/29/2017 | 10:52:32 AM
Re: Extremes Yes, the CSPs talk the talk but they don't walk the walk. In the main, that is.
mendyk 3/29/2017 | 11:42:53 AM
Re: Extremes Given the fact that there's a pretty big reality gap now regarding virtualization, this starts to feel a little like the current political climate, as in, "We know what we don't want, but we're not so sure exactly what it is we want or how to get it." It's not a great place to be.
amitawal 3/29/2017 | 2:44:02 PM
Integrated vs Separate HW SW Once HW is separated and common whitebox, then the CSPs have an interesting decision to make - how much to pay for what software capability and capacity. It will be increasingly transparent to see the price to be paid by feature bloat in RFP - have all ticks even if not required"" vs smaller vendors or open implementations which will focus on a subset of usecases & protocols.
lcw 3/30/2017 | 5:02:58 PM
white box Cisco claims that since they're Broadcom's biggest customer, their manufacturing costs are so low that they should always be able to undercut white box. So on the surface, it seems like disaggregation would be unnecessary to win purely on price.

But perhaps some customers are demanding disaggregation for reasons. E.g., take the the Intel Atom C2000 bug, which affects many white box switches... In theory, disaggregated SW might make it easier to swap the HW.
Sign In