Like a lot of people, I tuned in to John Chamber’s keynote presentation at Cisco Live last week to see how he views the world. Whether you care for him or not, he has a unique view of a significant chunk of the infrastructure market. His assessment of the market catches a lot of attention, especially from Wall St.
The slide everyone is abuzz about is this gem:
John articulated a case for why a lot of hardware vendors are not going to be around 4-5 years from now – and he’s not off base. You can argue about the exact list of vendors, but the broader commentary on hardware vs. software in the networking landscape is spot on. Basically, if you’re releasing some kind of networking chassis in 2014, you’re really missing the direction of the market.
Cisco has very clearly laid out a strategy where a non-trivial amount of their revenue will come from software and software-related services over the next several years. These claims are strongly backed by their announcements for multiple virtual networking products like the vASA, vWAAS, and especially the APIC-DC and APIC-EM. To drive the point home, they took over all of Moscone West conference center in San Francisco for their DevNet pavilion where hour after hour of presentations were lined up to teach network administrators how to write scripts and automate networks.
Chambers has made it clear: Cisco is doubling down on software in a big way and the market is coming with him.
A snappy retort would of course be “Didn’t Cisco just announce a whole new line up of hardware based switches in their Nexus lineup?” Yes – they did. And it was the right thing to do – doubling down on software should not come at the expense of designing the right kind of network. The reality is that servers are nowhere near dense enough on a port-by-port basis and just don’t have the performance needed to pull it off. Layer 2/3 switching still needs to be in hardware and the latest Cisco hardware shows that there is still room for differentiation in the Ethernet switching segment.
So while hardware makes sense for the forwarding plane, it doesn’t make sense for the application plane where services like load balancing, ADCs, firewalls, and WAN optimization live. That is evolving to software and if you don’t have meaningful software offerings today, you’ve missed the boat.
And that goes back to Chamber’s slide. There he shows what one of my friends called a dystopian vision of networking. The vendors listed in the column under “Today” that didn’t make it to the 2018 column still don’t get it. There is even a load balancing (ADC) vendor in there… and sure enough, they released new chassis units just this year. Guess who didn’t make the cut in 2018?
Do your own assessment. Decide for yourself – Chambers gets his business, but he doesn’t run yours. Do you see software taking a more front and center role? Do you think vendors that have focused on hardware are going to make it in the long term? Take a long hard look at your vendors and their competition. Look at the ones that have taken the virtual plunge and made their products not just software, but feature identical to hardware. Then look at the ones shipping big iron…
That chart from Chambers is starting to look less like a dystopian vision and more and more like harsh reality, doesn’t it?