Despite all the hype, disaggregation doesn’t make sense for most enterprise networks.
There’s a lot of talk today in the industry about network disaggregation. Internet giants like Facebook say it provides the flexibility that traditional networking equipment doesn’t. Networking vendors like Dell and Juniper are jumping on the disaggregation bandwagon, saying they want to provide their customers with more choice.
But I’m not going to beat around the bush: Despite what you've been told and sold, there is no compelling case to migrate to a disaggregated network in the enterprise.
Before I make my case, let’s look at what network disaggregation involves. Disaggregation offers the option to select an OS from one vendor and run it on compatible hardware from a different manufacturer. This situation has largely come about because of powerful merchant silicon Ethernet switching ASICs available from vendors like Broadcom, Intel and Cavium eating away at the performance edge previously enjoyed by custom switching ASICs developed by network hardware vendors.
Disaggregated Ethernet switch (“bare metal”) hardware for comes in two forms. “White box” switches are similar in some ways to buying a generic medication instead of the original name brand; the packaging may not look as exciting as the original and you've never heard of the manufacturer, but it's cheaper and with any luck contains pretty much the same ingredients as the original.
“Brite box” switches (a term coined by Gartner) are most easily thought of as white box switches veneered with the comfort of a brand you recognize. This is the equivalent of a pharmacy-branded medication; most likely both this and the generic are identical, but this somehow feels more trustworthy. It might cost a little bit more than the mystery generic, but if it turns out to be useless you know who to shout at.
Without an OS, white box switches are a little pointless, so to fill this void a number of startups like Cumulus Networks, Big Switch Networks, and Pica8 appeared. More recently, some of the big vendors --like Juniper-- have started making some form of their OS available to run on bare-metal hardware.
Why enterprises won't migrate
There are many supposed benefits of network disaggregation, but they don’t really apply in the enterprise.
- Lower costs: Forrester's report “The Myth of White Box Network Switches” estimated that the cost of a disaggregated switch platform calculated over a 6.6 year period was not significantly less than buying a vendor product when you include support and operational costs.
- Run one OS across every platform: Yes, but most likely only if those platforms are certified by your OS vendor, which may also limit hardware options to a particular brand of merchant silicon.
- Deploy switches using puppet/REST API: This sounds great, but is the business ready to fund development of the solutions necessary to automate the management of the network this way? How often are you deploying new switches?
- Reduced vendor lock-in: Certified hardware can be swapped in and out as desired, but does this simply shift the effective lock-in to the OS?
- You can run other software on the switch: Sounds good, but do you have a production use case in mind? Additionally, traditional vendors are starting to offer this option too, so it may stop being a differentiator.
Another consideration is that the capacity of a typical white-box switch (48 x 10 Gbps Ethernet plus some 40 Gbps uplinks) means it is aimed at the data center environment, and likely has little use outside the DC or the largest of user sites.
For these reasons, I don't see a compelling reason for a typical enterprise to migrate from their current platform to a disaggregated switching platform today.
One size does not fit all, and there are a few things that may alter the swing of the decision pendulum:
- As the cost of commodity hardware drops, switches may not need to be in the network for the 6.6 years used by Forrester in its report.
- Migrating from individually managed switches to an Ethernet fabric solution may reduce operational costs or improve service levels if your data center is large enough.
- Some enterprises will need fine-grained control of the OS or to run additional software on the switches, but they are also accepting the additional development cost and risk.
- There is a market for disaggregated switching as a network monitoring fabric (e.g., Big Switch), at least in larger data center environments.
What becomes evident with disaggregated switching is that size matters. The hardware available today is best suited to companies with large data centers or very large offices, but may not be particularly useful elsewhere. The larger the data center and the more the network is changing, the more relevant it may become to consider a strategy of disaggregation if the chosen OS offers a useful feature or allows faster deployment of services.
Small to mid-size enterprises are unlikely to see a significant cost benefit and are unlikely to be in a position to capitalize on the extra flexibility.
Your mileage may vary
Ethernet frames don't necessarily move any better -- or worse -- on a network built from disaggregated switches than they do on a network built using regular vendor hardware, especially when they are being switched by the same ASIC under the hood. Thus the decision to deploy disaggregated switching should be driven by a business need that can best (or only) be solved by that solution. There will without question be enterprises for whom this is the case, but there will be many more who can't or won't make the leap.
Learn about technologies and vendors critical to the Future of Networking at a two-day summit presented by Packet Pushers at Interop Las Vegas this spring. Don't miss out! Register now for Interop, May 2-6, and receive $200 off.