Network virtualisation and VMware's quest for the holy grail: Page 3 of 3
By Edwin Yapp August 12, 2014
The hurdles to adoption
Despite the promises that network virtualisation holds, Casado admits that the technology is still being adopted somewhat slowly, but argues that it’s not technology that is holding back progress, and is more about culture and change.
He says that while the technology is well ahead of the curve for most organisations, what is difficult is for them to change their thinking, specifically the need to redefine the traditional roles of networking and the breaking down of a ‘silo mentality.’
“Software changes fast and hardware changes slower, but culture change is by far the slowest,” he points out. “[But] that’s where we must focus because if we want to change the industry and make networking totally automated and flexible, we need to change the thinking of organisations.”
Despite this huge hurdle, the engineer in Casado remains cautiously optimistic that things will eventually change, especially when the market demands it.
“For me, things like operational agility is not so obvious [as a driver for change] as it’ll take more time to see this,” he explains. “But security is a compelling reason for change as the only thing growing faster is how much companies lose in security breaches.
“One of the reasons for this is that we don’t know how to put security control in the data centre. With network virtualisation and SDN, we can address this, and it becomes very compelling to customers,” he adds.
Analyst weighs in
Ovum’s Illsley (pic) was less sanguine about the adoption of SDN as he believes that the technology has fragmented into three main approaches: Proprietary, open, and hypervisor-based.
He says the hypervisor (overlay) approach is the most commercially ready and gaining the most traction, although it is said to suffer from ‘vendor lock’ fears because it is typically based on a stack from a single vendor, like VMware for example.
“However, we have seen the number of proofs-of-concept and pilot projects grow in the last 18 months, and this experience has seen the understanding by users mature, and has led to the three-pronged approach,” he adds.
Illsley believes that the proprietary approach is a stopgap measure by incumbent suppliers as a way to protect their revenue streams in the short term while the market and standards mature. On its part, the open approach is struggling to come up with a coherent strategy.
Asked what has changed since 2012 when SDN was being touted as the next wave of networking, Illsley concedes that it is difficult to say at this stage how the SDN technology will evolve, as open standards are being demanded by some sectors – telcos, for example – while other industries are less concerned and just want to get operational and hardware gains.
“Today much of that advice still stands, but I would add that the selection of approach to SDN should be based on a finite life expectancy, and the cost/ benefit equation worked out accordingly.
“So if you are a Cisco shop, then adopting Cisco’s approach may be the most pragmatic for the next five years, but if you are a VMware shop and your hardware is a mixture of suppliers and age, then the overlay approach may be best.
“The open approach would be for those with a long-term vision, as the investment in time and resources is the greatest,” he adds.
Ultimately, Illsley expects the choice of approach to change over time, but enterprises must factor in the cost of change as the market matures.