2013 Predictions: The End of Big Iron
These predictions were first published as a contributed article in VMblog’s 5th Annual Virtualization and Cloud Prediction Series.
For the last few weeks, the Nutanix marketing team kept nudging me to write a blog on 2013 predictions. I kept pushing it off, perhaps subconsciously, because it would have been a royal waste if the world really came to an end. 12/21/12 is safely behind us, and I’ve found my mojo to write. Something tells me I did better with my 2012 predictions than the Mayans. You be the judge. Tweet me a score between 1 and 10, 1 being an absolute dud and 10 a near perfect. @trailsfootmarks (hashtag: #Nutanix).
I look at 2013 as the beginning of the end of big iron. Public cloud, virtual datacenter appliances, Taiwanese ODMs, commodity flash SSDs, and 10 GbE will trump their expensive legacy counterparts.
The Year of the Public Cloud
In this constant yin-yang between own vs. rent, we’ll reach a tipping point in 2013 when customers — including F1000 — will start to build new applications in the public cloud. Developers outside of IT have always wanted agility, and they have begun to express their pent-up frustration bypassing IT. CEOs and CFOs have already begun to embrace the elastic and fractional consumption model of public cloud. CIOs will follow suit. Security will take care of itself by 2015 through innovation and fading skepticism. SLAs need work though. A growing and demanding customer base will make cloud service providers honest in the coming couple of years.
Google Compute Engine and Cloud Storage will begin to look good enough by 2013 end. Amazon will have a worthy competitor flexing its “economies of scale muscle.” Rackspace could find a worthy suitor this year. While Linux developers will gravitate to Amazon and Google, Windows .NET developers will default to Azure. VMware Cloud Foundry’s and HP’s Cloud fortunes are looking increasingly bleak this coming year.
The Year of “All In” Virtualization
While we continue to measure virtualization penetration in terms of servers (applications) being virtualized, there is a whole new class of datacenter and desktop services that are on the threshold of being virtualized. 2012 was the year of VDI. Secular forces such as BYOD, universal access, security and compliance, and Windows refresh fueled the VDI adoption fire, and will continue to do so in the coming year. 2013 will also be the year when almost all datacenter services — security (firewalls, VPNs, IPS/IDS), networking (switches, WAN optimizers, ADCs), and storage — will run as virtual controllers inside a hypervisor “sheet metal”. The days of special purpose pizza-box appliances are over. 2013 will witness enterprise-grade virtual appliance offerings from almost every datacenter appliance vendor who wishes to survive in the day and age of public cloud — one in which Amazon and Google rarely buy pizza boxes for specific purposes. Technologies such as PCI passthru make it possible for these virtual appliances to continue leveraging hardware offload engines for performance in on-premise private cloud environments.
This phenomenon, which VMware prefers to call Software-defined Datacenters, will result in some unusual technology alliances such as Citrix-VMware for NetScaler and Citrix-F5 for XenServer!
Shifting Sands in Alliances, M&A
Some key partnerships are dead in 2013 — Cisco-EMC, Cisco-VMware, and CommVault-Dell. Competitive acquisitions took a toll on these relationships. The stakes are too high now. VMware/EMC drew first blood by buying Nicira from right under Cisco’s nose. Cisco is seething, and they will act in 2013. Wait and watch.
2013 will put XenServer to sleep forever, as Citrix continues to wilt under Microsoft’s Hyper-v pressure. XenServer will not survive as the second open-source hypervisor, even if Cisco buys Citrix. There has never been room for two open-source products in the same category. Like consumer tech, the winner takes all the community mindshare. KVM will be the survivor, just like MySQL (over PostgreSQL) and Linux (over FreeBSD).
Next year is perhaps the last year when money is cheap, i.e., cost of borrowing is low for large leveraged buyouts. Expect to see friendly or hostile takeover bids for SaaS players like Workday, Service Now, et al.
“Good Enough” Continues its Rampage in 2013
Good-enough will eat its high-end counterparts for lunch with its ease-of-use, ubiquity, and network effects. Hyper-v will continue to play spoilsport with VMware’s high margin vSphere business. 10 GbE networks will continue to fade Infiniband and Fiber Channel into oblivion. FCoE will be dead, for good (phew!). MLC flash will finally euthanize its SLC and eMLC counterparts. Native mobile apps will continue to give way to HTML5. Taiwanese ODMs of x86 servers, with their economies of scale selling to cloud providers, will make a big dent into the enterprise. HP will continue to unravel, given its mess, the margin pressure from Taiwan, and lack of a coherent strategy.
And finally, big iron will give way to software-defined scale-out datacenters.