Note: The following is a piece I wrote recently for TechTarget. Sadly, I rather dropped the ball on the brief, and it didn't make the "cut". It was meant to be a technical tip, but quickly became an opinion piece about the current economic uncertainty we are facing - and its impact or not on the world of virtualization.
Sorry about the title. Generally I can't abide this fashion for putting software version numbers in things to describe the next generation - like Web 2.0. Sounds too much like some sort of media construct...
Over here in Europe we all seem to bracing ourselves for yet another round of the credit crunch, as European leaders once again meet to discuss what can be done to reduce the affect of the Euro sovereign debt crisis. One would think that the market uncertainty was creating gridlock with IT projects, with everyone batting down the hatches and tightening their respective belts. I took it upon myself to do an informal unofficial survey while at last week's VMworld Europe event in Copenhagen. The responses from the many people were many and varied but I did see a pattern emerging.
I found the mood amongst the unfortunately entitled “PIGS” (Portugal, Italy, Greece and Spain) was decidedly bearish. That’s entirely understandable given many of these countries have recently had their credit rating downgraded, and some such as Greece are enduring IMF enforced austerity packages. In the very week of VMworld EU there was a 24-hour strike in Greece occupied by riots, whilst at the same time the “Occupy” campaign moved into areas of high-finance to register their discontent with the “system”. My European colleagues in these beleaguered regions were very pessimistic about the immediate short-term – with many reporting that budgets are cut and projects are cancelled. Interestingly these guys still felt optimistic to predict a slow upturn within three to four years. Much of this optimism is based on an assumption that there will be certain amount of “churn” as businesses are forced to upgrade from one version of their software to another, or are forced to upgrade hardware to another. While it's common for companies to “sweat” their desktop assets where the failure of one client device only affects one user, when it comes to back-end infrastructure components a certain amount of upgrade work is unavoidable. Much of this comes from software support being withdrawn from legacy versions, or hardware falling out of its warranty. It's always been the case that it is more beneficial to pay for upgrades, and get performance and scalability improvements in return – rather than paying extortionate fees for support of legacy system that offer no tangible improvements in performance or features.
Outside of the “PIGS” zone I detected a much bullish attitude amongst my European connections with many reporting additional spending. While some of this is the same “churn” I was outlining above, much more of it is proactive spending on IT infrastructures. For many they have projects to extend the primary benefits of virtualization to applications and services that have previously been regarded as too sensitive from a politics or performance perspective. That attitude has definitely shifted with many of my contacts reporting that senior management is directing them to embark on the process of going beyond entry-level/low-hanging fruit virtualization. For many of these businesses the benefits have been delivered in the previous decade, and virtualization represents a positive action to reduce costs, and do more-with-less, even if this means signing of purchase orders to acquire more server or storage capacity. Just like with virtualization 1.0, it’s been seen that to save money you need to spend some, and that this represents a long-term strategy – one that many see as a prerequisite to a cloudy future.
In addition to this “steady as she goes” approach to more of the same – more virtualization. There are green shoots of experimentation. Although it is still the early stages I have witnessed an ever-increasing interest in building stretched virtualization clusters. This is where traditional single-site clustering technologies such as VMware’s “High Availability” are “stretched” to embrace two locations. Thus offering site-to-site workload distribution and availability. Make no bones about it – it isn’t for everyone. The technical challenges are daunting to say the least, and the cost equally so. In fact, what is so remarkable about this interest is that concern is more about the technical issues of such a configuration – not the costs. For those who find their geographical circumstances prevent the use of stretched clusters I’ve seen an increased interest in technologies such as VMware Site Recovery Manager, Zerto and VirtualSharp. These are DR and workload migration technologies that use replication to bridge the divide between two locations. This reflects a perception that as virtualization ratios increase, so do business concerns about the dependencies on their availability and protection from unexpected outage. In many respects the double-edged success of virtualization – as its popularity increases, concerns about the business dependencies on it increase also. The only depressing side of this interest is customers' continued capacity to conflate high availability and disaster recovery as if they are the same thing.
The other areas of increased expansion and spending are in the new generation of hardware needed to deliver on the next phase of virtualization. Increasingly I see customers purchasing blade technologies together with network consolidation tools. This represents a double-whammy of increasing switch away from conventional 2U and 4U servers for a more denser compute model, together with a simplification of the network down to a limited number of physical NICs bonded together with an adoption of 10G networking as a standard. It’s regarded that this generation of hardware is required to offer the best possible IOPS demanded by the resource intensive Tier 1 applications. Contrary to popular belief, VMware’s recent vRAM licensing debacle seems to be well and truly buried with the August 3rd announcement that increased RAM allocations to all its vSphere5 SKUs. I heard very little grumbling about the new licensing policy, with most regarding it as expected and inevitable. VMware, on the other hand, continues to ramp up the talk about “cloud”. You can sort of understand why. It would be hard for a CEO of any company to stand-up and say “You know the stuff you have been doing. Well carry on doing that.” The CEO is meant to offer direction , not “more of the same” as their vision. But the reality is that precisely seems to be the case. More virtualization seems to be the order of the day, rather than more cloud.
With all this said we shouldn’t forget the effect of the first credit crunch. For many organizations it took more than a year for the banking tsunami that began in the U.S. to reach their European shores. While some of my colleagues are working in market verticals that seem immune to current economic woes, many Northern European countries are reporting static growth, rising inflationary pressures and pathological lack of consumer confidence. All this is before the cuts that the U.K. government has enacted really take their bite. It seems clear that European project stands on the brink of currency and debt crisis, that will see either the forced creation of some kind of “United States of Europe” or the two-tier, two-speed Europe. While my friends in the
U.S. bemoan how much of their debt is backed by the Chinese government – it appears as if the E.U. bodies are banking on a bailout from the Chinese too. I’m no political ideologue – but doesn’t anyone else spot an irony here? The vanguards of liberal capitalist democracy (the U.S. and Europe) are increasingly indebted to a largely totalitarian regime which has scant regard for human rights and free speech.
All that is certain is, this time around the politicians won’t be able to stand around exclaiming how surprised they are – this time there has been plenty of notification.