2014: The year the cloud killed the datacenter
Summary: This year more and more CxOs pulled the life support on their own datacenter infrastructure in favor of ever-maturing public and private cloud offerings.
Back in the summer of 2013 I wrote a series of articles on what I called the "Datacenter Zombie Apocalypse."
The first was an observation of the troubles IBM was undergoing trying to shift its very datacenter-oriented services and outsourcing business to one that was more cloud focused and how it might serve as a bellwether for the IT industry as a whole.
The other was about how, as an IT practicioner, one might transition their skills to be more cloud-oriented.
Where are we now with all of this, over a year later?
IBM's situation has gotten worse. While it continues to invest in Cloud, its services revenue is flat, and has been forced to divest two separate hardware businesses — its x86 server business to Lenovo and its silicon manufacturing capacity that powers their System p and System z enterprise systems to GlobalFoundries.
The company will almost certainly rebound, but at a cost. It intends to take a $4.7B charge against revenue in the 3rd quarter and almost certainly more layoffs are in store for the Armonk technology giant, which has over 400,000 employees globally.
I hate to say I told you so, but yeah, I told you so.
Agile CxOs in the last year haven't been able to resist using Cloud-based services to knock out the low-hanging fruit that make no sense for them to continue to run themselves.
My colleague, David Gewirtz, wrote an excellent piece this morning about how CXOs need to factor in CAPEX vs. OPEXin terms of how they make infrastructure decisions in order to stay competitive. If you are a decision maker or an influencer within your organization and you haven't made any moves towards the Cloud as of yet, then I suggest you read it.
In the last year, competing public and private Cloud providers have tremendously upped the ante in terms of lowering their costs and also improving the feature sets of their services. And agile CxOs in the last year haven't been able to resist using Cloud-based services to knock out the low-hanging fruit that make no sense for them to continue to run themselves.
In 2013, for the enterprise, the low hanging fruit for Cloud was all about tackling Development and Test workloads which could take advantage of Self-Service provisioning and also de-provisioning.
In 2014 for CxOs it was all about investigating, piloting and actively moving production workloads into the Cloud.
One of these areas that was and still is a key pain point for shifting out of the datacenter and into the Cloud is messaging and email. Sure, there are some companies that still probably have good reasons to keep their email infrastructure private, but for any sizable organization the actual managing and running it by themselves is a continually headache inducing and money-draining proposition.
Special Feature
ZDNet and TechRepublic draw on their community of C-level executives and business thinkers to prognosticate where business technology is headed over the next 36 months. This includes advice, perspectives, and opinions on both creating and reacting to the future.
Services such as Office 365 as well as through numerous hosted Exchange providers that have value added services of their own, which include complete support and day to day administration can allow you to get that monkey off your back, permanently.
And with Office 365 there's also the benefit of being able to introduce new functionality such as Skype for Business, previously known as Lync, for VOIP/Video collaboration and instant messaging.
Office 365, of course, is not the only cloud-based messaging and productivity service. There's also Google Apps and a few others out there, particularly if you want to look at private cloud, 3rd-party hosted implementations.
But you really need to examine what functionality is essential and which of these services are the best match for your requirements.
Another consideration that is driving CxOs to consider tearing down some, if not all of their datacenter space is legacy application transformation. And the root cause of that is the end of life for Windows Server 2003, which at this point is less than a year away.
The decision point is quite simple -- if you have to undergo a migration of the server OS to something more modernized (such as Windows Server 2012 R2, or even Linux) then why even do it in your own datacenter when a public cloud provider such as Azure, AWS or a private cloud hoster charges you just for computational time, storage, and data ingress and egress?
Special Feature
Infrastructure as a Service providers make a very compelling argument for businesses to stop running their own data centers and simply purchase server capacity on-demand and scale up and down as needed. This is our deep dive on IaaS strategy and best practices
The other benefit is there are no (or minimal) infrastructure software licenses to pay for in a public cloud either, regardless of what OS you run on. The service provider is eating those costs and reporting that licensing on your behalf. You pay for utilization. Period. And in some cases, industry-specific, vertical cloud providers can also meter out ISV application workloads and assume the licensing as well.
And that's just for straight up, "Lift and Shift" and like-for-like IaaS-type migration scenarios. Your costs can go down significantly if you can utilize PaaS, or Platform as a Service, where you don't have to manage VMs and the underlying OS at all.
If any part of your line of business application can say, use DBaaS, such as through Azure SQL, then you can really start saving money. You go from usually much higher blended computational and storage costs to strictly transactional and storage costs in that tier of your app.
There are other low hanging fruit besides migrating enterprise messaging and legacy apps, of course. Cloud Integrated Storage (CIS) is getting a ton of adoption from shops suffering from storage sprawl. And you've got your choice of providers that are in a never-ending race to the bottom in terms of pricing that would love to earn your business to offload all that stuff on your arrays that you could better leverage for something else.
Of course, it doesn't have to be an all or nothing proposition. Hybrid cloud scenarios, such as making use of the aforementioned CIS to offload infrequently accessed documents, can give you the best of both worlds.
For example, let's say you have regulatory or data sovereignty issues that prevent you from putting your databases in the public cloud. You can put the Web and middleware tiers of the app at the public cloud provider, and use a point to point VPN from your datacenter to the cloud provider to connect to your on-prem database cluster.
While software VPNs (such as those built into Windows Server and Linux) may work for certain kinds of scenarios (such as for distributed branch offices) public cloud providers such as AWS and Azure offer dedicated, encrypted MPLS connections through telecoms and co-location services like AT&T, Verizon and Equinix that provide lower latency and higher speed connectivity.
AWS's solution is called Direct Connect, and Azure's is called Expressroute.
All of this is just scratching the surface of what public and private cloud infrastructure can do in terms of alleviating on-premises datacenter pain.
Without question, 2014 was a watershed event for widespread cloud adoption. 2015 will without a doubt put nails in the coffins of many more datacenters.
Did your organization migrate significant workloads to the Cloud this year? Talk Back and Let Me Know.
