The world of cloud computing continues to evolve on an almost daily basis and the recent exit of Rackspace from the IaaS market can be viewed as major change in the evolution of the cloud.
If one of the hyper-scale darlings of the IaaS space declares that it is now a commodity, and that they plan to move to managed, higher value/cost offerings, then what hope do Tier 2 and 3 players have for local markets or specific industry verticals? MSP/CSPs looking to survive, and yes I mean survive, beyond the next three years need to find a niche and develop their tailored offering quickly.
This transition from “Simple IT-aaS” to “Mission Critical-aaS” is what I am calling the second wave of cloud. If you consider x86 virtualisation technology as a pre-cursor to cloud, the adoption cycle follows a definite transition: first the virtualisation of the test and development environment in order to identify the benefits; next, as the processes mature, the roll out of VMware across the entire production x86 estate.
While this environment may achieve all the anticipated results planned for with this new technology, the UNIX and Mainframe backend systems haven’t been touched. All virtualisation has achieved was removing 50% of the cost out of 30% of the overall IT budget.
Cloud is following a similar adoption model. It may be a noble pursuit but moving an enterprise’s ‘Simple IT’ to a public cloud provider will barely move the needle when you take a macro perspective on measuring your overall IT budget. If you manage to take all in-house x86 systems and move them off premises, you will achieve savings and clear the decks of a number of issues and support nightmares in the process, but will you really affect your bottom line IT cost base?
No longer can an MSP simply throw a few x86-based VMware guests up on the web and hope to build a sustainable business. Cloud service providers must offer a specific software solution to a vertical. That solution must include, and uniquely understand, how to offer quality of service, availability and security. As this evolution to what I call the second wave of cloud workloads or ” Mission Critical IT-aaS” becomes the new battleground for MSPs, then the focus will be on innovation and daring to be different.
This next battleground centres on how to take the 99.999% availability, 24x7x365 applications that run your business, and deliver them as a service either on-premises or (preferably) off-premises in the public cloud. My belief is that only when you can deliver on this project will you truly be able to declare victory.
So how do you achieve 99.999% in the cloud when the current leader, according to Gartner’s Magic Quadrant for Cloud IaaS, is only able to deliver at best 99.95% reliability?
One approach I would advocate is to map out your mission critical workloads. Look for common components of the applications, such as database and middleware and look to decompose your applications into horizontal ‘utilities’ that can be delivered as a service. One example would the enterprise database. I am sure in any large corporate there are hundreds of servers running databases, such as Oracle and DB2 on UNIX servers. Why not try to consolidate these databases into a horizontal utility that can be provided to any application that requires a database?
This way you can centralise demand and offer Database as a service (DBaaS) to your applications. If you take this approach you can then, based on latency and other factors, decide whether this utility is on-premises or in the public cloud.
Following the crowd and hoping to be lucky could see some of the 35,000 MSPs currently operating globally to go to the wall, and out of business. This is to be expected as any marketplace matures new entrants are consolidated as the successful survive – call it the service provider equivalent of Survival of the Fittest.
So to all those MSPs out there not wanting to become the next dodo, I suggest you take a long hard look at the building blocks of your business. Consider your infrastructure and ask yourself one hard question: have you dared to be different?
Or have you followed the accepted practice and built simply a derivation of what everyone else in the industry is offering? However, a word of caution, if you look to aggregate multiple databases onto one utility service, then that service better scale, be available and be secure.
“No cloud platform can scale, be available and offer enterprise grade security on the levels I need to offer high I/O workloads such as database as a horizontal utility,” the dissenting voices may say.
You may not like the answer, but better to know now rather than let the market find you out: There is one cloud platform that can handle enterprise (class) cloud (workloads, in one) system… IBM’s new Enterprise Cloud System.