Cloud hits peak of hype cycle; better start planning for its downfall

Capacity planning and monitoring makes clouds concrete and usable

No matter whether you want an internal, external or hybrid cloud, or even if you're satisfied plugging a little SAAS here and there into a virtual infrastructure, you can't get away from the limits of hardware.

If you run any kind of IT operation, no matter how virtualized and abstracted you get, you have to deal with hardware somewhere along the line. Capital costing, hands-on-supporting, floor-space-taking, power-gobbling, big-fat-asset-depreciating hardware that drops 20 percent in value the microsecond you drive it off the lot.

Even Don Whittington -- the Domino Sugar CIO who has moved almost his entire IT infrastructure into clouds, hosted environments and various other levels of abstraction for a decade -- had desktops and laptops and departmental servers and switches and routers and cabling and all kinds of other things that are as good at gathering dust as they are moving data.

So, even though cloud computing is going to so completely change IT that veteran geeks will get lost on their way back from wasted visits to userland because the dank basements into which IT is crammed at many companies will have been cleaned out, air-freshened and replaced with banks and banks of outbound marketers on countless VoIPphones.

Actually, that's an exaggeration. In its most recent Hype Cycle report -- which simultaneously feeds on hype and feeds on it, Gartner reports that Cloud Computing has reached the peak of its hype cycle and is poised over a dramatic drop into the Pit of Disillusionment.

Which all sounds a lot more D&D than IT, but take my word for it, that's what it is.

Neither the hype cycle nor the technologies customers that ride its roller-coaster market-graph will ever escape the need for IT to plan for the "immediate" needs of business units, whether the technology to fulfill those needs is imaginary or not.

Either way, Hyper9 and companies like it are ready.

Hyper9 released version 3 of its capacity planning software with new features designed to cover virtual-server installations with more than 10,000 virtual machines, and others designed to cover capacity planning, usage monitoring and chargeback on (primarily) private clouds.

End-user companies that need capacity planning for clouds are generally the ones that have virtualized 60 percent or 70 percent of their servers and are building internal cloud platforms to make resource-use among the hardware supporting all those VMs as efficient as possible, according to Jonathan Reeve, VP of product strategy for Hyper9.

Automating performance and capacity management makes it easier for IT managers to assign specific costs or quotas to business units so Bob from Accounting doesn't go to the self-service portal and launch himself a new SAP instance every day (and never shut them down) just to get a fresh server.

It can also identify performance bottlenecks to make it simpler to shift resources from areas of the server farm that are overburdened to those chugging along with their little server feet up on the desk.

Hyper9 isn't the only game in town. Amazon provides planning tools; HP's Cloud Assurance is supposed to do it; it's the whole point of ITIL.

It's not that different than physical capacity planning; you just have to lead your target a little differently.

It's mainly designed to apply management, capacity, lifecycle and resource-management policies across thousands of VMs at once, from a single dashboard. It also workflow and monitoring for VM provisioning, usage statistics, performance metrics and comparisons of actual performance against service level agreements.

It does the same boring (but vital, ask your data center manager) stuff that keeps the data center from melting down every day, but expands to include both virtualized infrastructures and public cloud platforms.

In fact, if you highlight a chunk of your infrastructure, it will estimate how much it would cost to port the chunk to Amazon's EC2 or other cloud platforms, just in case you're fantasizing about getting rid of the hardware altogether.

Don't forget to blow off the dust, first.

Kevin Fogarty writes about enterprise IT for ITworld. Follow him on Twitter @KevinFogarty.

Top 10 Hot Internet of Things Startups
View Comments
You Might Like
Join the discussion
Be the first to comment on this article. Our Commenting Policies