Beyond virtualization in the private cloud

Last week, InfoWorld’s David Linthicum made what I think is a very astute and long overdue observation: The term “cloud computing” has approached the point of being meaningless. As he points out, that’s not because nobody can agree on what the cloud is or has been created to do, but because every vendor in the IT space is falling over themselves to tell you just how “cloudy” they are — and spending a mint in marketing dollars to do it. Along the way, they’ve diluted the term to the point where it’s hard to tell what anyone’s really talking about anymore.

Far from being an exception to this, the increasing popularity of private clouds has made the problem even worse. Whether it’s a server vendor hawking blades and automation software as an all-in-one cloud solution, networking vendors pushing next-gen converged networking gear as a foundation for the cloud, or storage vendors selling you “cloud storage,” it’s difficult to get any of them to describe how their gear fits into the bigger picture.

Maybe this is a good time for me to take my own advice and ask why — why is the concept of the private cloud attractive in the first place? What problems are we trying to solve by implementing one? And in the most practical sense, what real-life components do we need to put together to fully deliver on its promise? Make no mistake, every level of the data center infrastructure has a critical part to play in a well-oiled private cloud. But due to the marketing FUD permeating the marketplace, it’s sometimes difficult to see how they’re supposed to fit together.

Why private cloud? At the most basic level, most businesses want one very simple thing: to spend less time and money building and managing infrastructure and more resources directly adding value to the business. From a business standpoint, IT infrastructure is an expensive and time-consuming distraction to getting actual work done.

As IT professionals, we can easily deride such ideas as fanciful daydreaming. After all, you can’t deliver the kinds of integrated information management tools that the business needs to survive without an infrastructure for them to run on. However, this is the backdrop against which business units everywhere are drawn to the promise of public cloud-based offerings that require next to no lead time or capital expenditure to implement. The hyperelastic public cloud would seem to be the silver bullet that offers the agility and scalability craved by business.

However, the public cloud is also afflicted with a wide variety of pitfalls and uncertainty. Whether it’s concerns over data security, regulation, and ownership or simply not wanting to be caller number 9,999 out of 10,000 when a widespread cloud outage occurs, putting all of your eggs in someone else’s basket is scary enough to prevent most businesses from dumping their own infrastructure and heading to the cloud at full speed.

It’s clear that business needs a way to deliver the same kinds of agility and scalability on its own terms and with resources it can directly control. That means consolidating server and storage resources to yield better utilization, trimming management overhead wherever possible through automation, and offering business units the ability to provision their own services. Therein lies the promise and challenge of the private cloud.

How now, network and storage Though it needn’t be an overnight transformation, the journey of implementing a private cloud will inevitably touch every level of your infrastructure. Aggressive virtualization and automation software are the two obvious components that most people think of when they hear the words “private cloud.” However, stopping there leaves out some of the most important components of the data center: the network and storage.

Anyone who has virtualized all or part of their data center is already familiar with the enormous benefits of doing so. You can consolidate your workloads onto general-purpose server hardware, abstract the management of operating systems from the hardware they run on, implement secure multitenancy, and manage your compute resources as a single, commoditized pool that can be scaled nondisruptively.

These benefits in concert with the management flexibility to reshuffle and balance workloads among different server resources — and even from one kind of storage to another — make an incredibly powerful combination. It’s no wonder that virtualization is so popular.

With converged networking hardware, those same virtualization concepts can be applied to the network. By combining all I/O onto the same high-bandwidth network fabric, you can achieve the same types of operational efficiencies. Whereas a traditional network would require two parallel networks, often with many low-bandwidth links, the converged network delivers a single unified system that can be managed as a commoditized resource in much the same way as virtual servers.

Innovations in storage technology pave the way to eerily similar gains. Instead of allocating separate storage resources for every server, virtualized storage arrays spread storage load throughout available hardware. Likewise, multitier storage can be managed autonomously by the storage hardware, with frequently used data retained in faster storage, while less-used data is gradually migrated to slower, less expensive storage, free of admin intervention. As with virtual servers and converged networks, the pooling of resources and management results in a commoditization of storage resources.

All for one, one for all Better still, each of these layers of infrastructure provide symbiotic benefits to one another. The high utilization density that can be achieved through server virtualization allows you to make better use of those high-bandwidth converged networking ports you’ve deployed. Some converged networking vendors also allow you to integrate the provisioning of networking resources with that of virtualized workloads — further increasing operational efficiency and capabilities.

Similarly, tighter integration between modern storage offerings and virtualization platforms brings performance and management gains to the virtualization layer. These can be delivered through support for hypervisor-specific features such as VMware’s vStorage API, which allows many storage-intensive tasks to be off-loaded directly onto the storage system or through hooks into the virtualization management tools. Conversely, the intelligence in modern storage systems can take advantage of virtualization-based thin provisioning, which allows the storage to free unused space and eliminate most storage overprovisioning — one of the largest sources of waste present in storage infrastructures.

While the guts of a private cloud revolve around management software and a healthy serving of virtualization, stopping there ignores a large part of the value proposition. The private cloud is really a unifying concept built on consolidation of resources and management that draws together advances being made in virtualization, converged networking, and intelligent storage — all of which can combine to form a whole greater than the sum of their parts.

This article, “Beyond virtualization in the private cloud,” originally appeared atInfoWorld.com. Read more of Matt Prigge’s Information Overload blog and follow the latest developments in storage at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.