Post 4 of 6: xBlock- Designing a Repeatable Architecture
I have lost count of how many pre-integrated infrastructure offerings are now available through the various software, network, storage, and compute partnerships out there – all competing for their stake as your cloud’s foundation (for those who still think ‘the cloud’ is a hardware offering with some antiquated management slapped on top, please reexamine the definition of cloud). A well-designed infrastructure is a great (and required) start, but software is where the magic happens…but I digress. Whether it’s a vBlock (EMC, Cisco, VMware), FlexPod (NetApp, Cisco VMware), Dell’s vStart, or any of the dozens of combinations out there worth mentioning, one thing holds true…all (well, most) are examples of what a robust reference architecture can do for your cloud. Sure each offering will add it’s special sauce as a value proposition – proven, pre-engineered, specialized, best-of-breed, etc. – but, more importantly, these solutions provide the core components for your cloud’s infrastructure – Storage, Compute, Network, and the glorious Hypervisor – and use the Lego approach for scaling the infrastructure out as needed.
Scale Happens. Cloud infrastructures are meant to be elastic, agile, and designed to scale beyond your wildest imagination. [insert joke about the company that markets the “cloud in a box”]. Hint: rhymes with “poracle”. Using the right tools and technologies, the idea is to build small and scale big – to cloud scale – as the demand for resources grow and without the need to reengineer your infrastructure at each growth iteration.
Block-based architectures give IT departments the ability to seamlessly add the needed resources to the cloud without adding unnecessary complexity or spending more time than necessary on hardware integration. This is a tough task considering the latest and greatest chipset will become yesterday’s technology by the time I publish this post. Fortunately, we can plan ahead to avoid headaches in the future by utilizing vSphere’s ability to turn hardware into a huge aggregate of [compatible] resources and a cozy layer of abstraction between hardware and your workloads. Plan ahead and enable vSphere’s built-in Enhanced vMotion Compatibility technology to allow for smoother inter-cluster scaling. Or, better yet, grow your environment one full block at a time – by simply adding another cluster of compute, network, and storage resources to your cloud infrastructure.
Unless you know exactly what your workloads look like today and have planned for (and prematurely invested in) all the possible growth for the next several years, you will need to scale. The idea of block-based architectures is simply add another block and “bam!” you’re in business. Now you just need your software to scale along with the hardware seamlessly with little or no interruption. With vSphere virtualizing your back-end, that’s one less thing you have to worry about. Whether for green-field or growth, block-based architectures can help you reach the levels of scale and efficiency that were once inconceivable.
Take a look at VMware’s “Architecting a vCloud” white paper for some more ideas on how the block approach compliments cloud architectures.
In this Series:
4 – xBlock -designing a repeatable architecture
6 – Get a Grip! – managing your cloud