Hello, this is the first blog post in what I hope will be a large catalog of future brain dumps, how-to’s, and lively debates. MyPuresupport began in 2012 as medium of thoughtful exchange for users who were reviewing or had just purchased an IBM Puresystem/Flexsystem/Pureflex system aka Bladecenter 2.0. For those who played around with the Puresystem it could be argued that this was IBMs first attempted at a “Converged” Infrastructure combining Compute, network, and storage into the same 14U footprint. They also had a management node called the “Flex System Manager” which attempted to provide the user with additional tooling to provide faster deployment of servers, management components such as an “HMC-like” interface for anyone who used Power servers in their Flex system.
Why is this important? This began my long journey to discovering the future of Hyper-Converged infrastructure with…..NUTANIX.
When we talk about Nutanix we are talking about a new era that is inevitable. Think about the advancements made to cpu/gpu technology and increasingly dense memory in each server. Add that to ever denser flash memory and the future of flash in nvme and 3d cross point and we now have the ability to bring the storage controller up into the server and collapse the 3 tier stack. Legacy storage systems filled an important need when Systems required high iops and throughput by using raid arrays and striping across as many spindles as possible for maximum performance. Now with flash modules allowing for 20k iops a piece and getting up to larger capacities it no longer makes sense to restrict servers by limiting them to a single set of storage controllers when you can have each server have a dedicated controller vm while each controller vm also works as a unit for the virtual cluster. In addition to the performance benefits there are also huge gains in management overhead by at least 66% by removing two thirds of the three tier stack. Now there’s no need to manage firmware upgrades and checking interoperability between host systems and the storage array. Let alone actually performing the upgrade and crossing your fingers you won’t have issues while the storage unit goes through a degraded state rolling upgrade. The reduction of infrastructure management and money by collapsing into a 1 tier stack result in significant Datacenter savings. Nutanix is the first in the industry to provide a truly “webscale” architecture relieving the customer of the crippling limitations like cluster size and “piecemeal” components by using MapReduce framework for cluster job management and an “all-in-one” HTML5 dashboard that includes all of necessary features any administrator needs to manage their entire infrastructure from a truly single pane of glass.
What does this mean for MyPureSupport? As I tried to stay with the current trends of converged infrastructure through Puresystems I will now be blogging about anything and everything Nutanix and/or about the current trends in the HCI marketplace.