Today at VMware Partner Exchange, EMC announced their new EVO:RAIL solution called VSPEX Blue. Great segue into an article about what the market looks like for convergence which happens to also be “Resolution 2.” As usual, we need a few key definitions first. Converged platforms combine multiple aspects of the traditional infrastructure stack into a single appliance or appliance-like solution that can be quickly deployed in the data center. Though a reference architecture can technically fit this definition, I am going to skip that topic here and focus on the more traditional converged and the newer hyper-converged platforms on the market.
Looking across the EMC portfolio we see a few converged platforms available. Moving across the spectrum, we start with VBlock. These are engineered solutions that combine Cisco UCS servers and Nexus switches, VMware hypervisors and virtual networking, and EMC storage and data protection. They are designed for larger deployments (think 100’s of VM’s) and are built to customer specs based on a thorough sizing process. They have a fairly steep point of entry from an investment standpoint, because they contain everything needed to host your infrastructure. Growing that investment means simply adding more VBlocks and moving forward.
Next in the environment comes the new VSPEX Blue platform. This is built on VMware’s EVO:RAIL software stack and built according to their specs. It’s a 2U appliance designed for smaller workloads (think dozens of VM’s) and comes in either a small or large configuration. It includes a 4 node server chassis in that 2U and is simply plugged into your 10gb top-of-rack switch and powered on. Here’s where the cool happens. 10 minutes later, given some IP address planning, you have a fully configure virtual infrastructure that is ready to start hosting VM’s. This system literally configures itself with a VMware hypervisor and management station, VSAN storage, and switching infrastructure. It has fault tolerance built in and ready to go. What else is cool, is that the EMC implementation of this product includes data protection via VDP which can be written to a Data Domain device, RecoverPoint for VM’s which allows replication of the infrastructure to a second appliance for disaster recovery, and built in hardware management and alerting with ESRS call home functionality to allow administrators to detect and remediate hardware failures. Again, all of this in a 2U chassis.
The EVO:RAIL infrastructure is rigid, and provides less flexibility for the more advanced IT organizations that may want to use existing hardware, or are looking for different use cases. That is where ScaleIO can come into play. Using DAS on your existing servers, ScaleIO can turn that abandoned storage into a robust and highly performant enterprise storage pool to present to your compute infrastructure. This allows your existing servers to become converged with the addition of some software.
Different flavors, different approaches, but a new way of thinking about how to get an application online and how to solve the burden of the IT administrator. The days of racking a server, installing and OS, connecting to your storage infrastructure, creating LUNS, and then presenting that environment to your application team are gone.