It’s finally here! I’m glad to be publishing this report during the first day of HP Discover in Barcelona. This is a “down to earth report”, focusing more on value rather than the tech side and its benefits.

I think it best to introduce this paper by starting off with the executive summary that can be found at the beginning of the report:

“End users of all kinds are continuously looking at solutions aimed to reducing the complexity and to drive down the total cost of ownership (TCO) of their IT infrastructures. Especially in the SMB space, where resources are usually limited and skills aren’t often developed at best, there is a lot of attention to this issue.

Nowadays most infrastructures are virtualized, and they are based on x86 commodity servers with a standard ethernet networking. The highest costs, both in terms of TCA (Total Cost of Acquisition) and TCO (Total Cost of Ownership), usually come from data storage. In fact, it’s easy to find that about 50% of the cost of a virtualized infrastructure relies on an ordinary SAN/NAS infrastructure.

This paper will show a solution, completely designed on industry standard components and HP software, that allows the realization of a complete scale-out virtualized infrastructure without needing traditional shared storage. In a few words: a hyper-converged infrastructure. The converged solution that we are going to describe has all the characteristics that you could expect to address: TCO, ease of use, scalability, efficiency, it’s also open and with a low price. In fact, the big advantage comes from the virtualization of the internal disks on the same x86 servers that run the hypervisor, presenting a virtual SAN to the VMs.

Another big advantage is that the whole stack comes from a single vendor, enabling an easier way to get access to an end-to-end support service.

The scale-out model presented in the following pages could have a great impact to the user and, potentially, for the reseller of the solution. In fact, each node (a brick) has a precise amount of resources (CPU, RAM, disk space/IOPS) and a well-known acquisition cost too. Adding a precise amount of VMs to the infrastructure is only a matter of buying one or more bricks and connecting them to the existing infrastructure.

The primary objectives of this paper are to present a reference design, its potential scalability and a standard benchmark score showing its positioning.”

You can download the full paper here.

Comments are welcome!

Disclaimer: HP is a client of Juku Consulting srl