- Aging Legacy SAN
- High ongoing support costs
- Lack of IOPs
- Requirement to maintain and manage expensive Fibre Channel Fabric
- VMware vSAN (Software Defined SAN)
- HPE DL380’s with both flash and spinning disks
As a managed service provider, and owner of multiple Data Centres, this technology dependent organisation approached Port-P to assist with upcoming technology refreshes. Furthermore new client wins was adding further strain on infrastructure already nearing its performance and capacity limits.
At the centre of this and most critical to the customer was the need for high performance, reliable and scalable storage. IBM’s v7000 was the existing storage platform and was near the end of its initial supported life cycle and at an estimated 250k IOPs was feeling the strain from the increasing client demands placed upon it.
Working closely with the customer to identify the performance characteristics, as well as scalability required from any new server and storage platform, Port-P agreed that a Hyper-Converged approach, utilising VMware’s vSAN product would be an ideal platform to build upon, both for their own internal requirements as well as those required to form their own hosted and managed service offerings.
As an initial internal use pilot, Port-P designed an underlying hardware platform based on HPE’s tried and trusted DL380 servers. To provide the required performance, capacity and resilience, a mix of SSD’s and traditional spinning disks, spread across multiple 12GB SAS HBA’s as approved on VMware’s Hardware Compatibility List was proposed and deployed.
The results were outstanding; using Iometer the customer measured a peak performance of 1.6m IOPs across the new cluster of servers, at the fraction of the cost of achieving those figures from a dedicated traditional array.
Results and Conclusion
This organisation was at a critical point in their successful growth as a business. Though 21 years of experience and a thorough appreciation of all industry offerings Port-P were able to work collaboratively with them in the design of a high performance, scalable and future proofed hyper-converged solution, while reducing Capex spend by several hundred thousand pounds against the purchase of an equally performing physical external array coupled with new servers.
Furthermore ongoing OPEX savings were achieved through the reduction in physical rack footprint and the associated power and cooling overheads.
So successful was the initial pilot, it became a template ‘off the shelf’ solution that can be ordered on a consistent and repeatable basis as they acquire and adopt new managed service clients.