|
I am often asked to help my clients build a business case for virtualisation. Implementation of a technology which fundamentally changes the way in which IT systems are deployed and managed is not a small consideration, one which can be slipped past the board in amongst the usual lifecycle replacement requests and manufacturer support renewals. It must therefore be presented as a positive step forward, where the reasonably significant initial outlay will be recouped many times - both through reductions in ongoing costs, and also through the major benefits it has to the business in other areas.
Busines Benefits of a Virtual Environment with EqualLogic iSCSI SAN
- Reduced IT Inventory- the total number of physical servers will of course be reduced through consolidation of resources - it is commonplace to achieve consolidation ratio's of around 8-1 in the real world - that's seven less servers to accomodate, power, cool and maintain for each physical host. What is often not considered is the supporting infrastructure for those servers - that's cabling, switch ports, UPS and all their associated costs. The larger the deployment, the greater the saving.
- Reduced Administrative Costs - It can be difficult to attach an administrative cost to each physical server, but it is easy to appreciate the benefits of a single management console, with many automated tasks which would otherwise tie up administrative resource. Deployment of new servers in a suitably specified resource pool becomes possible in a matter of minutes - companies are no longer at the mercy of long lead times and the dreaded Microsoft status bar.
- Improved Resource Utilisation- Historically the most practical method of segregation for incompatible applications was to locate them on their own physical server. This led to enormous "server sprawl" with many (most) servers significantly over specified and under utilised. There may be a small proportion of servers on your network which are working hard, but during capacity planning with clients we are surprised to find machines working at more than 10-20% of their available capacity. It is no longer necessary to physically separate these servers - segregation and isolation can be simply achieved in a virtual environment. One of the key criteria when designing a virtual environment is optimising resource utilisation so that each physical host is operating at 75-80% of available capacity, whilst sufficient headroom is maintained to accommodate peak demand, host failure and future growth.
- Unplanned Downtime and Data Loss Savings- Unplanned downtime can have an exceptionally high cost to business, particularly at peak operating times. VMware, coupled with an EqualLogic shared storage platform can eliminate unplanned downtime, and minimise data loss in the event of corruption or accidental deletion. Local continuity is enhanced by services such as VMware Vmotion, High Availability (HA) and Distributed Resource Scheduler (DRS. Data on the EqualLogic SAN is protected by previously unimaginable numbers of snapshots per volume/pool (512/10,000). Snapshots have negligible performance impact, so they can be taken throughout the working day. These snapshots can be accessed and/or replace the original volume almost instantaneously so Recovery Point Objectives (RPO) and Recovery Time Objectives (RTO) can be massively reduced when moving from a traditional tape backup model.
- Simple, Practical and Cost Effective Continuity and Disaster Recovery- VMware effectively encapsulates entire servers into a set of files which no longer rely upon a specific hardware platform. This means that each virtual server or "workload" becomes exceptionally portable, enormously enhancing both local continuity and disaster recovery. Where these files are stored on a shared storage platform, workloads can be simply transferred to an alternative host locally, and importantly replicated to a disaster recovery site. Replication is included as standard in the EqualLogic iSCSI SAN platform, so disaster recovery with EQL ad VMware can be as simple as purchasing a second array and some additional VMware host servers for the DR site. The same economies apply at the DR site so hardware, rack space and administrative cost reductions are replicated too. Fail over is straightforward, with systems operation in minutes or hours rather than days or weeks. Disaster recovery tests can be conducted quickly and regularly without interruption to production.
- Realistic Research, Test and Development - You're not going to get a much more realistic test environment than a clone of the production environment. This can be simply achieved and segregated using VMware and EqualLogic. Server/application patches and updates can be thoroughly tested on exact copies of prodution systems before deployment. Proof of concept projects can be carried out without the cost and inconvenience of hardware procurement, resources returned to the pool in a couple of clicks if you decide not to go ahead.
- Discrete Benefits - Virtualisation has much to offer but the discrete benefits are easy to overlook. For Example: How many businesses rely upon aincient, unsupported systems running on legacy hardware. How long would it take to recover these systems if an eight year old motherboard or RAID controller failed? Would they even work at all? When these systems are virtualised, they can be hosted on new hardware with rapid response hardware warranties. Transfer of these systems can be carefully managed using the myriad of P2V tools available. Clones and snapshots of working systems can be now retained indefinitely for deployment onsite or in DR should the unthinkable happen.
Virtualisation and the Dell EqualLogic product range will continue to develop at a rapid pace - indeed in a recent conversation with our internal Dell EqualLogic contacts, we were informed that EqualLogics original five year roadmap had been accelerated to a two year timescale following the Dell acquisition.
VMware also continue to add new features and functionality - such as Fault Tolerance released in 2009 - this list is going to continue to grow, and this article will be edited to reflect this.
|