Posted by Richard Fichera on August 12, 2010
I’ve been getting a number of inquiries recently regarding benchmarking potential savings from consolidating multiple physical servers onto a smaller number of servers using VMs, usually VMware. The variations in the complexity of the existing versus new infrastructures, operating environments, and applications under consideration make it impossible to come up with consistent rules of thumb, and in most cases, also make it very difficult to predict with any accuracy what the final outcome will be absent a very tedious modeling exercise.
However, the major variables that influence the puzzle remain relatively constant, giving us the ability to at least set out a framework to help analyze potential consolidation projects. This list usually includes:
- Capital cost – Impact on capital cost will vary widely based on one major decision node – will you be buying additional servers and storage for this project. The outcome will also vary depending on the depreciation of the existing servers. For example, if you are consolidating onto new servers from an existing base of fully depreciated servers, then obviously your capital cost will rise. However, I think that capital cost is the least important of the potential effects of a consolidation program.
- Software licensing –Depending on exactly what components you have licensed from what vendors, and what terms you have, you can generally expect to spend between as little as zero or between $1,000 and $2,000 per 2-socket server for licenses. The choice will have a major impact on your software license costs and project economics.
- Server administration – Overall, your server admin costs should decline, since you will now have only 3 physical servers to administer, and the multiple VMs should be cheaper, by more than 50% (based on time and tasks) to support than a physical server image.
- Workload and VM management – The use of VMs should cut your costs for maintenance and ongoing workload management, recovery of failed systems, and capacity management by as much as 80% if you use even the basic ability to save and resume VMs. Use of a live migration facility will make operations even more efficient.
- OS provisioning and maintenance – The use of VMs will not make it any cheaper to provision and patch an OS image, although it will make it very much cheaper, by at least 80%, to generate a clone of an existing server image. Once the clone is generated, it will have to be patched as you would any other OS image, so you may want to look at how you deploy software if you plan to grow your server population quickly.
- Power consumption – Especially if you are purchasing new servers from one of the major vendors, you should be in a position to reduce the power consumption per server image by 70 – 90%, which can result in a significant offset to any increased capital amortization costs, depending on power cost.
Finally we come to the most important item on our list of variables, impact on operations. This is the BIG variable. Depending on how you change your IT processes to take advantage of the new capabilities you have acquired with your move to a consolidated and virtualized environment, you can effect major changes in operating costs at all levels, from provisioning to maintenance to business unit service requests.
Our net take – know your capital and other costs, but focus on reduced opex, particularly those enabled by the new technology and tools.
search forrester's blogs
The Age Of The Customer: Becoming Customer Obsessed
Dec. 10 at 2 p.m. (EST)
Register for the complimentary webinar »
2014 Tech Trends: Leverage Emerging Customer-Oriented Technology For Competitive Advantage
Dec. 10 at 8 a.m. (EST), 1 p.m. (GMT)
Register for the complimentary webinar »
- Amy DeMartine (1)
- Andre Kindness (20)
- Bryan Wang (11)
- Christian Kane (4)
- Christopher Voce (8)
- Courtney Bartlett (2)
- Dave Bartoletti (18)
- David Johnson (42)
- Doug Washburn (37)
- Eveline Oehrlich (9)
- Glenn O'Donnell (25)
- Henry Baltazar (5)
- Henry Dewing (5)
- James Staten (111)
- Jean-Pierre Garbani (12)
- John Rakowski (22)
- JP Gownder (70)
- Katyayan Gupta (16)
- Laura Koetzle (1)
- Lauren Nelson (7)
- Manish Bahl (1)
- Michele Pelino (4)
- Nupur Singh Andley (4)
- Rachel Dines (31)
- Richard Fichera (113)
- Sophia Vargas (2)
- Stephanie Balaouras (1)
- Sudhanshu Bhandari (2)