We are excited at the prospects of 2012 being a year where we are able to bring a renewed value to IT. Â The technologies upon us finally are making IT more accountable for more efficient use of resources. The world of virtual machines is really the most exciting of technologies as this has the greatest cost to benefit equation, with little if any additional risk. In fact, it can add more redundancy.
Make no mistake, it’s truly prime time here. We’re actually having fun moving numerous old physical machines to a single physical machine partitioned with VMWare. Look, virtual machines are not necessarily for every application, but the fact of the matter is, an average server only utilizes 15% of its CPU. Okay, that’s like an eight-hour-per-day employee working just over an hour a day. You wouldn’t put up with this in the world of people, why would you with the technology? This makes little sense. But exactly where are the savings you say? Ah, I was hoping you would ask!
Savings is where this technology excels. It’s about economy of scale as much as economy of smart. So where exactly would you save the money? Here’s the short list:
- SERVERS. Buying less hardware is just plain smart. Less is the new more. So you have three or four aging servers. Say they are $8k a piece to replace including a 5-year maintenance contract, which is $2k of it. Buy one slightly beefier box for $14k including the $2k maintenance. Using three-server-to-one consolidation model, your savings on hardware alone here is $10k! If you did four-to-one, assuming you kicked up the specs a notch and spend another $1k for another processor and more RAM bring it up to $15k total, your savings is $16k!
- ELECTRICITY. If you have one machine instead of four, all things being equal, you would only use one forth the energy in powering the machine. Additionally, the larger savings may be in the COOLING aspect, as it’s one fourth the BTU’s of generated heat.
- REDUNDANCY. This always an important, no critical, component when designing infrastructure. The “What if’s” and “Plan B Scenarios ” need to be covered out of the gate. So how does VMWare stack up? Welcome to RAID 6, that is RAID 5 with a second parity stripe. Read performance is unaffected and write takes a wee bit longer, but the plus side outweighs the overhead. You can lose not one disk, but two and still be running. This extra level of redundancy is important and should always be couple with redundant power supplies, one going to a reliable UPS unit and the other going to the local power grid.
- MANAGING. Yet another area where this technology excels. There is little easier than managing the pool of resources. You give the resources to VMWare and it manages it all. I suppose it might be easier, we just haven’t figured out how that would even be possible. It’s all easy and on the fly with no reboots. Meet the nirvana of management, VMWare.
- DOWNTIME during MIGRATION. This has historically been both a painful and costly thing. But if you’re simply moving an old server to new hardware, you do it all without incurring downtime and switch over when ready. Likely your only change will be the static IPV4 address on the NICs as it will grab from the DHCP pool during the migration. But remember also, that this is not for every machine, but we can help you figure out what makes the most sense here.
In summary, this technology makes sense for most servers, except heavily utilized database servers and other machines that truly require dedicated hardware, but in most cases, pooled resources are more than adequate if the virtual machine host is properly-sized.
Also, this video (hyperlink to below) from VMWare may help you better understand as a picture and a movie are easily worth a thousand words! http://www.vmware.com/virtualization/
Many to one, now we’re talking!