Drivers for server change
What do you mean, saving money is not at the top?
Workshop Over the last few weeks we have looked at many facets of the server environment. A major question faced in every organisation concerns compelling events that necessitate changes. What is driving “change” and do you consider the changes you make to be strategic or tactical?
As input to the report we’ll produce at the end of this workshop series, we recently asked you to what degree a number of technical IT issues are a challenge in your environments, and what business drivers are having an impact. Considering the technical isues first, a quick glance at the figure below shows the range of challenges is clearly quite broad, and looking at the data behind the chart we see that the overwhelming majority cite significant issues (level 5 or 4) in at least three areas.
Almost half said that operations staff are over-stretched and 40 per cent allude to challenges meeting business expectations for service availability. Issues around power, space and cooling are in there too along with cost / efficiency plus the old chestnuts of server management and monitoring, still a concern despite being areas in which considerable efforts have been expended over the years.
Some interesting results were highlighted when we flipped things around and asked what business drivers most effect on your server environments. The usual suspects of “the requirement to support new applications” and “changing requirements” for existing applications were high on the list - classed as high impact by over 40 per cent of organisations - but they did not top the list.
Security and compliance drivers were hovering around the 30 per cent mark with “improvements to service levels” just behind this. So what’s on top as the highest impact server environment changer? The answer is storage, and more specifically “the growth in data to be served”. That data volumes are increasing day by day goes almost without saying. But the fact that professionals answering a survey targeted at server estates made this the most widespread major business challenge very interesting. It illustrates just how much of an issue it has become to manage storage to meet business needs.
It should also be noted that “environmental drivers” are firmly established at the bottom of business drivers - a factor for only around one in five organisations. But perhaps the most fascinating return, especially in these challenging economic times is that the “need to reduce costs” was a high impact factor for around 25 per cent of those responding to the survey, placing quite far down the table. That said it is always going to be a factor considered in all solutions proposed, if not in and of itself the primary driver.
The business oriented change drivers are mostly concerned with meeting increasing demand. Whether it is the need to manage more data and therefore throughput, or to support new or changing application functionality, the big imperative is to allow more to be done. While service level improvements are also relatively low down the list (as most are delivering an adequate service), maintaining those levels is clearly going to be important.
So given this background, what are the main approaches you are considering in the evolution of your server estate? What role are you finding for the hot topic of server virtualization and where does storage virtualization fit in the grand scheme of things? We welcome your input in the comment section below.
We were able to save quite a bit by going with newer gen gear. Replacing something like 120 dual proc single core Xeons (2005 technology I think purchased in 2006 though), just started coming off of warranty/support early this year) With about 45 newer quad core boxes, and the newer boxes have plenty of headroom for growth. Replaced something like another ~50 servers from the same era with a 16-blade chassis running ESX. Pretty crazy going from something like 4 racks to about 1/4th of one rack.
Even more savings by going to a next generation storage system from a legacy system(about the same age as the servers). Cut power/space in half, and support by about 3/4ths. Not even taking into account all the things the new system can do that the old one could not.
Planning on retiring even more gear next year with another round of blades with more CPU cores and memory.
What do you mean, saving money is not at the top?
We mean that the cost "saving" of buying an expensive new "energy efficient" server as opposed to keeping the existing one running doesn't appear to exist in a lot of cases when you look at it with the degree of scepticism that people subjected to con artists/sales droids on a regular basis do.
Scrap your existing (in warranty) hardware and buy our new "energy efficient" server! (for £5k+) Its so much more efficient (per watt, but it draws more in absolute terms than our existing equipment) that it could save you thousands on electricity costs (over 25 years, which is 5 times more than the maximum warranty period)