 |
 |
 |
 |
 |
|
 |
 |
|
 |
 |
|
 |
Page II: When is the right time to consolidate your IT resources...or is there ever a good time?
And in general...
The low hanging fruit for consolidation are file, print, e-mail, and database servers, says Favetti. They provide improved performance and functionality, and once in place IT staff can explore the potential within the organisation for new features.
But picking the low hanging fruit is a tactical approach, warns Albe. "It's only going to buy you a little time--the effectiveness isn't going to be there." Instead, he counsels a more strategic approach, such as examining existing direct-attached storage and looking for economies of scale that can be achieved through centralisation. If you can get the required performance levels, the economies are fantastic, he says, but it's important to understand who owns the data and you ensure its availability meets their needs.
"Assumption is a very risky thing," he warns, so engage the right stakeholders and put a good project manager on the job. Being insufficiently farsighted can cause problems--if consolidation is done simply as a cost cutting measure (eg, by reducing headcount), you might overlook internal plans for mergers and acquisitions, changes to disaster recovery capability, or new applications. Any of these can turn IT staffing costs into secondary considerations.
This sentiment is echoed by Favetti. "Consolidation is not just a technical thing, it's a business thing as well," he says, so executive buy-in is important. "It's a question of ownership," he suggests, and organisations should consider the business and financial perspectives, risks and benefits.
An end-to-end view is necessary for successful consolidation, suggests Paul Muller, head of HP software marketing, strategy, and alliances, so it is important to take the time to understand your environment at the outset. If you don't understand the chain of applications, servers, storage, networks and so on, you may not be able to predict the effects of any individual changes you make. It's better to plan than react, he says.
"Good management is the core underlying driver for consolidation," says Muller. While cost reduction (or at least cost control) is important, organisations are not necessarily looking for the lowest cost but for the right cost. He draws a parallel with the airline industry, where success is possible with either low cost and low service or high cost and high service.
Albe suggests a SWOT (strengths, weaknesses, opportunities, threats) analysis before consolidation, looking at data and its attributes today and into the future.
"Rather than 'do more with less', 'do the right thing' and plan first," he says. Although people are looking at cost savings, they might get them more quickly by avoiding tactical decisions and instead doing the right thing at the right time, using the right people.
Christian agrees, suggesting consolidation should be regarded as a strategic decision, made with a two-to-five year perspective.
Kevin McIsaac, research director Asia Pacific, META Group recommends a stepwise approach to consolidation in the data centre.
- How many different locations do you use? Can you consolidate to a smaller number? This will allow centralisation of staff and economies of scale, but depends on the networking capabilities of the software, the cost of bandwidth, and internal politics. "It's not a huge return, but it's one most people start with," he says.
- Storage consolidation. SANs are expensive, but there are certainly advantages to external storage, including quicker recovery of valuable data, an improvement in utilisation from between 40 and 60 percent to 80 percent or so, and modest hardware and operational cost savings.
- Stop and think. Putting everything onto one big box doesn't make sense. The cost of servers with more than four CPUs is between two and three times the cost of the same amount of processing power purchased as one to four-way servers. Also, these extra capital costs are certain while any operational cost savings are only anticipated. One 16-way server might be cheaper than 30 single CPU boxes and it can be capable of delivering the same workload, but it's not easy to run a server at 60-80 percent utilisation, so more sophisticated management tools are needed.
6%
3%
