With this current economic situation, already tight budgets are tightening more and IT managers are asked to work with a smaller IT budget. "Doing more with less" is not just a catch phrase these days. For many businesses, it's an exercise in survival.
Since taking care of business remains the top priority, cutting critical services isn't an option, and unscrewing half of the light bulbs in the office isn't practical. So the question remains: Where can you cut back?
When I considered how I would answer this, I came up with ways to save money in the short term and the long term. I also reminded myself that sometimes you have to spend money to save money. To that end, let's look at what and where IT can scale back to save.
Server reduction via virtualization
If you haven't considered server reduction for your organization lately, perhaps it's time to do that now. Virtualizing servers requires more robust machines (processors, memory, disks), but since most application servers are only in use 30% of the time, it's easy to justify virtualizing -- especially if it's time to upgrade.
Virtualizing servers can save money in various ways:
- Licensing: Windows Server 2003 Enterprise gives you four free server licenses for virtual machines, and for Data Center, unlimited licenses are provided.
- Support and maintenance costs: Fewer machines require fewer support contracts, fewer parts to replace and so on. Again, four virtual machines on a single physical host require only one maintenance contract.
- If servers are due for an upgrade, virtualization will reduce the additional investment.
Note: There are some downsides to virtualization. Microsoft and other vendor support can be an issue. Microsoft likes to play the "we don't support VMware" card from time to time, but often the company will support it. And while the company does support its own virtualization products, it may force you to reproduce it on a physical box.
Support contracts and service-level agreements: What's the cost?
Can you afford your service-level agreement? I've heard of one company whose email recovery SLA forced it into a backup/storage configuration that was not a standard. This caused many technical problems because the configuration was so unique. It's best to be in the 90% bracket rather in than the peculiar 10% grouping. For email, consider the initial SLA to get email running again, and then have a longer time to restore the mailboxes. And, if there are critical users whose mailbox restore is vital for business, then put them in a separate database and restore them quickly – the others can wait. The restore will be faster and you'll save money because you don't have to invest in expensive storage solutions to get the recovery speed. Another strategy is to use disk to disk as a front line restore and then use tape after that.
Does your software support contract satisfy business needs?
When was the last time you looked at your software support contract costs and compared them to your needs? To determine what is truly needed and what can go, you must ask, analyze and get answers.
- Ask if you used all of it. Whether it is incident-based or hourly-based, make sure you used all of your software support. If not, try to reduce it in the next contract.
- Analyze what it was spent for – upgrading could save money.
- Do certain servers require a lot of support? Try upgrading or replacing hardware and/or software.
- Are certain applications problematic?
- If these are legacy apps, are they worth it? These are expensive apps to maintain, and this might be a good excuse for change.
- Can you upgrade faster? Migrations are costly because they consume many resources. Perhaps dedicate more resources or move purchasing up for faster migration. If the business would benefit and people would become more productive, this is an attractive option.
- Are your legacy Windows OSes worth it? You'd be surprised to learn how many Windows NT environments still exist and how many shops still run Windows 2000. They may have a valid reason for it, but Microsoft support contracts are very expensive. And troubleshooting legacy OS problems uses more resources, i.e., time, since support engineers have to use inadequate legacy tools and must try to troubleshoot problems they are not familiar with.
- Get a breakdown from your provider on how many incidents were used, time per incident and so on.
- If the time per incident is high, go to a per-incident contract if the price is right.
- If time per incident is low, going to an hourly contract may be cheaper.
- Most organizations require you to define a very narrow problem as one incident and open another incident if the problem expands. Paying an hourly rate eliminates that, but it could be more expensive depending on how they charge.
Invest in staff training
How about training your staff to solve those cases? Knowledgeable staff members can solve problems faster than phone support because they know the environment and they have a vested interest in problem resolution. It is difficult to find troubleshooting courses, but there are a lot of Internet resources that can help.
Microsoft has some troubleshooting courses, but you can solve a lot of problems using Google searches, www.eventid.net, www.experts-exchange.com, Microsoft TechNet blogs and other forums. Simply forming the right query will bring answers. Also, hire people who have experience. Hiring high school students because they passed a Microsoft certification exam won't help when it comes time to troubleshoot.
These are just a few ways to create a cost management plan so you can do more with a smaller IT budget. These pointers may inspire you to be creative and find other cost savings.
Gary Olsen is a systems software engineer for Hewlett-Packard in Global Solutions Engineering. He authored Windows 2000: Active Directory Design and Deployment and co-authored Windows Server 2003 on HP ProLiant Servers. Gary is a Microsoft MVP for Directory Services and formerly for Windows File Systems.
