Times are tough, everyone is asked to reduce spending. Some IT departments are seen as an expendable cost center - if they are not saving money, they are not doing their jobs. I think this can be a healthy attitude to keep those sometimes-unaccountable IT departments honest and focused on the corporate business. I have a few high-tech ideas for cost savings that can yield real results. If I were the CIO of your company, this is what I would do to reduce costs in the data center and desktops:
1. Use commodity server hardware - do you really need that proprietary big iron Unix server? Wouldn't a Linux blade work instead? Intel x86 architecture can run faster and cheaper than that proprietary hardware, and you'll be surprised at the difference - you might get a 3X performance boost and cut your costs 75% at the same time! Unbelievable? I've seen it myself.
2. Use a commodity server OS - do you need Unix? Linux would work for 70% savings. If you are not truly married to Windows Server, switch that to Linux too, you're operations people will love it.
3. Use an Open Source app server - use JBoss, its better and cheaper than Oracle or IBM app servers. If you can re-write your Microsoft apps, do that too. Or just halt all future new Microsoft software projects in favor of Java / JBoss apps.
4. Use Open Source database servers - Postgres and MySQL have evolved in the last 5 years to be true powerhouse database engines. If you are using Oracle or IBM, you could save millions by switching your old apps and using Open Source databases for new apps. You may have a master license agreement that gives you "unlimited copies" of Oracle, well... how generous of you to line the pockets of the Oracle salesman. You'll find out that MySQL and Postgres are much easier to maintain too, so you will be able to reduce DBA costs at the same time. Want to get world-class support? Pay Sun for MySQL support or EnterpriseDB for Postgres support.
5. Outsource your email servers to Software as a Service (SaaS)- stop paying for Microsoft Exchange or IBM Notes, use Google Apps. If you don't like Google there are hundreds of alternatives, all with professional grade email and your customers and colleagues will never notice the change when they get emails from you. Pull the plug on your in-house email servers and reuse the hardware or turn them off to save electricity.
6. Outsource your whole data center - do you truly need to worry about having 30 servers in a concrete bunker with UPS, disaster recovery, HVAC, fast internet lines, and monitoring? Hire a hosting company that can do that better and cheaper than you ever could. You can host your test servers, your database servers and your internal web apps. Users won't even know that their apps are really running in a separate data center. You can get dedicated server hardware and fast links to their data center if you need it. Or save a huge amount on VPS services.
7. Use Open Source Office apps on the desktop - OpenOffice.org has truly arrived, with version 3.0 just released, it can read the new docx formats that Microsoft Office 2007 uses. Admittedly it does not behave exactly the same, but I used OpenOffice almost exclusively for Documents, Spreadsheets, and Presentations. I can easily send them to my colleagues who use Microsoft Office without compatibility problems. For email clients you can use Thunderbird to replace Outlook. Google Docs is making some waves but it is still immature and buggy. I look forward to seeing Google improve these apps enough to compete.
8. Use an Open Source OS on the desktop - Ubuntu and Fedora have really come far in the past few releases to offer truly solid Linux desktop experiences. Most Windows users will have no problem with the new UI. This may seem like a risky move, but if you are serious about cost-cutting, I think you'll find that Windows is a commodity in your office already. Now that email and office apps will run on Linux desktops, is Windows a requirement for your business? You may hit some web sites or Word documents that need Microsoft technologies, but I bet that's more rare than you think.
These are concrete, no-fluff ideas that any IT department can use. Obviously smaller, more agile IT departments and companies will find these changes easier to swallow. And some of these ideas might seem pretty risky, but if you want to keep risk down then try things out on new projects first. Later, switching the older systems to complete the cost savings. Its OK to mix Linux and Unix machines in the same data center, and there's no need to switch everything over night. You can get savings project by project to keep some sanity in the environment.
Making these changes won't be easy, many will resist these changes. These excuses are common:
* We already know how these old systems work, lets just buy more and keep it safe
* Open Source is risky, who can we blame for problems, what if we get sued?
* Nobody ever got fired for buying IBM (/Oracle)!
All of these are weak arguments to prevent the hassles involved with any change. Spending new money on old expensive solutions is wrong, and you can pay for professional support from RedHat, Sun, Canonical, and still come out way ahead on costs. At IBM and Oracle's prices, you should be fired for choosing them when lower cost competitors that would fit your business even better.
Cost saving pitfalls. I do not advocate cost-cutting in non-commodity areas of your business or IT. For example, Offshoring software development or customer service can have bad consequences. Quality varies wildly at home and abroad for software developers and customer service experts. Offshoring those essential skills just multiplies the risk by 6 or 10 time zones. I also would not expect the Sales department to outsource the sales department - company image and product knowledge are too important to the business to trust to others who are not invested in your future.
History shows us the way. I remember the switch from mainframes to Unix and Windows in the 90's economy. It did not happen overnight, and we heard plenty of grumblings from the luddites who resist all changes. It started with commodity apps like email and word processing (anyone remember the mainframe spell checker? yechkkk!), and one business application at a time to reduce the risk. We mixed mainframes and server apps for a decade, and we realized that we could reduce costs pretty quickly. As a side-effect we got more choice and better apps.
The same thing is happening in Open Source and the Software as a Service (SaaS) industry. Lower costs, more choices, better products and services. And it starts with commodities like operating systems, database engines, app servers, and apps like email and word processing. I use all of these technologies to save costs at my company.
Don't cling to expensive Microsoft, Oracle, and IBM products that haven't changed much in 10 years. Or you'll suffer the same fate as those mainframe-clingers in the 90's. Instead, use the latest technologies to save money and help your business in these trying times.
- Jay Meyer
2 comments:
Exactly what I'm thinking :) I know from inside how hard it is to keep these big ol' platforms progressive ...too many involved persons and too many non-IT folks around the decision making boards
Here's an eWeek blog that shares some of my suggestions:
http://www.eweek.com/c/a/Linux-and-Open-Source/Hot-Open-Source-Projects-for-Enterprise-Use/
Post a Comment