Some (vague) analyst advice for an OS switch

If you are thinking about switching operating system platforms but are not sure about the best course of action don't count on the analyst community for a clear-cut answer. Sure, where they lack in specifics, they make up for with decision approaches and criteria lists that can factor in things you may overlook.

score

If you are thinking about switching operating system platforms but are not sure about the best course of action don't count on the analyst community for a clear-cut answer. Sure, where they lack in specifics, they make up for with decision approaches and criteria lists that can factor in things you may overlook. But in most cases, you'll hear something like this: "Only after a business conducts a thorough cost/performance/risk analysis should it decide whether to remain with Windows, do a wholesale switch to Linux or install multiple operating systems." This quote is from a recent Yankee Group report that shows SMBs how to choose between Linux and Microsoft. Sounds great, but sounds stale. Fortunately the researcher has something more provocative to say:

There is no such corresponding, clear-cut evidence that indicates a switch from Windows to Linux will significantly reduce TCO or substantially increase performance, reliability and uptime, particularly among SMBs with significant budget constraints. In fact, March 2005 independent Yankee Group survey data indicates the performance, reliability and overall TCO of Windows improved dramatically in Windows Server 2003. The survey, which polled 500 North American respondents, showed nearly 50% found Windows Server 2003 to be 1.5 to 3 times better than comparable Linux distributions in performance and reliability.

But wait a second,

Anecdotal case study data also showed that for the roughly 20% of SMB companies with a high degree of technical proficiency, Linux and open source distributions provide stellar TCO and a near immediate ROI.

Ok, so say you are a tech savvy firm and ready to get on board with Linux. The next step is to pick a vendor, which is typically going to be Novell/SUSE or Red Hat. Gartner tackles this one with a set of selection criteria. Here is a taste:

Red Hat's strategy is to build pure open-source software (OSS) stacks, but with additional subscription services for clustering, high availability, cluster file systems, identity management and authentication (Netscape), and application servers (JOnAS). In contrast, Novell/SUSE Linuxs preference is a hybrid of OSS and proprietary solutions, including YaST (offered as OSS), ZENworks (desktop management), Polyserve (proprietary, non-open-source software for clustering and high availability), identity management (non-OSS from NetWare), eDirectory (non-open-source, but also support for OpenLDAP as OSS), iFolder (part OSS) and JBoss (OSS and supported by Novell). Both vendors will provide full solutions based on combinations of open-source and non-open-source products. In Red Hats case, all the non-open-source components are delivered by partners (which may also compete with Red Hats open-source alternatives); in Novells case, some are from partners and some are from Novell. The crossover point from open source to non-open-source when moving up the stack will differ between the two vendors.

Unfortunately, the distinction between these two is nothing like that of Linux andWindows so Gartner shouldn't even try and list any best use scenarios. So you'll need to look at smaller differentiating factors (applications, partner relationships, quality-of-service) to come up with a decision. And when you do, don't forget a cost/performance/risk analysis.




Show Comments