In-Depth
Big Iron in a Class by Itself
What’s the most reliable server operating platform of them all? If you answered z/OS, you’re right. The rest of the tally might surprise you, however.
Mirror, mirror on the wall: what’s the most reliable server operating platform of them all?
If you answered z/OS, you’re right.
In a year-end survey of 18 server operating environments -- including Microsoft Corp.’s Windows and Apple Inc.’s Mac OS X -- IBM Corp.’s bread-and-butter mainframe operating system easily outdistanced the rest of the field.
Big Blue’s only competition, in fact, came from its own AIX platform, which topped the also-rans of the tally, ahead of competing entries from high-availability specialists Stratus Technologies Inc., Fujitsu, Hewlett-Packard Co. (HP), and Oracle Corp.’s Sun Solaris. The rundown of reliable server OSes surprisingly included Microsoft Corp.’s Windows Server 2008, which outpaced offerings from HP, Oracle Sun, and other prominent vendors.
Those are just a few nuggets from the latest installment of the “2010-2011 Global Server Hardware and OS Reliability Survey,” published late last year by Information Technology Intelligence Consulting (ITIC).
Perhaps the most interesting finding from the ITIC survey concerns the ascent of Microsoft’s Windows Server 2008 platform, which ITIC now places among the Top 3 mainstream OSes in terms of overall reliability.
“Microsoft’s Windows Server 2008 and Windows Server 2008 R2 served up the biggest surprise in the survey, scoring impressive reliability gains and making it one of the top three most reliable, mainstream server OSes,” writes industry veteran Laura DiDio, a principal with ITIC.
DiDio concedes that Windows Server 2008 made up ground very quickly. “Windows Server 2008 R2’s reliability renaissance is especially impressive since Microsoft’s Windows Server OS noticeably lagged behind the majority of the UNIX, Linux and Open Source distributions in ITIC’s 2008 and 2009 Server Reliability surveys,” she writes. “This was particularly evident when it came to chronicling the most severe Tier 3 outages which typically last for four or more hours, involve data loss and require multiple members of the IT department to perform remediation.”
In the survey, for example, Windows Server 2008 trailed AIX by only a single percentage point in ITIC’s most critical reliability metric.
It likewise finished ahead of several prominent RISC/EPIC-Unix platforms, including HP-UX and Oracle Sun Solaris. For example, 84 percent of AIX respondents reported experiencing less than one severe (Tier 3) incident per server last year. Similarly, just 83 percent of Windows Server 2008 R2 users said they’d experienced a Tier-3 outage (once again, on a per server/per annum basis). That was tied with Novell Inc.’s SuSE Linux Enterprise Server 11 and ahead of both HP-UX 11i (79 percent) and Sun Solaris 10. In the latter case, shops running Solaris-on-Intel reported surprisingly reported fewer Tier-3 outages (73 percent) than did shops running Solaris-on-SPARC (63 percent).
Big Iron Supremacy
All of these platforms have far to go if they’re to challenge Big Iron for reliability supremacy. “IBM’s z/OS … which is optimized for advanced real time capabilities and is in a class by itself, outpaced the field,” notes DiDio, who says that 90 percent of z/OS users didn’t experience Tier-3 outages of any kind.
What’s more, she continues, most (76 percent) Big Iron shops experienced less than five minutes of downtime per year -- which works out to 99.999 (or “five nines”) availability. This performance was matched only by dedicated fault-tolerant hardware, such as Stratus Technologies’ ftServer line, which likewise achieved 99.999 percent reliability over the survey period. Three-quarters of Stratus shops reported experiencing less than five minutes of downtime on a per-server, per-annum basis.
HP’s Integrity line -- which is based on assets it inherited from non-stop computing specialist Tandem Computer Corp. (which was itself acquired by the former Compaq Computer Corp. in 1997) -- likewise turned in a highly respectable showing, with nearly three-quarters (74 percent) of shops reporting less than 5 minutes of downtime -- per server, per year -- during the survey period. Fujitsu’s Primequest and Primergy turned in similar numbers.