ITIC » Application Availability

IBM, Stratus, HP, Fujitsu Top ITIC/GFI Software Hardware Reliability Survey

Laura DiDio — Thu, 27 Jan 2011 15:02:38 +0000

For the third year in a row, IBM AIX Unix operating system (OS) running on the company’s Power System servers scored the highest reliability ratings among 19 different server OS platforms – including other Unix variants, Microsoft’s Windows Server, Linux distributions and Apple’s Mac OS X.
Over three-quarters or 78 percent of survey respondents indicated they experienced less than one of the most common, minor Tier 1 incidents per server, per annum on IBM’s AIX v. 5.3 and AIX v 7.1 distributions
Those are the results of the ITIC 2010-2011 Global Server Hardware and OS Reliability Survey. ITIC partnered with GFI Software (formerly Sunbelt Software) to conduct this independent Web-based survey. It polled C-level executives and IT managers at 468 corporations from 23 countries worldwide from November through January.
The survey data indicated that the reliability and uptime of all the major server OS and server hardware distributions has improved significantly over the past several years.
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. 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 the ITIC/Sunbelt 2008 and 2009 Server Reliability surveys. 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.
An overwhelming 92% majority of Windows Server 2008 R2 users experienced less than one or one Tier 3 outage per server, per annum followed closely by the 90% of respondents using IBM’s AIX 7.1 who said they experienced one or less than one severe Tier 3 incident, per server per annum. Some 86% of Novell SuSE Linux Enterprise Server 11 and 84% of HP UX 11i v3 users also testified to the reliability of those platforms, reporting that they experienced either one or less than one unplanned Tier 3 outage per server, annually.
The survey found that all server OSes continue to make year-over-year reliability gains. The essay comments and first person customer interviews revealed that the majority of the moderate and severe Tier 2 and Tier 3 outages were attributable to integration and interoperability issues such as incompatible drivers, trouble applying patches, (particularly in highly customized environments), misconfigurations and the lack of a specific component or software fix for a particular platform.
Some IT managers also acknowledged that complexity and the IT department’s unfamiliarity with new products, software versions and new technologies like virtualization and private clouds prolonged downtime. This is particularly true in instances where corporations lacked the time or the funds to certify and re-train the appropriate members of the IT staff on new technologies.
The Sun Solaris 10 now owned by Oracle had respectable reliability statistics, though the Solaris on SPARC systems lagged behind most other OS distributions. Nearly 73 percent of respondents reported that Sun Solaris 10 recorded less than one Tier 1 per server, per annum outage, while only 63 percent of Sun Solaris 10 SPARC users achieved those same reliability results. The numbers were similar for the more moderately serious Tier 2 outages with 70 percent of users running Sun Solaris 10 on SPARC systems reporting less than one incident per server, per year. Sun Solaris 10 on x86 systems fared slightly better with 71 percent recording less than one Tier 2 incident per server on an annual basis. With respect to the most severe Tier 3 outages, 70 percent of Sun Solaris 10 on SPARC survey participants say they experienced less than one incident on each server during the year, compared with 74 percent of Sun Solaris 10 running on x86 platforms who reported less than one severe Tier 3 incident per server, per annum.
Overall, with respect to the most severe and prolonged unplanned Tier 3 outages, Sun Solaris 10 also lagged behind all of the major OS distributions with 70 percent of customers reporting less than one outage. That is the approximately the same percentage of organizations that are still using the eight year-old Windows Server 2003 server operating system. Some 69 percent of Windows Server 2003 users reported less than one per server, per annum Tier 3 outage.
IBM Tops in Server Hardware Reliability
IBM hardware was also best in class in terms of reliability, stability and performance. IBM’s System z mainframes recorded the least amount of downtime; 76% indicated System z machines experienced just one-to-five minutes of unplanned outages per server, per year, the equivalent of 99.999% or better availability.
Stratus Technologies’ ftServer 6300 and 4500 series and Fujitsu’s Primequest and Primergy Servers also made impressive showings. Some 75% percent of Stratus ftServer 6300 and 4500 users say they experienced one-to-five minutes of per server, per annum downtime, for five nines of availability. Some 74% of HP’s Integrity and Fujitsu Primequest and Primergy server said they experienced less five minutes or less of unplanned annual server downtime.
Among the other survey highlights:
• A 57% majority of respondents said their server hardware is between one and three years old. One-in-five corporations – 20% – said their servers were three-to-four years old.
• One-quarter – 25% — of businesses refresh their main line of business server hardware “as needed” and 10% said they upgrade a portion of their servers annually.
• Only a very small 2% minority of organizations aggressively upgrade their servers every two years. The majority of companies are on a three, four or five year server refresh cycle with 15% of participants stating they upgrade servers every two years; 15% upgrade every three years and 17% are on a protracted five or six year server upgrade cycle. Another 15% said they have “no specific” server upgrade timetable.
• A higher percentage of users prefer to apply patches manually rather than automatically. Nearly three out-of-10 organizations – 30 percent say they opt to apply patches manually, all or most of the time. Another 35 percent of survey participants say they “sometimes” apply patches manually. Only 16 percent of respondents never apply patches manually.
• Some 26 percent of respondents who always use group policy to apply patches and 16 percent who sometimes utilize group policy methods compared to 52 percent of survey respondents who eschew group policy.
• The manual patch method does take longer than applying patches automatically or using group policies. Overall 61 percent of those polled said they spend more than one hour applying patches to their server platforms for each specific upgrade . Of that figure, just under half – 29 percent – revealed that it takes them in excess of four hours to apply patches for each incident.

The length and severity of Tier 1, Tier 2 and Tier 3 unplanned outages and the patching actions related to each correspond to specific line item capital expenditure (CAPEX) and operational expenditure (OPEX) costs for the business. Reliability, measured by downtime, can positively or negatively impact TCO and accelerate or delay the time it takes to realize ROI.
Improvements or declines in reliability also mitigate or increase technical and business risks to the organization’s end users and its external customers. The ability to meet service-level agreements (SLAs) hinges on server reliability, uptime and manageability. These are key indicators that enable organizations to determine which server operating system platform or combination thereof is most suitable.
Overall, these survey responses provide crucial, comparative reliability metrics to enable customers to make informed choices on which server hardware and server operating system or combination thereof, best suits their specific business and budgets needs.
Conclusions and Recommendations
In summary the ITIC 2010-2011 Global Server Hardware and Server OS Reliability Survey findings indicates that all of the server operating system platforms have achieved a high degree of reliability. However, the IBM AIX 7.1 operating system, followed closely by Windows Server 2008 R2, HP UX 11i v3 and Novell SuSE Enterprise Linux 11 are the top four most reliable server OS distributions.

Application Availability, Reliability and Downtime: Ignorance is NOT Bliss

Laura DiDio — Wed, 29 Apr 2009 19:28:04 +0000

Two out of five businesses – 40% – report that their major business applications require higher availability rates than they did two or three years ago. However an overwhelming 81% are unable to quantify the cost of downtime and only a small 5% minority of businesses are willing to spend whatever it takes to guarantee the highest levels of application availability 99.99% and above. Those are the results of the latest ITIC survey which polled C-level executives and IT managers at 300 corporations worldwide.

ITIC partnered with Stratus Technologies in Maynard, Ma. a vendor that specializes in high availability and fault tolerant hardware and software solutions, to compose the Web-based survey. ITIC conducted this blind, non-vendor and non-product specific survey which polled businesses on their application availability requirements, virtualization and the compliance rate of their service level agreements (SLAs). None of the respondents received any remuneration. The Web-based survey consisted of multiple choice and essay questions. ITIC analysts also conducted two dozen first person customer interviews to obtain detailed anecdotal data.

Respondents ranged from SMBs with 100 users to very large enterprises with over 100,000 end users. Industries represented: academic, advertising, aerospace, banking, communications, consumer products, defense, energy, finance, government, healthcare, insurance, IT services, legal, manufacturing, media and entertainment, telecommunications, transportation, and utilities. None of the survey respondents received any remuneration for their participation. The respondents hailed from 15 countries; 85% were based in North America.

Survey Highlights

The survey results uncovered many “disconnects” between the levels of application reliability that corporate enterprises profess to need and the availability rates their systems and applications actually deliver. Additionally, a significant portion of the survey respondents had difficulty defining what constitutes high application availability; do not specifically track downtime and could not quantify or qualify the cost of downtime and its impact on their network operations and business.

Among the other survey highlights:

A 54% majority of IT managers and executives surveyed said more than two-thirds of their companies’ applications require the highest level of availability – 99.99% — or four nines of uptime.
Over half – 52% of survey respondents said that virtualization technology increases application uptime and availability; only 4% said availability decreased as a result of virtualization deployments.
In response to the question, “which aspect of application availability is most important” to the business, 59% of those polled cited the prevention of unplanned downtime as being most crucial; 40% said disaster recovery and business continuity were most important; 38% said that minimizing planned downtime to apply patches and upgrades was their top priority; 16% said the ability to meet SLAs was most important and 40% of the survey respondents said all of the choices were equally crucial to their business needs.
Some 41% said they would be satisfied with conventional 99% to 99.9% (the equivalent of two or three nines) availability for their most critical applications. Ninety-nine percent or 99.9% does not qualify as a high-availability or continuous-availability solution.
An overwhelming 81% of survey respondents said the number of applications that demand high availability has increased in the past two-to-three years.
Of those who said they have been unable to meet service level agreements (SLAs), 72% can’t or don’t keep track of the cost and productivity losses created by downtime.
Budgetary constraints are a gating factor prohibiting many organizations from installing software solutions that would improve application availability. Overall, 70% of the survey respondents said they lacked the funds to purchase value-added availability solutions (40%); or were unsure how much or if their companies would spend to guarantee application availability (30%).
Of the 30% of businesses that quantified how much their firms would spend on availability solutions, 3% indicated they would spend $2,000 to $4,000; 8% said $4,000 to $5,000; another 3% said $5,000 to $10,000; 11% — mainly large enterprises indicated they were willing to allocate $10,000 to $15,000 to ensure application availability and 5% said they would spend “whatever it takes.”

According to the survey findings, just under half of all businesses – 49% – lack the budget for high availability technology and 40% of the respondents reported they don’t understand what qualifies as high availability. An overwhelming eight out of 10 IT managers – 80% — are unable to quantify the cost of downtime to their C-level executives.

To reiterate, the ITIC survey polled users on the various aspects and impact of application availability and downtime but it did not specify any products or vendors.

The survey results supplemented by ITIC first person interviews with IT managers and C-level executives clearly shows that on a visceral level, businesses are very aware of the need for increased application availability has grown. This is particularly true in light of the emergence of new technologies like application and desktop virtualization, cloud computing, Service Oriented Architecture (SOA). The fast growing remote, mobile and telecommuting end user population utilizes unified communications and collaboration applications and utilities is also spurring the need for greater application availability and reliability.

High Application Availability Not a Reality for 80% of Businesses

The survey results clearly show that network uptime isn’t keeping pace with the need for application availability. At the same time, IT managers and C-level executives interviewed by ITIC did comprehend the business risks associated with downtime, even though most are unable to quantify the cost of downtime or qualify the impact to the corporation, its customers, suppliers and business partners when unplanned application and network outages occur.

“We are continually being asked to do more with less,” said an IT manager at a large enterprise in the Northeast. “We are now at a point, where the number of complex systems requiring expert knowledge has exceeded the headcount needed to maintain them … I am dreading vacation season,” he added.

Another executive at an Application Service provider acknowledged that even though his firm’s SLA guarantees to customers are a modest 98%, it has on occasion, been unable to meet those goals. The executive said his firm compensated one of its clients for a significant outage incident. “We had a half day outage a couple of years ago which cost us in excess of $40,000 in goodwill payouts to a handful of our clients, despite the fact that it was the first outage in five years,” he said.

Another user said a lack of funds prevented his firm from allocating capital expenditure monies to purchase solutions that would guarantee 99.99% application availability. “Our biggest concern is keeping what we have running and available. Change usually costs money, and at the moment our budgets are simply in survival mode,” he said.

Another VP of IT at a New Jersey-based business said that ignorance is not bliss. “If people knew the actual dollar value their applications and customers represent, they’d already have the necessary software availability solutions in place to safeguard applications,” he said. “Yes, it does cost money to purchase application availability solutions, but we’d rather pay now, then wait for something to fail and pay more later,” the VP of IT said.

Overall, the survey results show that the inability of users to put valid metrics and cost formulas in place to track and quantify what uptime means to their organization is woefully inadequate and many corporations are courting disaster.

ITIC advises businesses to track downtime, the actual cost of downtime to the organization and to take the necessary steps to qualify the impact of downtime including lost data, potential liability risks e.g. lost business, lost customers, potential lawsuits and damage to the company’s reputation. Once a company can quantify the amount of downtime associated with its main line of business applications, the impact of downtime and the risk to the business, it can then make an accurate assessment of whether or not its current IT infrastructure adequately supports the degree of application availability the corporation needs to maintain its SLAs.