For the 12th straight year, IBM’s Z mainframe and Power Systems, achieved the highest server; server operating system reliability and server application availability rankings, along with Lenovo’s ThinkSystem servers which delivered the best uptime among all Intel x 86 servers for the last seven consecutive years, in ITIC’s 2020 Global Server Hardware and Server OS Reliability survey.
ITIC’s latest independent survey data finds that the most reliable mainstream server platforms – the IBM Power Systems, Lenovo ThinkSystem, Hewlett-Packard Enterprise (HPE) and Huawei KunLun deliver up to 26x more uptime and availability than the least dependable unbranded “White box” servers.
The superior uptime of the above top ranked mission critical hardware makes them up to 34x more economical and cost effective than the least stable White box servers.
High end mission critical servers from IBM and Lenovo both registered under two (2) minutes of per server, per annum unplanned downtime due to inherent flaws in the underlying hardware or component parts. Cisco, Hewlett-Packard Enterprise (HPE) and Huawei server platforms were close behind: each recorded approximately two minutes or a few seconds more downtime attributable to inherent issues with the hardware. Among mainstream servers, IBM POWER8 and POWER9, along with the Lenovo x86 ThinkSystem servers; the HPE Integrity Superdome X and Huawei’s mission critical KunLun servers continue to deliver the highest levels of reliability/uptime among 18 server platforms. (See Exhibit 1).
The least consistent hardware – unbranded White box servers – averaged 53 minutes of unplanned per server downtime due to problems or failures with the server or its components (e.g. hard drive, memory, cooling systems etc.). This represents an increase of four (4) minutes of downtime compared with ITIC’s 2019 Global Server Hardware, Server OS Mid-Year Update survey.
ITIC’s independent Web-based survey polled over 1,200 businesses worldwide from November 2019 through March 2020. The study compares and analyzes the reliability and availability of over one dozen mainstream server platforms and one dozen operating system (OS) distributions. To obtain the most accurate and unbiased results, ITIC accepts no vendor sponsorship.
IBM’s System Z server is in a class of its own. It maintained its best in class rating among all server platforms. An 83% majority of IBM respondent organizations said their firms achieved five and six nines – 99.999% and 99.9999% – or greater uptime. Nine-in-10 IBM Z customers reported that the mainframe recorded just 0.62 seconds of unplanned per server downtime each month and 7.44 seconds annually due to inherent flaws in the server hardware or its component parts. Less than one-half of one percent of IBM Z respondents said the mainframe experienced unplanned outages exceeding four (4) hours of annual downtime.
The economic annual downtime cost comparisons among the top performing and the least reliable server hardware platforms is staggering.
A single hour of downtime estimated at $300,000, equates to $4,998 per server/per minute.
According to that metric, organizations using the most reliable IBM POWER8 and POWER9; Lenovo x86-based ThinkSystem; HPE Integrity or Huawei KunLun servers that experienced just under or just over two (2) minutes would spend $9,996 in annual per server downtime costs due to inherent flaws in server hardware or component parts (See Table 2).
By contrast, corporations using Dell PowerEdge servers which experienced 26 minutes of per server/per minute downtime at the same $300,000 per hourly downtime rate potentially would rack up yearly outage costs of $130,026 for a single server.
Corporations deploying the least reliable unbranded White box servers that registered 53 minutes of per server, per minute downtime can expect to incur possible downtime losses of $264,894 specifically related to server hardware flaws and bugs in the OS and applications. The four additional minutes of downtime – from 49 minutes per server in ITIC’s 2019 poll, to 53 minutes of per server outage time in 2020, represents a cost increase of $19,992 compared with the White box server 2019 per server, per minute downtime price tag of $244,902.
Time is money.
The higher monetary costs associated with unbranded White box servers are not surprising. The unbranded White box servers frequently incorporate inexpensive components. And some businesses recklessly run unsupported or pirated versions of operating systems and applications. The aforementioned hourly downtime examples are for just one server. Downtime costs can mount quickly and reach into the millions for corporations with dozens or hundreds of highly unreliable servers.
Survey Highlights
Among the other top survey findings:
• Reliability: IBM Power Systems and Lenovo ThinkSystem hardware and the Linux operating system distributions were once again either first or second in every reliability category, including server, virtualization and security.
• Availability: IBM Z mainframe, Power Systems, Lenovo ThinkSystem, HPE Integrity and Huawei KunLun all provided the highest levels of server, applications and service availability. That is, when the servers did experience an outage due to an inherent system flaw, they were of the shortest duration – typically one-to-five minutes.
• Technical Support: Businesses gave high marks to IBM, Lenovo, HPE, Huawei and Dell tech support. Only 1% of IBM and Lenovo customers and 2% of HPE and Huawei users gave those vendors “Poor” or “Unsatisfactory” customer support ratings.
• Hard Drive Failures Most Common Technical Server Flaw: Faulty hard drives are the chief culprits in inherent server reliability/quality issues (58%) followed by Motherboard issues (43%) and processor problems (38%).
• IBM, Lenovo and Huawei KunLun Servers Had Fewest Hard Drive Failures: IBM, Lenovo and Huawei’s KunLun platforms experienced the fewest hard drive quality or failure issues among all of the server distributions within the first one, two and three years of service. Less than one percent – 0.4% – of IBM Z mainframes, for example, experienced technical problems with their hard drives in the first year of usage, followed by the IBM Power Systems and Lenovo ThinkSystem with one percent (1%) each during the first 12 months of deployment.
• Security is Top External Issue Negatively Impacting Reliability: Security and data breaches now have the dubious distinction of being the top cause of downtime.
• Minimum Reliability Requirements Increase: An 88%majority of corporations now require a minimum of “four nines” of uptime – 99.99% for mission critical hardware, operating systems and main line of business (LOB) applications. This in an increase of five (5) percentage points from ITIC’s 2018 Reliability survey.
• Patch Time Increases: Seven-in-10 businesses now devote from one hour to over four hours applying patches. This is primarily due to a spike in wide ranging security issues such as Email Phishing scams, Ransomware, CEO fraud as well as malware and viruses.
• Increased Server Workloads Cause Reliability Declines: The survey data found that reliability declined in 67% of servers over four (4) years old, when corporations failed to retrofit or upgrade the hardware to accommodate increased workloads and larger, more compute intensive applications. This is up 23% from the 45% of businesses that said uptime declined due to higher workloads in the ITIC 2018 Reliability poll.
• Hourly Downtime Costs Rise: A 98% majority of firms say hourly downtime costs exceed $150,000 and 88% of respondents estimate hourly downtime expenses exceed $300,000. Just over one-third of ITIC survey respondents – 34% – estimate the cost of a single hour of downtime now tops one million ($1,000.000).
Server hardware, server operating system – and by extension, virtualization reliability, uptime and availability are the core foundational elements of the overarching health of an organization’s entire Digital Age ecosystem and the life blood of daily business operations.
The core reliability of corporate servers, server operating systems and the mission critical applications that run on them are absolutely imperative. The inherent reliability of enterprise hardware, OS and applications are necessary to maintain daily, uninterrupted business operations; ensure secure access to proprietary assets; mitigate risk and drive revenue.