As of last year, the Linux operating system was running 90 percent of public cloud workloads; has 62 percent of the embedded market share and runs all of the supercomputers in the TOP500 list, according to The Linux Foundation Open Mainframe Project’s 2018 State of the Open Mainframe Survey report.
Despite a perceived bias that mainframes are behemoths that are costly to run and unreliable, the findings also revealed that more than nine in 10 respondents have an overall positive attitude about mainframe computing.
The project conducted the survey to better understand use of mainframes in general. “If you have this amazing technology, with literally the fastest commercial CPUs on the planet, what are some of the barriers?” said John Mertic, director of program management for the foundation and Open Mainframe Project. “The driver was, there wasn’t any hard data around trends on the mainframe.”
Cloud vs. Mainframe
Respondents were asked how they view cloud computing in comparison to mainframes, “and overwhelmingly, they said it’s going to augment the mainframe and not replace it,’’ Mertic says. Eighty-five percent of respondents said they are using cloud in addition to the mainframe, while 15 percent said cloud is replacing the mainframe.
This parallels the rest of industry, he says, noting that “everything’s going hybrid. You’re going to have some degree of private cloud, public cloud, some SaaS, some IoT solutions — and the mainframe fits in there. Isn’t it great we have a menu of choices in architecture, and that they complement each other.”
In terms of security, only five percent of overall respondents said cloud solutions are at least as secure as mainframes, which Mertic said surprised him the most. Even among self-identified cloud users, only six percent agreed that cloud is at least as secure as mainframes.
“These are people who feel cloud is the future, but then say, ‘Wow, if security is a top priority, cloud isn’t there compared to mainframe, which is still far more secure,’’’ he said. “That surprised us, especially among people who self-identified as being on the cloud bandwagon.”
General findings include:
- Both users and non-users of Linux on the mainframe agree on the mainframe platforms strengths of performance, security, and cost reduction.
- Cloud computing is perceived to augment the mainframe, not replace it.
- Mainframe will continue to have a key role in a hybrid computing world.
- 28 respondents said cloud solutions are significantly less costly to implement, compared to 31 percent of self-identified cloud users.
- 47 percent of respondents said cloud solutions provide much greater flexibility in adding power or scalability, compared to 52 percent of self-identified cloud users.
- 15 percent of both self-identified cloud users and overall respondents said that over the long run, cloud solutions are far less expensive to maintain, while 57 percent of general respondents disagreed with that, compared to 59 percent of self-identified cloud users.
Standing the Test of Time
Key workloads for Linux on the mainframe are application servers, database servers and data analytics, Mertic says. Even though IT infrastructure has changed compared to what it was 10, 15 and 20 years ago, the mainframe has withstood the test of time, he says.
“The mainframe has always stood for four guiding principles: availability, scalability, security and performance from the introduction of the mainframe in 1964 on through to today,” Mertic says. “IT in 2018 has tons of choices, and if a first-class open source platform with those guiding principles are what your organization requires at a ‘10’ level, then modern mainframe is an architecture you should consider.”
The survey of 145 mainframe professionals was conducted between Aug. and Oct. 2017.
Stay tuned for part two of the Open Mainframe Project survey in the fall, which will provide broader insights into how the general IT market views the modern mainframe.
- How Blockchain Changes the Nature of Trust - 01/22/2019
- Machine Learning, Biased Models, and Finding the Truth - 11/27/2018
- AI in the Real World - 11/15/2018