Servers

NASA: Powering down the last mainframe

With the consumerization of IT, legacy systems are going the way of U.S. manned spaceflight. NASA shuts down its last mainframe -- a two-ton, 56 square-foot monster running Red Hat Enterprise Linux 4 (RHEL4).

Image courtesy: NASA

You know it's the end of an era when even NASA doesn't need a two-ton, 56 square-foot computer anymore.

Crazy! The National Aeronautic and Space Administration (NASA) is powering down its last number-crunching mainframe, says NASA CIO Linda Cureton in a blog post this weekend.

The outgoing mainframe is an IBM Z9 approaching two tons and running Red Hat Enterprise Linux. And NASA is showing it the door. At its site, IBM lists the specs:

In her post, NASA CIO Cureton found it necessary to "define what a mainframe is" for her "millennial readers." Another sign of the times. Here's an excerpt:

Marshall Space Flight Center powered down NASA’s last mainframe, the IBM Z9 Mainframe. For my millennial readers, I suppose that I should define what a mainframe is. Well, that’s easier said than done, but here goes -- It’s a big computer that is known for being reliable, highly available, secure, and powerful. They are best suited for applications that are more transaction oriented and require a lot of input/output – that is, writing or reading from data storage devices.

They’re really not so bad honestly, and they have their place. Things like virtual machines, hypervisors, thin clients, and swapping are all old hat to the mainframe generation though they are new to the current generation of cyber youths.

In my first stint at NASA, I was at NASA’s Goddard Space Flight Center as a mainframe systems programmer when it was still cool. That IBM 360-95 was used to solve complex computational problems for space flight. Back then, I comfortably navigated the world of IBM 360 Assembler language and still remember the much-coveted “green card” that had all the pearls of information about machine code. Back then, real systems programmers did hexadecimal arithmetic – today, “there’s an app for it!”

Keeping the mainframe going likely cost thousands of dollars a month in storing and cooling costs, a price NASA probably can't justify in this age of ever smaller, faster and nimbler computing devices. Even so, NASA CIO Cureton says mainframes like the IBM Z9  shunting are "still a requirement for ... many other organizations ... the end-user interfaces are clunky and somewhat inflexible," Cureton continues, "but the need remains for extremely reliable, secure transaction-oriented business applications.

Somewhat inflexible?

NASA has yet to indicate what hardware it is replacing the Z9 with. I've got a call into NASA.

In the meantime, check out this picture. It's the IT equivalent of Mad Men era computing. Break out the martinis, tobacco and horned-rim glasses. Pick your poison.

About

Gina Smith is a NYT best-selling author of iWOZ, the biography of Steve Wozniak. She is a vet tech journalist and chief of the geek tech site, aNewDomain.net.

14 comments
Grumpy_IT_bod
Grumpy_IT_bod

...as space consuming as those pictures. Mainframes have moved on. IBM has newer mainframe models. Besides newer mainframe models can integrate with newer apps. May I add that mainframes still have a better performance and load handling capability than web based platforms. When you add to that the security of your data then it still has a strong case for being kept on as part of an organisation's legacy systems.

bobmatch
bobmatch

Programming and care for these over grown calculators were to be the economy???s saving???s grace. As they fell by the way side and PCs took over that was the start of the boom years. Who cares that manufacturing went overseas, we got IT and PCs. Now it???s cheaper to buy a new laptop than get the old one fixed, or just do your computing on your phone you got for free with a 2 year agreement. Need A Job? Wish there was an App for that! But not to worry, the next big thing is just around the corrner, we hope...

Zorched
Zorched

...with the remnants of Control Data and had worked with supers from them and SGI. There's a lot of technology that could transfer over to PCs yet, even decades after their haydays. A fair number of mainframe capabilities have been coming out on PC chips. The big thing we could benefit from is something called security rings. Think of the CPU core commands being surrounded by differing size rings. The closer ring you get the more access you have. These are hardware enforced so can't be bypassed by software. If they had these on PCs to protect against software install I think we'd have a lot less compromising of systems. Graphics from SGI on their scalable supers were the main renderers for big animated movies up till the middle 2000's when Desktop cards started incorporating a lot of the things SGI had had for over a decade. It's amazing how people these days go, "Wow, that's new and cool!" Well, to you maybe.

melinda.baccus
melinda.baccus

Yeah, Gina, but they weren't 100 lbs overweight and they had... P E N S I O N S.

Loggies
Loggies

"...14 digit multiplications at 330 miilion per minute"......How does this compare to today's desktops using Intel i7's...." ?

mperata
mperata

The image of the guy in a business suit with white shirt and tie loading a tape into a 2420 brought back memories of my first job out of the Navy in 1968, when I was an IBM CE and working on the 65s and 85s at SLAC in Menlo Park, CA.

MeijerTSR
MeijerTSR

I still wouldn't give it to Iran.

JohnOfStony
JohnOfStony

I first encountered assembler on a Plessey Myproc 16-bit microprocessor consisting of a circuit board heavily populated with TTL chips with their part numbers erased to prevent copying of the design. This was in 1977. Then I moved to a full-time software development job using the Intel 8085 and a development kit which had no hard drive, just twin 8 inch (160k) floppies. Next came the Motorola 6800 and 5.25 inch floppies leading to the Zilog Z80 and CP/M. Then came the DEC VAX - the nearest to a mainframe I've ever worked on but not at assembler level. But it was back to assembler and C on the ARM in 1986-7 before I left assembler forever. However, ARM assembler is certainly the nicest assembly language I've ever used - a superbly designed instruction set in which every instruction is conditional, thus eliminating a lot of short branches and thus speeding up program operation as every branch breaks the efficiency of the pipelining fetch, decode, execute process. I am very glad I've lived through such an exciting and rewarding period of computing history.

hellingerga
hellingerga

I fondly remember my 360 Assembler programming class at BGSU in 1978/79. Gotta love the boxes of punched cards too!

FuzzyIce
FuzzyIce

That chip can do 158.4 GFLOPS - 158 x 10^9 floating point operations per second. Half of that in double precision (according to this doc: http://www.intel.com/standards/floatingpoint.pdf) . Do the math... But the thing is, nowadays mainframes are still fast, really fast. Only problem is the high cost of ownership

Stargzer
Stargzer

... installing cables under the raised floor in the computer room.

rdrcomp
rdrcomp

I was with Sperry Univac back then, and the picture does look familiar

Stargzer
Stargzer

My college had a 360 Model 25 with a whopping 32K of core, two 2311 disk drives (7.5MB (MEGA-Bytes) each, as large as a washing machine), card reader, punch, and printer. Sophmore year they upgraded it to a full-blown 48K, the most it could handle, and IBM had to fly in the guy that designed that model because they couldn't get the memory to install! A far cry from the 360/95! My Agency still runs a couple of Z-series mainframes. Nothing can beat a mainframe for real data crunching, storage, I/O, and security. The Word of the Day is MANDY: Mainframes Are Not Dead Yet.