Lost in the digital wilderness at an early age, Tom Lyon was adopted by feral punched card equipment that taught him to speak FORTRAN and inducted him in the cult of the mainframe.
After being rescued by a search algorithm of his own devising, he handily won a tri-state programming contest and was made to promise not to re-enter. In high school, he routinely abused the programmable calculators when he was not helping the teachers with their programming assignments. A summer’s NSF program in Hoboken opened his eyes to the potential of interactive computing (and digital porn). While in New Jersey, he also conducted a surreptitious census of the PDP-11s at the Bell Labs Holmdel site, and his faith in the mainframe was badly shaken.
Placed into the world’s finest liberal arts university, Lyon spent most of his time in the basement of the computer center. A chance encounter with a fellow student who was yelling ‘UNIX is coming!’ led him to discover that the use of operating systems did not have to involve sacrifices of body parts or loved ones.
Lyon proceeded to become a UNIX evangelist, yet still believed in the mainframe. He set out on a project to bring UNIX to mainframes (a fool’s errand which, sadly, continues to this day with Linux) and was noticed by the high priests of UNIX at Bell Labs. As a summer intern, Lyon helped Thompson, Ritchie, and Johnson liberate UNIX from the PDP-11 and create a ‘portable’ version of UNIX. The first target of the port was the Interdata 8/32, which, like many minicomputers of the ’70s, faded quickly into irrelevance.
Lyon’s Silicon Valley career was launched when he happened to be in the room with Ken Thompson when someone from Amdahl called inquiring about the possibility of UNIX on the mainframe. After a long, irrelevant year to complete his degree, Lyon joined Amdahl in Sunnyvale, CA, and literally brought UNIX to Amdahl – a splinter sect of the mainframe cult that believed mainframes were possible without IBM. Amdahl then produced what was one of the earliest commercial UNIX products, which was greeted by the market with widespread apathy.
Through his work with UNIX in California, Lyon had met Bill Joy, the leader of the BSD sect of UNIX. When Joy became a founder of Sun Microsystems, Lyon leapt at the chance to join him, becoming employee #8 of the promulgator of open systems. At Sun, Lyon wrote innumerable device drivers and contributed
egregious but long lived hacks to the NFS and SPARC architectures. As a member of the SunKit Racing Team, he helped implement the world famous April Fool’s Day hacks at Sun. He was also a key architect of the SunLink,
SunScreen, SunThis, SunThat, and SunWhatever product lines.
In the 90s, Lyon caught the ATM networking bug [note: the kind of ATM you only put money in to], architected one of the first ATM chips, and created the ATM AAL5 standard, which is still used today if you have a crappy ADSL connection. Having learned nothing about alchemy from trying to mix UNIX and mainframes, Lyon then proceeded to found Ipsilon Networks to mix ATM and IP networking. His IP Switching caught the world by storm and was an early example of how to gain total market awareness while still having no revenue. Ipsilon was rescued from its hubris by a humble band of radio engineers from Finland who sought to use its technology in wireless networks. Meanwhile, the folks at Ipsilon ‘pivoted’ into the firewall appliance market and built a $500M/yr business that was
ultimately unable to be integrated into the radio business.
Lyon left the Finns to start his own consulting firm, Lyon-About, LLC – a moderately successful effort to get venture capitalists’ money without actually doing any work. His real efforts during this time revolved around learning Linux
and NetBSD, and accidentally reanimating the abandoned Ricochet wireless network.
After a few years, Lyon became ashamed of not abusing his talents to the fullest, so he grabbed a cool million from the VCs to start Netillion, Inc. Netillion built distributed shared memory technology on 10 gigabit Ethernet, but it was too late for the former and too early for the latter.
While trying to find a home for the dying Netillion, Lyon was contacted by ‘MPLS’ – the legendary team that made Cisco what it is today. Ironically, MPLS is also the name of the technology which was the successful competitor of IP Switching. Joining with MPLS, Lyon became the sole legal founder of Nuova Systems, while the MPLS folks figured out whether they were or were not still Cisco employees. Nuova became a spin-in for Cisco, and developed the UCS server product line, as well as the Nexus 5000 switch family.
Lyon retired from Cisco at the end of 2010 and entered a prolonged period of omphaloskepsis. Near the end of 2012, he was convinced by his friend Satya to re-enter the world of technology and, in March of 2013, DriveScale was born. (Although the birth certificate says Samya Systems and neither founder fully accepts responsibility as the father). DriveScale started in a humble one-room executive office because neither founder had room in their garage and has grew to have five times as many computers as employees, with nary a mainframe or punched card in sight. At DriveScale, Lyon pursued alchemy via disaggregation, promoting a world in which servers, storage, switches, and
software can all assume their proper roles in the datacenter
After years of showing once again that having good products is not a sufficient business model, DriveScale was rescued from near death by a flock of birds from San Francisco, and Lyon was encaged in an organization where he can be more closely ‘followed’ to prevent future mishaps.
3 thoughts on “My Extended (Irreverent) Biography”
“Amdahl – a splinter sect of the mainframe cult that believed mainframes were possible without IBM”
Amdahl 470 mainframes had the same instruction set as the IBM 370, did they not? And they were plug-compatible with the IBM 370, weren’t they? (Though they were air cooled, unlike the IBM mainframes.) Anyway, this is what I recall from oh-so-many years ago. If I recall correctly, then the Amdahl mainframes would not have been possible without IBM.
Thanks for this great read. I worked at U S WEST in the ‘90s, testing ATM and Frame Relay equipment in their !nterprise lab. Liked reading about Datakit in your Bell Labs write-up as that was something I tested, along with Cascade, Stratacom and Newbridge.
Nice biography. I think I might even have my “Back the bit!” badge (er, button) somewhere.
(Guy from across the corridor in MTV14)