This is a story I have been tracking for some time now. Oracle is one of the companies I was enamored with early in my career, Sun microsystems as well. They are both legacies here in Silicon Valley, absolutely. I can now confirm that Oracle is getting out of the chip business. Oracle got into the chip business with the $7.4B acquisition of Sun Microsystems in 2009 (which was a diving catch in my opinion) and has continued to make SPARC CPUs down to TSMC 20nm. Unfortunately, they fumbled 20nm and never made it into the FinFet era which was their undoing, my opinion.
Sun really took the EDA business by storm in the 1980’s replacing mini computers with Motorola 68000 based systems that not only out performed the legacy DEC and Data General systems, they ran a native UNIX OS and were much less expensive. Sun also started the “personal CAD workstation” revolution with the SparcStation that was based on the Sun SPARC RISC CPU paired with a GPU.
About the time IBM PC clones were storming the market Sun licensed SPARC in an attempt to create a Sun clone market. One of those companies was Solbourne Computer (I worked for Solbourne). Solbourne (born of sun) did quite well in the server market with their multi cpu server business. They were the first to SPARC multiprocessing and made huge amounts of money selling (4) CPU servers. Back then EDA software was system based so you could almost get four times the EDA throughput on a Solbourne system using the same software license as a single CPU Sun system. This was short lived of course since the software vendors changed their licensing when Sun started shipping multiple CPU systems. Back then a CPU was a board versus a chip.
Sun played some dirty tricks on the clone vendors (holding back OS releases, nasty sales FUD, and such) so the clone business died a slow and expensive death. Sun went on to hyper-growth during the dotcom boom then hyper-bust when the dotcom money dried up. Sun was never really the same after that and was ready to go belly up when Larry Ellison came in for the save. Larry and Sun co-founder Scott McNealy were friends so it was an inside deal. Larry wanted to go up against database competitors HP and IBM by providing turnkey DB systems but Oracle is now all-in on cloud computing.
The big failure here was how Sun operated internally in my opinion. One glaring example was the CAD group. Sun was notoriuosly unfair to EDA vendors. One time we were told point blank that, “Sun did not buy EDA software”. We had to give them the software for free but they would pay a 15% support fee. Sun was also slow to adopt commercial tools and to this day Oracle still has a very large CAD tool development group. Yes, Oracle still develops some of their own EDA tools.
Sun really changed the way we designed chips taking us from mini computers in the back room to workstations on our desks. Similar to what Apple has done taking computing from the desk to our pockets. R.I.P Sun Microsyetms and thanks for the memories (and CPUs)…
Please share your Sun Stories in the comments section…. This really is an end of an era.
Share this post via:
Next Generation of Systems Design at Siemens