A brief history of BLADE SERVERS: From the Big Bang to the, er, 'unblade' Nowadays we take blade servers for granted, but a lot of moving parts had to come together for us to get where we are today. Tracing their history can help us make better judgments about how the technologies of tomorrow will evolve. The development of VMEbus architecture in about 1981 was perhaps the beginning of technological …
It was largely obsoleted by ATCA, especially in the telecomms space where blade servers are the technology of choice.
The article also omits one major practical advantage of blades, and the reasons that busses were specially designed for them: Hot Swap. One huge advantage of a blade over a conventional rackmount system is that when it goes wrong you just pop it out and slot in a new one. No cables to disconnect and (mis-)reconnect, no disks to swap or reimage, the blades are booted off external network storage. Blades have IDs related to the backplane slot, so you can configure a DHCP server to assign address 'X' and software load 'Y' to "the blade in slot 3", instead of having to reconfigure boot parameters to use the MAC address of a newly-replaced rackmount tray.
There was a book released by DEC that told you how to design interfaces to connect to the UniBus in the proper way. Along with the Schematics that cam with the Systems, yes Unibus was propetary but it was darned easy to build an interface for and DEC didn't stop you doing it. We had a dodgy interface that messed up all the DEC disgnostics even if it was placed after a Bus Repeater so we made sure that it was always disconnected when their Field Service came to do a PM on the 11/45 we were using.
Them were the days.
"it was darned easy to build an interface for [UNIBUS] and DEC didn't stop you doing it. "
Shortly after UNIBUS came Qbus, a cut down version for smaller systems.
DEC would happily sell you Qbus transceiver chips and higher level stuff (DMA stuff etc) to make it simple to build your own. Business model may have been applicable to Unibus as well, can't remember.
These chipsets were a bit like a precursor to the PLX9060 etc "generic PCI interface chip". Except the DEC ones came from a different era and were consequently a lot simpler to use.
It's almost as though Mr Pott submitted his usual well informed stuff but was told that Microsoft had rejected it, and he could either tone it down or write something else instead in the five minutes before deadline time.
Definitely not up to his usual standard, which is a shame.
And it's very likely PICMG not PCIMG.
ps
before the UNIBUS was the OmniBus (PDP8).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
