I took liberty of x-posting URL to (linkedin open) "Old Geeks"
http://lnkd.in/YzVS6A
and (linkeding closed) "IBM Historic Computing"

note that NJE/NJI ... used left over entries in the 255 entry psuedo
unit record table to define network nodes ... typically around 160
... however the internal network had relatively early exceeded 200 nodes
... forcing JES2 to purely boundary nodes (JES2 would trash traffic
where the origin and/or the destination nodes didn't appear in its
table). 

fortunately the internal network was based on rscs/vnet which also had a
very clean layered architecture ... as a result it was relatively
trivial to implement a wide variety of drivers ... including ones that
simulated NJE. This was used to great advantage in JES2 because the
implementation had jumbled together job control fields and network layer
field ... and JES2 systems at different release levels ... that talked
directly to each other ... would frequently result in taking
down/crashing the whole MVS system. It then became responsibility of
internal VNET/RSCS to have a whole library of JES2 drivers ...  and it
became the responsibility of VNET/RSCS to start the JES2 driver
... talking directly to a JES2 system ... to convert all JES2 headers to
the format exactly required by that specific JES2 system release (as
countermeasure to different JES2 systems at different release levels
constantly crashing each other). There is the infamous case of Hursely
MVS systems crashing because of files arriving from San Jose MVS JES2
system ... and the local Hursely VNET/RSCS system being blamed
... because it wasn't correctly converting the header fields (as
countermeasure to JES2 systems at different release levels crashing each
other).

Prior to December 1976 ... VNET/RSCS wasn't going to be announced as
product ... because it was in the wake of Future System failure and POK
was in the process of convincing corporate to kill the vm370 product and
transfer all the people to POK to work on MVS/XA. JES2 NJE wasn't going
to be announced because it had to be charged for ... and requirement was
customer charges times the forecast had to cover the costs ... and there
was no customer forecast times any price that covered the JES2 NJE costs
(company was still adapting to 23Jun69 unbundling announcement where
non-kernel software was being charged for). A deal was then cut to make
a joint JES2/NJE plus RSCS/VNET product announcement. RSCS/VNET total
product costs was so small and the forecast was so large ... that it
would have been possible to ship at $30/month and still meet all product
pricing requirements. As a result, a joint product announcement at
$600/month ... resulted in joint revenue that covered joint costs (in
effect RSCS/VNET revenue was subsidizing the JES2/NJE product announce).

trivia: the internal network was larger than the arpanet/internet from
just about the beginning until late 85 or early 86. by the time JES2/NJE
got around to supporting 999 nodes, the internal network was over 1000
nodes and by the time JES2/NJE extended to 1999 node support, the
internal network was over 2000 nodes.

misc. past posts mentioning internal network
http://www.garlic.com/~lynn/subnetwork.html#internalnet
misc. past posts mentioning Future System
http://www.garlic.com/~lynn/submain.html#futuresys
misc. past posts mentioning HASP, JES2, NJE/NJI, etc
http://www.garlic.com/~lynn/submain.html#hasp


-- 
virtualization experience starting Jan1968, online at home since Mar1970

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
  • JES History Lizette Koehler
    • Re: JES History Anne & Lynn Wheeler

Reply via email to