| PLEX86 | ||
Arpa address 3484some number of homogeneous operations also tend to have lots of stories about scheduled shutdown of the service while all components in the infrastructure are syncronously upgraded ... like the big 1-1-83 conversion for arpanet. sna and jes2 folklore has huge number of such stories. Arpa address 3489 actually there are two, somewhat separate heterogeneous capability provided by gateways ... one is somewhat the technology... jes2 was particularly senstive to homogeneous operation ... minor jes2 release changes would result in incompatible network communication between otherwise identical jes2 systems. the jes2 homogeneous operation even went so far that relatively minor jes2 release compatibilities could precipitate the failure of a jes2 operation, including bringing down the related mvs operating system. this problem became so bad on the internal network that the heterogeneous rscs-vnet support was enlisted to compensate for the homogeneous jes2 operation requirement. not only was a large library of jes2 nje drivers developed for vnet ... vnet was even enhanced to recognize when two different incompatibile jes2 systems were attempting to communicate and do the necessary format conversion. this eventually evolved in vnet essentially having a canonical representation of jes2 format and converting to the specific format needed by a directly connected jes2 operation (there is an early infamous story about a jes2 software upgrade on a system in san jose resulting in MVS system crashes in hursley) Arpa address 3486 Anne & Lynn Wheeler If a node is an IMP, then 100 nodes is not inconsistent with 250 hosts. Remember the mapping of IP addresses on net 10 (and 26... there are a large raft of similar customer stories about large SNA installations ... where large number of different components would have to be upgraded simultaneously ... i.e. shutting the service until all components have been upgraded ... as infrastructures got larger and larger ... it was becoming less and less possible to contain such a syncronous migration over even an extended weekend. part of the vnet infrastructure was a kind of gateway mechanism that allowed heterogeneous operation as well as the deployment of incremental upgrades w-o having to syncronize such upgrades across the whole infrastructure. some misc. RFCs ... from my rfc index. normally in frame's mode, the rfc summaries appear in the bottom frame. it is possible to "click" on various fields, clicking on ".txt=nnnn" retrieves the actual RFC 101 Notes on the Network Working Group meeting, Urbana, Illinois, February 17, 1971, Watson R., 1971-02-23 (14pp) (.txt=30343) (Updated by 108) ... lists there being 15 IMPss as of 2-23-71 (although many in purely test operation). Arpa address 3485 Anne & Lynn Wheeler from CSNET (Computer Science NETwork) is funded by NSF, and is an attempt to connect... 638 IMP-TIP preventive maintenance schedule, McKenzie A., 1974-04-25 (4pp) (.txt=4088) (Obsoletes 633) ... this appears to imply that Tuesday mornings from 7am to 9am are reserved for doing network software maintenance on all IMPs across the whole infrastructure. In addition there is somewhat rolling hardware maintenance schedule for different IMPs. This lists 45 IMPs as of 4-25-74. This RFC and several other RFCs imply that there was syncronized software upgrades of all IMPs across the whole infrastructure (implying a form of network infrastructure homogeneous operation) Arpa address 3487 re: so a location might have an IMP node for the network, a mainframe ... considered a HOST, another IMP working as a terminal server ... also considered a HOST... --
|
||||
Alt Folklore Computers from Newsgroups The #1 Usenet Provider on the Internet
|
||||