home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.dcom.cell-relay
- Path: sparky!uunet!portal!kyma
- From: kyma@shell.portal.com (Matt J Young)
- Subject: Implications for IBM
- Message-ID: <BzAAwL.CGy@unix.portal.com>
- Sender: news@unix.portal.com
- Nntp-Posting-Host: jobe
- Organization: Portal Communications Company
- Date: Tue, 15 Dec 1992 04:53:08 GMT
- Lines: 26
-
-
- Prior to deployment of the new networks, we have to make expli-
- cit their effect on the traditional mainframe business. To para-
- phrase the CEO of SUN, nobody benefits if IBM is dismembered.
-
- The advocates predict very low cost, low latency, high
- bandwidth. Hence, the user can add single processor boards to
- the data center for less than $3000, with an attached disc. As
- demonstrated by Sequent, these clusters make for powerful data
- base machines, and the design of the semantic data model favors
- simple, single tasking operating systems on each processor.
-
- However, since the new network processes asynchronously migrate
- through this semantic model, extracting some structured data to
- be duplicated onto the desktop, the desktop needs multi-tasking
- to manage its interaction with the structures, yet continue to
- offer an interface to the user. The desktop participates as a
- partition in the larger semantic model.
-
- The paradox then is a resurgence of simple DOS like systems in
- the data center with an imperative for multi-tasking on the desk-
- top.
-
- The new opportunity is for an OS/2 market on the desktop, while
- the demand for risc in the data center begs for a simple, single
- tasking kernel.
-