🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: comp.sys.dec
Article #1376

Re: Info on Storage Subsystems for Vax Clusters

#1376
From: Stephen Tihor
Date: Wed, 19 Aug 1992 17:07
43 lines
2945 bytes
I am going to be a little careful as much of what I have been told was udner non-disclosure.  Fortunately not all.

Note that DEC has a LOT of information available undernon-disclosure in the US
since some of this stuff is getting close to announcement.

                I am looking out for equipment that will replace the HSCs
        or co-exist with them on the CI Bus, that is, connected to the Star
        coupler directly be faster in terms of I/O requests /sec and have
        some measure of redundancy maybe a RAID solution.

Digital has publicly mentioned that they are working on RAID solutions.  
THe only public mentions I recall were on software based solutions ala
Volume Shadows phase II but for higher levels of raid, ultimately all the way
up.  

                Have these Companies made agreements with DEC to interface
        on to the CI bus, or would we be likely to see some legal issues
        comming to the surface as they have in the past ?
Check careful (get in writing) wether the firmware is based on DECs new 
CI licensing program or will be converted to it.  There were some problems
with early products not being fully compliant and some of us have fought 
for a long time to get DEC to license out its valuable technologies.  
There are going to be some licensed products.  If a vendor can't cover license 
fees and wont; guarentee compatibility or replacement with a compatible 
solution then something is wrong.

Paralell disk arrays appear at the Files 11 level as one large disk, today.  
That puts some obviouslimits on its potential size.  Of course you can 
split those into pieces at the vax end and a full hardware soltuion 
like failsafe can lie about its configuration anyway it wants but there 
are some real advantages to keeping the physical to logical disk mapping 
relatively straightforwards.

                What happens in the Backup area?  If chunks are distributed
        across multiple spindles is backup in the traditional sense meaningful?

A storage solutions expand the notionof backup becomes much harder.  
At first blush you go to logical backups of the disks if you find the 
reliability insufficient or for other purposes.  Then youg et into
heirarchical datastores and have to add a whole new level of thinking.

Either it is a special case of date replication or it is nextus of tools
including a file system level journal and set of data dumps.
Technical Details
Message-ID: <59010003@acf3.NYU.EDU>
Path: arc1-ca.pugleaf.net!sparky!uunet!charon.amdahl.com!pacbell.com!mips!swrinde!cs.utexas.edu!rutgers!cmcl2!acf3!tihor
References: <1992Aug18.190607.89795@vaxc.cc.monash.edu.au>

Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!charon.amdahl.com!pacbell.com!mips!swrinde!cs.utexas.edu!rutgers!cmcl2!acf3!tihor
From: tihor@acf3.NYU.EDU (Stephen Tihor)
Newsgroups: comp.sys.dec
Subject: Re: Info on Storage Subsystems for Vax Clusters
Message-ID: <59010003@acf3.NYU.EDU>
Date: 19 Aug 92 17:07:00 GMT
References: <1992Aug18.190607.89795@vaxc.cc.monash.edu.au>
Sender: notes@cmcl2.nyu.edu (Notes Person)
Organization: New York University
Lines: 43
Nntp-Posting-Host: acf3.nyu.edu