Async mirroring would be possible.  The bad / good thing is your
source dasd track could be updated several times between passes.  What
the controller does is look at the updated track table, 1 bit per
track, copies the tracks in order, resetting the bits when read and
acknowledged, then proceeds to the next track.  When it gets all those
tracks copied, it reads the updated track table for the next pass.

If you process the updated tracks fast enough, the number of tracks to
send drops with each pass until updates are sent almost immediately.
If you process the updated tracks too slow, then the number of tracks
to send keeps growing.

The key is the transmission rate to the endpoint.  I think the last
time I heard the limit was continent wide, because the links between
continents simply didn't have enough throughput capacity to get the
secondary up to speed.  Even with the land bridge, N/S America is
pretty slow, and African links too, despite cables across the
Mediteranian.  Maybe Europe / Asian would not be too bad.  Antarctica
is satellite only.

Australia is a long separation.  I had a friend in Australia who kept
getting notifications that her password was changed in Shanghai China,
so she changed it, got the message again, and repeated a few times.  I
pointed out that that was probably where the Australian underwater
cable came ashore and the location registered, because the date time
of the change was when she changed her password.

https://www.submarinecablemap.com/

On Fri, May 21, 2021 at 2:34 PM Cameron Conacher <[email protected]> wrote:
>
> Hello folks,
> I am trying to find out what we could consider to be a "safe" maximum
> distance for mirroring DASD from one site to another.
> We have a situation where site #1 will be thousands (9,000) of miles away
> from site #2 (across the ocean.
>
> I realize there would be a great deal of latency involved.
> But is this even possible or is it simply a non-starter?
>
> The second question I have is, suppose that I waved a magic wand and
> suddenly all my data was sitting comfortably in site #2.
> Could I (again ignoring latency for the moment) run a batch JOB or CICS
> transactions in site #1 that accessed the data 9,000 miles away in site #2?
> My thought here was if the data must absolutely be relocated, and
> people are willing to accept there will be latency, can we just access the
> DASD at site #2, rather than building a large data centre and performing
> all of the processing in site #2?
> Even if we run the processing at site #2, there will be many interfaces
> between site #1 and site #2.
>
> I know there are hard distance limitations.
> I think Global Mirror is limited to about 1500 miles.
> And I would hazard a guess that remote DASD would not be viable either.
>
> But I wanted to see if someone really "knows", rather than something I may
> have mis-remembered.
>
> Any thoughts/opinions are appreciated.
>
> Thanks
>
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN



-- 
Mike A Schwab, Springfield IL USA
Where do Forest Rangers go to get away from it all?

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to