Wanda brings up the most important point.  This isn't just about
backups; this is about restores.  I had a customer once that backed up
over 250 clients, each a different site around the whole country, many
connected by only a 256kbs frame relay circuit.  On many clients the
first full incremental took 2 weeks to get done, but after that the
incremental only took a few hours, and was acceptable.

But they did as Wanda suggested.  They had spare servers at the central
site, and if a significant restore had to be done they would restore it
to the local server and ship it overnight.  It happened so rarely that
the hassles of doing it were acceptable compared to having to support a
backup infrastructure at each remote site.

But in your case, would this even work?  How long would it take to ship
a server and get through customs, etc.   I used to work for an IBM
business partner, and some IBM equipment was manufactured overseas, like
Germany and Hungary.  It sometimes took a few weeks to ship to the US.
Maybe a small server would not be so slow, and going out of the US might
be faster than coming back in.  But if I were you I would try shipping a
server over there just to set everyone's expectations. 

Best Regards,

John D. Schneider
Lead Systems Administrator - Storage
Sisters of Mercy Health Systems
3637 South Geyer Road
St. Louis, MO  63127
Phone: 314-364-3150
Cell: 314-486-2359
Email:  John.Schneider AT Mercy DOT net


-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Wanda Prather
Sent: Wednesday, March 19, 2008 9:03 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: [ADSM-L] Looooooooooong distance backups


I've got customers doing a few trans-atlantic backups.
TCP/IP is a pretty forgiving protocol; there really isn't much
difference in the issues doing a backup across 7K compared to 1K.  It's
all about the amount of data and the size of the pipe.

I assume they already have I/P connectivity to you at some level; find
out what kind of link they have - T1, T3, something better?  Figure out
how much bandwidth that gives you, compared to how much you need to back
up per day (and who else needs the bandwidth and when).

Use everything you can to reduce the amount of data to be transmitted;
make agressive use of excludes, turn on client compression, use subfile
backup if it's feasible on your client.

And figure out what you are going to do in case of a system failure.  If
the total client is many GB, how long will it take you to restore the
whole thing? Frequently on a WAN, incremental backups are practical, but
full restores aren't - a good strategy is to keep a spare server in your
data center; in case you need a full system rebuild, do it locally and
then Fedex the box to where it needs to go -- it will get there faster!


On 3/18/08, Zoltan Forray/AC/VCU <zforray AT vcu DOT edu> wrote:
>
> We at VCU (Richmond, VA) have a "campus" at Doha, Qatar  (approximate 
> distance is 7K miles.
>
> We are discussing the viability/possibility of having TSM clients in 
> Qatar backing up to our TSM servers, here in Richmond!
>
> Is this possible?  Feasible?
>
> What are the implication of such a long distance backup?
>
> Anybody else doing this kinda thing?
>
> I gather we would probably need to do things like encryption (who 
> knows how many countries/routers the link would go through) as well as

> CRC checking!
>
> I also question the legality/licensing/export issue?
>
> Thoughts?  Ideas?  Suggestions?  War-Stories?
>