Re: Huge Replication Scenario
- From: Hilary Cotter <hilary.cotter@xxxxxxxxx>
- Date: Wed, 30 Jul 2008 04:20:17 -0700 (PDT)
Basically for this sort of a large geographically dispersed
environment there is no good option.
What others have done is a terminal services application, using a SAN
solution with something like RecoverPoint or SRDF and having one site
active at one time. These are expensive solutions and require low
latency links.
Replication will work, but the latencies can be lengthy (ie several
hours) depending on your work load. They also tend not to be scalable
to large numbers of nodes - ie you will have problems after 10 nodes
but your WAN bandwidth may be saturated much earlier.
On Jul 29, 4:00 am, Peter <Pe...@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
If that isn’t a good option, I would consider it bad. So is there a good
option?
As to the latencies, that’s exactly the bottleneck in performance, which
users experience now. The data is now literally travelling across all the
oceans on this planet. I was thinking we would overcome this by having local
SQL Server installations. I’m kind of confused now.
Regarding the limited number articles, how do I determine the number of
articles needed for our system? Can it be estimated or even better calculated?
Thanks
Peter
"Hilary Cotter" wrote:
I don't think this is a good option for you. Latencies across the ocean can
be huge.
also there is a limit for the number of articles in a transactional
publication.
http://msdn.microsoft.com/en-us/library/ms143432.aspx
"Peter" <Pe...@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:B316A01C-1A6C-4E8F-9CCB-EC213E123B5F@xxxxxxxxxxxxxxxx
Hello,
We have several operational companies doing projects in the gas and oil
industry. These companies use a smorgasbord of applications running
against a
central SQL Server in Central Europe.
To overcome latency issues we are planning to have about a dozen SQL
Servers
all over the world and set up a bidirectional, transactional replication
between all these locations.
As all our subsidiaries work on their own projects conflicts seem to be at
a
minimum. What concerns us is our huge database, though. It encompasses
around
a thousand tables and several hundred views etc.
I spent some time sifting through this newsgroup but never found a post
pertaining a similar sized scenario.
Does anyone have industrial experience with replicating huge database
dispersed across various locations worldwide?
Any comment is highly appreciated.
Peter
.
- References:
- Re: Huge Replication Scenario
- From: Peter
- Re: Huge Replication Scenario
- Prev by Date: Re: Bidirectional Transactional replication
- Next by Date: Re: Huge Replication Scenario
- Previous by thread: Re: Huge Replication Scenario
- Next by thread: Re: Huge Replication Scenario
- Index(es):
Relevant Pages
|