Unanswered: Simple Distributed Partioned Views (DPV's) Replication & Sheep
Last Night I had a Dream
And it involved my favourite Design Technique - Simplicity
So to achieve Simple Replication across geographically disparate Servers we could use:-
300 plus SQL2000 Servers enabling 1000 Concurrent Active Clients (10% actual light Activity)
WAN = National Private Secured (ping 300ms max) (128kbps leased line Min)
One Interface Server constantly Running A Simple Dynamically Built Partitioned View (removing down servers)
One Stored Proc (or more) that Synchronizes the DPV Updateable Partitioned View with
Interface Mirror Table DeNormalized Holding a Physical Copy of each of the Subscribers/Publishers/Client Sql Servers Data
The Question is:-
Was my Dream a Nightmare OR A Dream Come True ?
I Know it's down to the Network Quality to a great degree but (That's the suck it and see part of the question)
but as a form of replication it seems a very simple platform that could possibly tackle our friend
The DCP (Data Consistency Problem) with Client Update DateTime Column & frequent activation (30 Secs).
Has anyone had much experience with this type of Scaling out over a WAN ?
Is it worth a Try in the Real World ?
"Everything should be made as simple as possible, but not simpler." - Albert Einstein
"Everything should be made as complex as possible, so I look Cleverer." - Application Developer