[
Lists Home |
Date Index |
Thread Index
]
I agree that one can aggregate that way.
Or, depending on how close to real time you need that
data, maybe these services should wake up in the off
hours (whatever those are in your geo-temporal-social
location), get all the data it needs for that day in
one coarse transaction, copy it to the local node,
and process it. How are rates being charged for using
the service? How binding are the contracts for the
goodness of the service and the information it provides?
One can go in either direction and aggregation is one
way to do it. So is replication. There are issues
that go beyond the speed of the network.
http://www.infoworld.com/articles/uc/xml/02/01/03/020103ucwhatis.xml
Scary stuff when it is applied from shrinkwrap to naked
real-time transactions.
Or this: How the Web Was Lost
http://www.clint.ca/oops/index.htm
(This is rated R: for language)
len
From: Amy Lewis [mailto:amyzing@talsever.com]
I'm not entirely sure that I agree with these best practices, although
they are likely to be good rules of thumb.
Consider this scenario:
|
