I work for a company with a large investment in SAP, and we also have dozens of large .Net systems (mainly for engineering systems) and Java platforms (mainly for external web applications). So we have big development stores in ABAP, C # and Java EE.
We have more than 20 large objects distributed over very large distances, 2 data centers and 600 or so small objects in urban, rural and remote areas (I think, close to the North Pole).
We have decent standards for which the platform should be used in each functional domain, but we do not have much clarity regarding local processing and the use of distributed (central) processing. (i.e. thick client versus thin client)
I would like to build a decision tree based on best practices to help with this issue. Some of the criteria I've seen in the past include:
- Is a reliable network available wherever the application is installed? (If not, local processing may be the best option)
- Does the application require real-time data? (If so, centralized processing might be the best option)
- Does the application require intense processor power or high I / O? (If so, centralized processing or thick local processing may be options)
- Is the app required offline? capabilities? (If so, then local processing may be the best option)
- ,
? ( ,
)
- ,
? ( ,
: A
, ,
).
- ,
,
? (
)