Concept and Logic of ML
Technical Aspects of ML
Howto enter in production with ML
Perspective of ML
Here are some ideas that we could implement later on:
Now, when you import or delete or access a document, it is done according to the parameter that gets one Ip Port from the table of services. However one could want to access to the closest Legacy component implementing one Legacy in ML. This can be done using an algorithm that check the local IP address and getting the closest IP address that fits some rules and where the service is opened.
Or another way could be to try from one OpenLSD Server (let say the first one) and if not ready try the next one.
One could also want to use something like a load balancer between ML repositories. Although I don’t think it could be a good idea since they should be distant so as to enforce the security aspect of the physical storages.
-
We now can make a file containing database orders (insert, update, delete) such that the replication from an application point of view should be easier. This file could be generated for each step of the Op Handler.
We can improve the number of functions from standard OpenLSD to ML support.
-
We could improve production approach by including a test to see when we launch the Op Handler if this is the only one running.
