Performance with high db update rate

classic Classic list List threaded Threaded
2 messages Options
guherbozdogan guherbozdogan
Reply | Threaded
Open this post in threaded view
|

Performance with high db update rate

This post has NOT been accepted by the mailing list yet.
Hello,

I would like to query the following architectural query. The context of project employs the Apache Data Grid system as in memory grid for a web system and the cache data will also be persisted to a nosql db server cluster (data grid will be connected to a persistent nosql db cluster). What i would like to ask is, in a very high db update rate, would my following initial cautious  thinking be unnecessary; would   employin a message queue system like ZeroMq beside a Cron architecture like Apache Chronos to control/manage the rate of db updates be necessary or not? Are message queues of rest server and internal subsystems of Data grid are designed to work with overload data? In overload cases, are there performance penalties? Would we need to employ a pre controller system for batch db update requests? For a scalable system, do we cautiously need to deploy ZeroMq like queues for controlling continunous overload?


Thanks in advance
Best Regards
vkulichenko vkulichenko
Reply | Threaded
Open this post in threaded view
|

Re: Performance with high db update rate

guherbozdogan wrote
I would like to query the following architectural query. The context of project employs the Apache Data Grid system as in memory grid for a web system and the cache data will also be persisted to a nosql db server cluster (data grid will be connected to a persistent nosql db cluster). What i would like to ask is, in a very high db update rate, would my following initial cautious  thinking be unnecessary; would   employin a message queue system like ZeroMq beside a Cron architecture like Apache Chronos to control/manage the rate of db updates be necessary or not? Are message queues of rest server and internal subsystems of Data grid are designed to work with overload data? In overload cases, are there performance penalties? Would we need to employ a pre controller system for batch db update requests? For a scalable system, do we cautiously need to deploy ZeroMq like queues for controlling continunous overload?
Hi,

Ignite is designed to be scalable, stable and performant when working with lots of data and under high load. Actually, this is our main goal - to make things faster :)
Adding any additional layers to your architecture will most likely slow you down, so I would recommend to have them only due to some business logic requirements.

-Val