persistence - Are there databases that bases durability on redundancy and not on persistent storage? -


sorry title isn't obvious, couldn't word better.

we right using conventional db (oracle) our job queue, , these "jobs" consumed number of nodes (machines). db server gets hit these nodes, , have pay lot software , hardware database server.

now, occurred me other day that,

1) there multiple nodes in system
2) "jobs" may not lost because of node failures, there no reason have sitting in secondary storage (no reason why couldn't reside in memory, long not lost)

given this, couldn't 1 retain these jobs in-memory, making sure @ least n number of copies of job present in entire cluster, thereby getting rid of db server?

are such technologies available?

did take @ gigaspaces? on internet scale, not need persist @ all. have know sufficient copies around. if have low latency connections places not on same powergrid (or have battery power), pushing out transactions duplicates enough.


Comments

Popular posts from this blog

windows - Why does Vista not allow creation of shortcuts to "Programs" on a NonAdmin account? Not supposed to install apps from NonAdmin account? -

c++ - How do I get a multi line tooltip in MFC -

unit testing - How to mock PreferenceManager in Android? -