CODEWITHSUNDEEP
javaweb-servicescachingjettyin-memory-database
Soumya Simanta
Soumya Simanta

Reputation: 11741

Options for In-memory databases (Open source and Java-based)

I've a web app that makes external web service calls on behalf of it's clients. I want to cache the data returns by some web services in the web app so that other clients can reuse this data and run filters and queries on this cached data.

The current architecture of the web app uses Apache Camel, Spring and Jetty. I'm looking for options (pros/cons) of in-memory database options.

Upvotes: 2

Views: 2332

Answers (3)

Nikita Koksharov
Nikita Koksharov

Reputation: 10763

Try Redisson - distributed and scalable familar Java data structures (Set, Map, ConcurrentMap, List, Queue, Lock, AtomicLong, CountDownLatch, Publish / Subscribe) on top of in-memory db Redis.

Upvotes: 0

ali haider
ali haider

Reputation: 20182

Hazelcast (Java API) - you can distribute the in-memory datagrid (with map, multimap, sets, lists, queues, topics) over multiple nodes very easily & use load/store interface implementation with a disk based DB. You can do something similar with EHCache.

Redis is another option (use the Java client to access it). You can simply configure the conf file to write data to disk (or avoid it altogether) & should not have to write your own load/store classes.

Besides these, there are a number of options you could use. Not sure if you are only looking at open source options, looking at distributed options or not.
Hope it helps.

Upvotes: 2

Jonathan B
Jonathan B

Reputation: 1061

Have you considered using MemCached? It is not a database, but a caching system you can control from inside your application.

Here are a few more thoughts about in-memory databases. First almost every modern RDBMS has a memory caching system inside it. The more memory you give to the database server (and configure it for caching) the more that it will store in memory for later. If you put together a system with enough memory to cache all the tables, you will have an "in memory" cache without the overhead of another database.

Most total "in memory" databases are used for high volume/large data systems where performance is totally key. And, because they are for extreme performance systems, you are going to pay for them. Or more specifically, pay extra for them. For example, the SAP/Sybase DB's that support full in-memory can cost you from 40% to 300% more than our existing products.

So, in answer to your question, do you really need one?

Upvotes: 0

Related Questions