I simply found "Has anybody released a detailed comparison between different in-memory RDBMSs?" that's related to my question.
It features a lot going for this including.
- Fast, consistent response time
- High transaction throughput
- Standard SQL, no application rewrite
- Persistent and recoverable
- High availability with no data loss
Yet it's listed from the achieve on most people ($41,500.00 / Processor).
What exactly options exist, and what exactly are there benefits and drawbacks.
(I'm using .Internet whether it changes the way to go)
As 1000 individuals have seen this quesions, I've made the decision to include a bounty to ascertain if a much better answer could be got.
A well known in-memory database product is McObject's eXtremeDB, see
There's additionally a good whitened paper, offered for download in the webpage of McObject's Site, known as, "In-Memory Database Systems: Misconceptions and Details". It clarifies some questions that arise when talking about the sun's energy, for example, "how's an IMDS not the same as a 'traditional' database running on the RAM disk?"
Generally, within this area people also take a look at kdb from Kx Systems. I believe these two systems really had they hay-day in early 2000's. I'd find out if a updated MySQL having a solid condition RAID array would provide you with the performance you'll need.
BTW, I remember when i authored a period series database which outperformed both TimesTen and Kdb. It supported a SQL-92 syntax. The important thing towards the performance was binding the schema to C++ template based classes, and taking advantage of memory planned files. Permitting for on-the-run schema changes is very costly, and determining the schema at compile time, data searches might be retrieved in a couple of set up instructions.
An alternate might be solidDB, an evaluation between your two might be found here: TimesTen and solidDB comparison