Another high speed open source in memory database option? Postgres with GPU backend
This was pointed a few weeks ago but this is just another but I want to verify that Redis is still the more viable option due to it being less maintenace with somewhat complicated coding with GPU or CUDA
Oracle 12c successfully install for my requirements for high speeding trading with message queuing in memory and more features
Sepcial NOTE: on installation: The Server install seems to be work ok but the Workstation some wierd permissioning problems for the database when installed.
Successfully and easily installed on my Windows 2012 Server. It is a strange world we live in now that it sort of works. After digging through, Oracle 12c could meet all my requirements.
Why Oracle?
When you read about Liquidnet being used, it got me wanting to try it out.
One terabyte may not sound like a lot, but because TimesTen also provides columnar compression, it’s equivalent to about five terabytes of addressable storage. Because it’s in main memory, the entire data store can be scanned within five seconds, according to Oracle. Additional Exalytics boxes can be added to increase total system capacity.
BTW, I once wrote a time series database which outperformed both TimesTen and Kdb. It supported a SQL-92 syntax. The key to the performance was binding the schema to C++ template based classes, and using memory mapped files. Allowing for on-the-run schema changes is quite expensive, and defining the schema at compile time, data lookups could be retrieved in just a few assembly instructions.http://stackoverflow.com/questions/1593987/alternative-to-the-timesten-in-memory-database
NOTE that this is all experimental but so far, I can report it seems pretty decent. As compared to SQL Server 2014 default, it seems not much of a memory hog as compared to SQL Server.
Is the FREE kdb + 32 bit in memory database decent for high speed tick data or is this open source Kona solution good enough?
This topic came out for some within London!
Read this as a potentially good in memory database option for tick data
https://news.ycombinator.com/item?id=7522327
This was mentioned by the contact:
https://github.com/kevinlawler/kona
Any opinions of this Kona?
Update: It was suggested the codingto learn how to reveerse engineer this Kona is not wise due to complex coding both in the interpreter and database languages.