I am trying to identify possible methods for storing 100 channels of 25 Hz floating point data. This will result in 78,840,000,000 data-points per year.
Ideally all this data would be efficiently available for Web-sites and tools such as Sql Server reporting services. We are aware that relational databases are poor at handling time-series of this scale but have yet to identify a convincing time-series specific database.
Key issues are compression for efficient storage yet also offering easy and efficient queries, reporting and data-mining.
How would you handle this data?
Are there features or table designs in Sql Server that could handle such a quantity of time-series data?
If not, are there any 3rd party extensions for Sql server to efficiently handle mammoth time-series?
If not, are there time-series databases that specialise in handling such data yet provide natural access through Sql, .Net, and Sql Reporting services?
thanks!