3 ms·
"We realized that this problem needed a new kind of database, and so we built TimescaleDB: the first open-source time-series database designed for scalability a
by cstanley 9y ago
"We realized that this problem needed a new kind of database, and so we built TimescaleDB: the first open-source time-series database designed for scalability and complex analysis."
I'm not a DB expert, but what have financial firms been using for decades? Likely dealing with much larger securities datasets...
Or, is it just that TimescaleDB is 'opensource'...
- beagle3 9y agokdb+ from Kx systems is popular among users who have mostly tick data (billions of trade/bid/offer updates). It has bare bones SQL, and its own query langauge "q" which in many cases simplifies the queries immensely (with two features, "foreign key chasing" and "as-of joins" which are somewhat at odds with Codd's original pure relational model so are unlikely to get into the SQL standard)
- stalller 9y agoYes, they have. kdb+ has been the standard bearer within finance for more real-time trading applications. Still, q as a query language tends to marginalize more of the organization from having access to the data. TimescaleDB offers full-SQL at scale, opening the database up to more people. It's used to power longer-term trading platforms (for e.g., commodities trading) and interactive analytical dashboards (for e.g., research analysts). We also see applications in compliance and risk management. And yes, TimescaleDB is open-source. So we see TimescaleDB as complementary to kdb+, which has been around for much longer as a proprietary system, is certainly full-featured and powerful, but is used more in real-time trading applications.