We are looking at InfluxDB to store large numbers of streamed measurements (1-2 tera-samples). Additionally, we would like to be able to also store audio and video streams corresponding to the measurements (not all of them but many). To me at least this makes sense, since it is all time base data. But I don't see any discussion of this online.
I imagine that the video data could be broken up into frames. And that the audio data could be broken up into 100msec audio frames.
Has anyone tried this? Any recommendations?
Most time-series databases are optimized for storing floating point values, with the occasional string here and there. Storing BLOBs beyond perhaps 1KB is likely not a good use case for InfluxDB, although we haven't done much performance testing with larger binary data.
That said, I don't quite follow your use case. It seems more like you need to index audio and video, rather than store and analyze time series data. TSDBs aren't just optimized for storing things with time as the primary axis, they are also optimized for aggregating those values and looking for change over time. Your use case doesn't seem to involve any aggregation or pattern searching, just a simple look-up table by time.
I would think a NoSQL database would be just as good for this, or perhaps OpenTSDB, which builds on top of Cassandra.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With