TSDB vs HBase : What to choose in long term? [closed]

Question

Closed. This question is opinion-based. It is not currently accepting answers.

---

Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.

Closed 6 years ago.

Improve this question

The problem we are trying to solve : Store all the datapoints that could be queried later point to deduce performance of a webservice. Give a visual representation of data.

Here is my understanding so far from what I have read on Hbase and on TSDB sites

• TSDB stores data in Hbase, albeit in a little non-standard way (wide-row format)

• TSDB is blazingly fast, and gives ways of adding graphs very quickly.

• HBase while slow compared to TSDB, gives you option of using SIMBA ODBC driver to connect it to Tableau, which is amazingly agile and beautiful visual data-representation.

One more issue with TSDB is it doesn't work with current HBase .96 (I know tsuna@ is working on it but I dont have timeline for a stable release of tsdb which works against current Hbase.96)

So your simple answer could be Hbase or TSDB, or in addition to choosing one over another you can justify your answer and help me choose one over another.

I am open for other solutions, if they can support about trillion data-points stored over a year.

Answer 1

In my honest opinion, If you need plotting data, you should consider using OpenTSDB. I'm not sure in a long term period what it could be better because OpenTSDB is continuosly developing and in November it was released the RC 2.0 version. There is also a fork of OpenTSDB that is called KairosDB and it uses Apache Cassandra as database, so that you can store points up to milliseconds, while OpenTSDB allows you to store points up to seconds.

So I suggest you Timeseries Distributed Databases like OpenTSDB/KairosDB for plotting data in a quite simple way. Not sure about the long term issue, maybe something better could be developed over Hbase in the next month because of its opensource nature.

Answer 2

If you're set on Hadoop ecosystem, Axibase Time-Series Database is another product that runs on top of HBase. We have a built-in integration with Tableau and JDBC driver that allow you to fetch time-series data (raw and aggregated) with SQL-like queries, which we support.

Source code and examples available on https://github.com/axibase/atsd-jdbc