Speaker Interview: Rafia Sabih

Parallel query in PostgreSQL: How not to (mis)use it?   Friday 09:30   New York

Twitter: RafiaSabih Blog: rafiasabih.blogspot.com LinkedIn: rafia-sabih-b38721b9

Could you briefly introduce yourself?

At EnterpriseDB, I am actively involved in the development of upcoming storage engine for PostgreSQL, namely ZHeap. Previously, I was working in parallel query, where apart from developing some new parallel operators I did extensive TPC-H benchmark for parallel operators. This provided some strong evidence for the requirement of a few operators which were later developed and committed.

How do you engage with the PostgreSQL Community?

Primarily, I keep an eye on hackers community mails and review a few patches of my interest every now and then. Other than that I attend and present in postgres conferences.

Have you enjoyed previous pgconf.eu or FOSDEM conferences, either as attendee or as speaker?

This is my first time in PGCon Europe and I am all excited. Last two years, I spoke at PGCon India and absolutely enjoyed it.

What will your talk be about, exactly? Why this topic?

My talk is about utilising parallel query in PostgreSQL without over-utilising it or over expecting from it. Since, the advent of parallel query in PostgreSQL, people are excited about it because of the dramatic performance improvements it offers. But at the same time it is important to realise it’s limitation and tuning your server right for this feature to get the most of it.

I chose this topic, primarily, when I came across discussions about the ideal values of work_mem for parallel query, or other parallel query specific configuration parameters. It was in those discussions I realised that we have advertised it with well tuned parameters for benchmarks but never went in to the details of proper tuning or even discussing the scenarios where/where not to expect performance improvements with this feature. So, this time at PGConf Europe I’ll address the same.

What is the audience for your talk?

Anybody, who is either involved in using or development work of PostgreSQL will be appropriate to attend this talk. From users point of view, in this talk I’ll try to cover the use cases and tuning for most of the parallel operators. For developers, this talk will have details of the limitation of this feature and prospective directions in which as a developer we might want to go ahead.

What existing knowledge should the attendee have?

Just a basic understanding for PostgreSQL usage would be enough.

What is the one feature in PostgreSQL 11 which you like most?

Without a doubt, parallel-query!

Which other talk at this year’s conference would you like to see?

There are quite some exciting topics lined up for this conference, for which I’d like to congratulate the selection committee of the conference. I am planning to attend Review of Patch Reviewing by Stephen Frost and Fast, not Furious - What a benchmark can tell you by Rubens Souza.