PyPgDay 2013 Schedule
Please note: the schedule below is not final. We are still waiting for confirmations from some speakers, and the schedule may be rearranged to avoid conflicts.
Links: Registration, PyPgDay Info
PyPgDay 2013 Is Presented By:
|Time ||Speaker ||Category ||Talk |
|08:30 to 09:00 ||Coffee ||Breaks ||Coffee provided for attendees |
|09:00 to 09:30 ||Josh Berkus, |
|Introduction ||Introduction |
|09:30 to 10:15 ||Evan Kiltzke, |
|Migrating to PostgreSQL |
|10:30 to 11:15 ||Greg Smith, |
|Monitoring PostgreSQL with Python |
|11:30 to 12:15 ||Jignesh Shah, |
|Best Practices for HA and Replication in Virtualized Environments |
|12:15 to 13:15 ||Lunch ||Breaks ||Not catered |
|13:30 to 14:15 ||Jeff Davis, |
|PgPy Applications ||Using Range Types in Your Application |
|14:30 to 15:15 ||Mike Clarke, Disqus ||PgPy |
|Sharding with the Django ORM |
|15:15 to 15:45 ||Coffee ||Breaks ||Coffee Break |
|15:45 to 16:30 ||Christophe Pettus, |
|Extending PostgreSQL Into Python |
|16:45 to 17:30 ||Fazal Majid, |
|Scaling Mobile Analytics with PostgreSQL and Python |
|17:45 to 18:30 ||Various ||Various ||Lightning Talks |
|19:00 to 21:00 ||Salesforce ||Party || |
Party at Fault Line Brewing
Coffee Break Sponsor:
Talk Descriptions and Bios
Migrating to PostgreSQL
Evan Kiltzke, Uber.com
Uber has undergone a rewrite of its core API to migrate from MySQL to PostgreSQL, in support of various business needs. In this talk we'll discuss the reasons for our migration, the various steps we took to prepare our codebase, infrastructure and personnel, and cover the actual night of the switch.
Evan Kiltzke dropped out of college to spend four years building Yelp, took a stint at Google, landed at Uber. Lover of Hip-Hop, Bay Area native.
Monitoring PostgreSQL with Python
Greg Smith, 2ndQuadrant
PostgreSQL is the database backing store for more Python applications every year. But PostgreSQL's default configuration isn't optimized for good performance, and the job of monitoring how the server runs is left to external tools. This talk will take a look at the most vital performance issues and present at least one helper tool for each. That includes both pure Python tools (Graphite, pgtune) and helpful libraries that include Python interfaces to manage data (collectd, RRDtool). By the end you should have a clear list of initial actions to take with any PostgreSQL database for good performance tracking.
Greg Smith leads the US-based work for international PostgreSQL services firm 2ndQuadrant. He's the author of "PostgreSQL 9.0 High Performance", available from Packt Publishing.
Best Practices of Postgres HA and Replication in Virtualized Environments
Jignesh Shah, VMware
Learn the best practices with PostgreSQL HA and Replication to meet uptime requirements of typical enterprise deployment SLA - Service Level Agreements in the kind of virtualized datacenters which are the rule today. We will also cover the most common "gotchas" observed by database client applications.
Jignesh Shah is Sr R&D Manager of vFabric Postgres at VMware Inc. He has been involved with PostgreSQL project and PostgreSQL performance for many years.
Range Types in Your Application
Jeff Davis, Aster Data
PostgreSQL now upports native range types -- ranges of dates, timestamps, integers, etc. This is a powerful new feature, particularly for applications involving time and scheduling:
- No more "date_from/date_to" fields; have a unified range of dates in one field instead
- better performance: better indexing, better statistics
- less error-prone: 1 more long queries juggling NULLs to mean "hasn't ended yet"
- "range unique" exclusion constraint
Jeff Davis is a major contributor to the PostgreSQL project and community. Author of Sychnronized Scans (8.3), Exclusion Constraints (9.0), and Range Types (9.2).
Sharding with the Django ORM at Disqus
Mike Clarke, Disqus
Disqus was built from the ground up on Django and its ORM. After achieving widespread adoption, horizontal scaling & sharding data across multiple databases became critical for facilitating future growth. This talk is a deep dive into the sharding "magic" that enables Disqus developers to continue to use the ORM without concern for the underlying implementation details.
Mike Clarke works as the operations lead at Disqus, building infrastructure to support the nearly 1 billion monthly unique visitors across our network.
Extending PostgreSQL Into Python
Christophe Pettus, PostgreSQL Experts Inc.
PostgreSQL is the most extensible relational database in the world... but how do you take advantage of it from Python? Find out! We'll go over extending Postgres using functions, types, including taking advantage of these from frameworks such as Django.
Christophe has been working with PostgreSQL since 1997, and Python (and Django) since 2005. He currently consults on performance, horizontal scaling and application design for PostgreSQL Experts.
Scaling Mobile Analytics with Python and PostgreSQL
Fazal Majid, Apsalar
A description of how Apsalar uses Python and PG to run its mobile analytics service (several billion API calls a week). We use the Skype architecture (PL/Proxy, londiste) and a sharded DB setup to scale, do most of the processing in-DB in PL/PGSQL, and most of the rest in Python, including quite a bit of batch processing the PG optimizer can't deal with.
Fazal Majid is Co-founder and CTO of Apsalar and Kefta. 6 years' experience with PG, 18 with Oracle, 19 with Python. Scalability and telecoms background.
If you can't say it in five minutes, you can't say it at all! Join us for seven intense and rapidfire presentations on a variety of topics involving PostgreSQL, Python, and software development, including:
- Python for Social Content Filtering - Satish Viswanatham, Cloudable
- Simple Python Interface to PostgreSQL using JSON - Robert Geiger, Translattice
- Deployng Postgres in the Cloud with Ubuntu - Jorge Castro, Canonical
- Query Better with HTSQL: Catherine Devlin, Dell KACE
Come party at Fault Line Brewing with elephants, snakes, and our favorite "no software" company! Salesforce is sponsoring a party for all the PyPostgres geeks. Instructions on getting to the party will be given during the lightning talks.