From OpenSQLCamp

up to Boston 2010 main page

The sessions will be chosen in an un-conference style on the day of the event, but session topics should be posted here to give other people an idea of what to expect. If you would like to lead a session, please describe it here. This may be a more formal presentation with slides, or possibly just leading a group discussion on a hot topic. Also feel free to toss ideas out to the mailing list to get feedback in case you are not sure. We want to encourage people from all levels to lead sessions, share your thoughts!

How to add a proposed session

Copy and paste the "Title - Presenter" template just under this text and put the name of the title and presenter(s) in the heading, and the description in the body.

Title - Presenter

Description goes here; just a few sentences to illustrate the idea.

Perl and Database Programming - Patrick Galbraith

Are you a LAMP developer? Patrick Galbraith recently wrote a book about programming with memcached, MySQL, Apache and Perl and is also the maintainer of DBD::mysql. He would like to have an informal session with developers and DBAs who use Perl on a daily basis to see if there might be some cross pollination of ideas on how to solve common problems or develop great applications. Other language lovers are welcome too.

Membase - Patrick Galbraith

Membase is an open-source (Apache 2.0 license) distributed, key-value database management system optimized for storing data behind interactive web applications. Have you ever wished that memcached had durability? Well, this is Membase! Patrick Galbraith will discuss just what Membase is, as well as provide a demonstration of installing and utilizing membase.

Checking MySQL's replication sync with Ma'atkit tools - Sheeri Cabral

The theory has been around since an article by Baron Schwartz in October 2008. Sheeri Cabral tackled this daunting task for clients of the Pythian Group and found a few surprises -- from options that don't work as they are supposed to, to learning that a lot more data was in sync than originally guessed. Sheeri will lead folks through the theory of setting up continual replication sync, and the actual process, including the important aspects to be aware of when doing this on your own systems.

mysqltuner 2.0: What the next generation can do - Sheeri Cabral

mysqltuner 2.0 is a launchpad project that can do a lot more than the original mysqltuner. Through a simple config file, you can choose what to check, what thresholds are problematic, and even what to print out if a threshold is met. Gathering metrics is made very easy by mysqltuner's simple model where a string is matched to status/system variables and replaced with the value -- mysqltuner can be easily extended to use other sources for data, including SHOW ENGINE INNODB STATUS). You also control how often to run mysqltuner, which makes mysqltuner great for a one-time sanity check of performance variables, or for frequently collecting metrics.

Ideas with no concrete presenter:

What is different about the Percona build? (no presenter)

Would like to see a presentation on what is different in the Percona build so I can assess if I want/need to use it.

Feature overview of the Percona Server - Matt Yonkovit

A review of the current features that exist in XtraDB and the Percona Server. Analysing topics like: How do these differ from the stock MySQL builds? How these can be used now to solve current problems? Which features are the most compelling and widely used? Can these be used to solve the energy crisis? and more.

Common MySQL Performance Mistakes - Matt Yonkovit

Who does not love to tune their databases? Isn't it what we all dream of every night? What are the problems and issues we see on a week to week basis... Let's talk about some of the tricks and techniques Percona uses on their consulting assignments that assist customers in reaching their goals. High level topics include:

 * Goal oriented optimization
 * tools of the trade ( i.e. maatkit, cacti, etc )
 * Hardware and OS tuning 
 * Problems and pitfalls to avoid

Solar Power Data Monitoring with MySQL and other open source goodness - Matt Yonkovit

I have spent the last year working with a Solar Power Inverter Manufacturer working on a Government project to monitor solar power usage and consumption around the US. If anyone wants I can walk people through some of the gotcha's with a large project such as this, the architecture and infrastructure we put into place, and some of the cool things we have done or are planning to do in the future.

Discussion: Solving the Counter Problem for Distributed Databases - Josh Berkus

One of the fundamental issues facing distributed databases in maintaining consistency is the ability to have an "absolute clock" which allows each node to determine what order changes are to be applied in. While intuitively simple, implementing an absolute clock or cluster-wide counter is either technically improbable or a major performance bottleneck. I'd like to have a group discussion on the steps various distributed / clustered databases have taken to solve this.

NOTE: this session depends on having appropriate people there from various transactional clustered databases. I see Mason Sharp, but we'd need some folks from NDB/Cassandra/RIAK/Hypertable/etc.

What happened at the MariaDB Developer Meeting - Colin Charles

A summary of what was discussed at the MariaDB Developer Meeting in Istanbul, Turkey, just last week.

What's in MariaDB these days - Colin Charles

MariaDB is something that folk have been hacking on for over a year. What has been achieved? What is being worked on? From simple community contributions like Croatian collations, to some more interesting stuff like table elimination that Monty Program has worked on, there's features to be looked back on.

Data integrity : comparing the SQL and NoSQL approach - (Giuseppe Maxia + NoSQL Presenters wanted)

This is an open session that should be helpful to every relational practitioners, to understand how NoSQL products deal with common problems of data integrity. Ideally, it should be a panel where an expert from each product explains how such cases are handled:

Insert data that contains repeating contents (e.g. a sales invoice containing many products details and the customer's billing info)
update data concurrently, making sure that nobody modifies stale data.
update two related pieces of data atomically (e.g. the savings account of the donor and the receiver)

Note: Alex Esterkin can present NoSQL

Teaching SQL - Brainstorming - Josh Berkus

One of the reasons for many developers turning to NoSQL databases, aside from specific features, is that they don't know or like SQL. Is there some way that we can help developers to be less afraid of SQL? Maybe even to like it? Have you taught SQL to developers in the past, and what techniques did you use?

Making ORMs Not Suck - Discussion - Josh Berkus

One of the critical issues in database applications today is that every one seems to be built on an ORM, and every ORM seems to be half-baked. I'd like to discuss a number of questions:

Are some ORMs better than others? Which ones?
Is it possible to build a really good ORM? What would it look like?
How can we improve existing ORMs to not make our databases look bad?
Is there an alternative to ORMs other than asking developers to write SQL (which they won't)?

MVCC Implementations - Panel - Josh Berkus

Several databases implement the MVCC model (append-only tables) for concurrency. This would be a compare-and-contrast panel, with questions about:

VACUUM and other cleanup
Handling long-running transactions
Handling DDL (where applicable)
Other MVCC issues
MVCC vs. Undo Log and other approaches

Multi-Tenancy in Relational Databases: Someone from Drizzle

Apparently Drizzle is going to have a brand-new implementation of multi-tenancy catalogs, resource control, and other apparatus for their eventual final release. I'd love to have them talk about it (JMB).

Cassandra and Lucene - T Jake Luciani

This session will include and intro to Cassandra both operations and data modeling. Then we will look at how Cassandra can be used as a storage backend for Lucene.

Replication and the quest for a global transaction ID - Giuseppe Maxia

A proper failover in a MySQL replicates server requires identifying the points where the slaves stopped replicating from the master. MySQL doesn't have this feature by default. You can have a patch (Google, in 5.0) or an external tool (Continuent), or you can hack your own solution (as suggested by a recent book). And you can hack the hack, as shown in this session.

MVCC Unmasked - Bruce Momjian

Multiversion Concurrency Control (MVCC) allows Postgres to offer high concurrency even during significant database read/write activity. MVCC specifically offers behavior where "readers never block writers, and writers never block readers". This talk explains how MVCC is implemented in Postgres and highlights optimizations which minimize the downsides of MVCC. This talk is for advanced users.

Rapid Upgrades with pg_upgrade - Bruce Momjian

Pg_Upgrade allows data to be transferred between major Postgres versions without a dump/restore. It does this by transfering the user data and version-dependent data separately. This talk explains the internal workings of pg_upgrade and includes a pg_upgrade demonstration.

The Magic of Hot Streaming Replication with Postgres - Bruce Momjian

This talk explores the much-anticipated Postgres 9.0 features of hot standby and streaming replication. It explains how these features work, how to configure them, and their current limitations. It includes a hands-on demonstration that can be done either by the instructor or by students.

Database Hardware Selection Guidelines - Bruce Momjian

This introductory talk explains why database servers have hardware requirements different from other infrastructure software, specifically unique demands on I/O and memory. The talk also covers various I/O options and their benefits.

Resource management through admission control policies (Josh Berkus)

Are any other databases using, or attempting to use, admission control (i.e. queueing requests by resource usage, resource "pools") to regulate system resource usage among database requests? PostgreSQL is just getting into implementing this, and I'd like to discuss what has and hasn't worked for other database systems. Particularly:

What are you regulating with admission control? Memory, CPU, I/O, others?
How well is it working?
What are the tunables?
How are you dealing with locking on the admission control counters?

The Databases of the Future Discussion (facilitator: Josh Berkus

What kinds of databases do you think people will be using in 2020? What features will they have? What kinds of databases will people stop using? Discuss.

Understanding Indexing - Zardosht Kasheff

This talk presents three simple and effective rules on how to construct indexes that result in good query performance.

Optimizing for $/txn - Ariel Weisberg

Define what it means to optimize for dollars per transaction, and explain at a technical level six ways to do this and why it can't be done any other way.

Data must fit in memory
Stored procedures as the unit of transaction
Single threaded execution of transactions to completion
Establishing a global timestamp order for transactions
Partitioning (not sharding!)
Transactionally export old data

Using SQL Databases as NoSQL Persistent Store - Alex Esterkin

This talk will present the case for employing MySQL and/or other RDBMS as NoSQL Database Persistent Store.

The attendees will learn why NoSQL databases make sense and how they work.

Some NoSQL databases allow for using relational DBMS as persistent store. The attendees will learn how to use MySQL in that role.

If there is time, the talk will outline the challenges of building massively partitioned MPP databases in outer space.

Note: this talk can only be presented on Sunday.