As usage of on-demand applications becomes the default setting for many business users and consumers, traditional ISVs as well as new market entrants must determine how best to meter and price consumption of their applications. What models exist and what are the pros and cons of each? What real-world experiences have you had with your approach, and what are you thinking of trying in the future? Join other providers of MySQL-based applications to explore these issues.
This is a chance for those new to Drizzle to ask questions, learn about development and new features, find out the best ways to get involved and contribute. We will also be speaking to Database Administrators who want to know what is different, and how they can leverage Drizzle's differences for their projects.
This will be a precursor to the Drizzle Developer Day to occur on April 13th following the conference.
If you want to get involved, what better way than to start off with the BOF session to learn about the project and then spend Friday hacking with the developers?
The Drizzle Charter:
o A database optimized for Cloud infrastructure and Web applications
o Design for massive concurrency on modern multi-cpu architecture
o Optimize memory for increased performance and parallelism
o Open source, open community, open design
Drizzle Project Scope:
o Re-designed modular architecture providing plugins with defined APIs
o Simple design for ease of use and administration
o Reliable, ACID transactional
To learn more about the Drizzle project, you can read Brian Aker's reflections of the 3 years leading into the Drizzle 7 release (http://krow.livejournal.com/700783.html).
The Drizzle Documentation is a good read too, docs.drizzle.org, and has sections such as A Brief History of Drizzle and Notable MySQL Differences.
From a single MySQL server instance to a cluster of MySQL servers on the cloud.
This discussion will explain the technology that transforms MySQL to a
cluster of integrated database and storage nodes that provides dynamic
scalability and High Availability on private and public clouds.
Andrew Aksyonoff the CEO/Founder of Sphinx Search:www.sphinxsearch.com will lead a community discussion to delve into Sphinx's advancements over the past year; from enterprise ready Real-Time indexes to the lesser understood features such as keywords dictionary or expression-based ranker. Furthermore, he will offer insight and answer questions on numerous search related topics including Sphinx vs. other Full-Text Search engines, such as MyISAM/InnoDB Full-Text Search or Lucene/Solr Search.
Polyglot Persistence: Pulling MySQL, Couchbase, memcached, Hadoop and Others Together
One is never enough. With the many different demands on different ways of using and accessing our data, methods of data management and analysis never exist in isolation for modern applications.
These applications use a combination of document stores for super fast, super simple data access, other structured systems for analytics and even distributed file stores alongside SQL databases. The MySQL community has long been open to embracing many other ways of managing and storing application data with the core SQL database capabilities.
This birds of a feather session will offer discussion and insight into how data can flow between a caching document store, a store optimized for analytics, different caching layers and the ever-flexible relational database.
One example will be shown with Couchbase Server as the fast, flexible schemaless store, with analytics for the data in Cloudera's Hadoop distribution. For richness of transaction capabilities and indexing, we have MySQL. Using all of the tools for persistence available today, one gains many new capabilities and higher scalability. Each does not deprecate the other; many deployments persist their data in two or more stores to optimize different ways of accessing and processing data.
Synchronous replication. Multi-master conflict detection. High performance replication across cloud availability zones or datacenters. Familiar InnoDB behavior. What's not to love about Percona XtraDB Cluster, the combination of Percona Server with XtraDB and the Galera write-set replication technology?
Query Optimization -- bring your naughty SELECTs. I will run this BOF as a clinic for fixing YOUR queries. I will quickly walk through what I see and how that leads me to an optimization strategy. Almost anything goes: InnoDB, MyISAM, PARTITION, JOIN, OFFSET, subquery, group-wise max, huge deletes, utf8, data warehouse, Replication delays
Bring, ready to display:
* the Query -- Please prettyprint it!
* SHOW CREATE TABLE tbl \G -- one per table (needed for datatypes and indexes)
* SHOW TABLE STATUS LIKE 'tbl'; -- (table size)
* EXPLAIN SELECT ... ; -- (not \G)
* (optional) One entry from the slowlog.
Some techniques are found here in http://mysql.rjweb.org . (I list it here so I can point at it in the BOF.)
This BOF will discuss strategies for implementing sharding and multi-master configurations, issues that are encountered by participants and best practices for addressing those issues.
MySQL is offered as a service by many cloud service providers, both at the Infrastructure as a Service (IaaS) level such as Amazon Web Services and OpenStack, and at the Platform as a Service (PaaS) level, such as Red Hat OpenShift, Heroku, VMware Cloud Foundry, and ActiveState Stackato.
This BoF will be for cloud service provider developers and cloud service provider user developers to meet, talk about the current state of practice, tips and tricks, and hopes and suggestions for future practice.
If you operate many database servers, how do you keep track of errors, which may be logged anywhere? Do you use Syslog? Nagios? Splunk? Or do you just shield your eyes and pretend nothing bad ever happens? Let's discuss best practices for monitoring error logs.
We would like to invite users of Continuent Tungsten software (both open source as well as commercial editions) to join the Continuent engineering team to discuss current and future plans for Tungsten, as well as interesting problems you have encountered. Please bring your ideas for improvements, problems you would like help in solving, and general feedback on our replication and clustering products. Or just stop by and say hi. The Tungsten team will be there and ready to brainstorm solutions. We look forward to meeting you!
Let's talk about MySQL and Ruby On Rails! Your host is Oleksiy Kovyrin, who's worked on many large-scale Rails+MySQL applications, including Scribd and LivingSocial.
Flash is a hot topic for MySQL. Vadim is Percona's resident flash/SSD expert, and has benchmarked and stress-tested most of the devices MySQL users buy. In addition, several flash vendors should be participating in the discussion.
