Caching is a powerful weapon in the arsenal of an enterprise data architect. Designing an optimal caching strategy for your organization can increase performance and help save cost, but a poorly designed caching strategy often hurts more than it helps. We will cover strategies encompassing popular options available for caching.
Spider and Vertical Partitioning Storage Engine will be introduced by Giuseppe in this Conference.
In this session, I'll introduce some additional tips for joinable database sharding using Spider and Vertical Partitioning.
Many people view topics like Map/Reduce and queue systems as advanced concepts that require in-depth knowledge and time consuming software setup. Gearman is changing all that by making this barrier to entry as low as possible with an open source, distributed job queuing system. This session looks at advanced use cases that demonstrate the power and flexibility of distributed architectures.
Production MySQL deployments face constant challenges to keep databases up, avoid data loss, and scale performance. Tungsten offers a unique open source implementation of master/slave clustering that integrates replication, management, and SQL routing. We will present a tutorial introduction that shows how to set up and use Tungsten at home. Come see what the excitement is about!
Are you ensuring that a slave has the same data is the same as its master? Using mk-table-checksum from Ma'atkit as the basic tool, we will show you how to check that your data is in sync -- in a non-invasive way. We will share the tricks, tips and gotchas we have learned deploying this method in several different environments and schemas.
The Drizzle Project is a fork of the MySQL 6.0 server. One of the many goals of Drizzle is to enable a large plugin ecosystem by improving, simplifying, and modernizing the application programming interfaces between the kernel and the modules providing services for Drizzle. This tutorial serves to showcase the new APIs for Drizzle's replication through a series of in-depth examples.
Galera is emerging synchronous multi-master replication method, providing tightly coupled InnoDB clustering solution.
Galera has nice features like:
* High Availability - no lost transactions
* Performance - scalability even with 100% write rate
* Transparency - direct MySQL connections
MySQL/Galera 0.7 is production-ready GA release
Users of Gearman know that it is a distributed client/server system that
allows great flexibility to developers and DBAs.
Advanced users of this tool have come up with amazing examples of cunning hacks that will possibly help you in your daily tasks immediately or get you started in your path of creating your own hacks.
This talk is for those that know the basics of replication but want to take it to another level. We will show you how replication can solve complicated problems, give you better performance, and avoid excess downtime.
Gearman is a distributed client/server architecture that can handle tasks independently from programming language and operating systems. Using it, you can create flexible and efficient applications. Through MySQL UDFs, Gearman functions can me made available to a MySQL server directly.
War and peace stories from Facebook - how to detect, analyze and fix problems below, inside and above overloaded MySQL servers.
This covers concurrecy problems that appear in real world environments, beyond labs and benchmarks.
Facebook runs MySQL on commodity servers. Commodity servers have become extremely fast. Learn about the methods we use to identify performance bottlenecks in MySQL and the projects we have completed or started to keep pace with hardware advances.
Want to learn what the difference between a 10 server deployment and a 200 server deployment is in terms of administration? Best practices in large deployments? Got questions on centralized management and change management? Capacity planning? All these subjects and more will be covered in this session
Implementing multi-master setups with MySQL is fairly trivial in basic terms, but for an effective environment much more is involved. There are some tools (such as MMM, Multi-Master MySQL Manager) that can help, but there's still a lot to think of and each deployment is different.
Memcache is one of the tools of scaling read heavy architectures, but keeping it in sync with MySQL can be a challenge: scalability vs. consistency ? Do it in the app or the database ? Storage engines or Triggers + UDF ? Let's rethink it and make it scalable and consistent: let's replicate from MySQL to memcache.
Discuss what to look at when regular monitoring information is not descriptive enough. Review how to obtain useful status snapshots and how to interpret the results using real world examples covering items like: replication health, memory allocation, network issues and performance.
The binary log is a critical part of replication, but it can also be used for point-in-time recovery and auditing purposes. In this session, we present a thorough description of the binary log structure, the replication events, and how to use the mysqlbinlog tool to investigate and work with the contents of binary logs.
In this talk, the developers behind MySQL Replication walk through some of the new Replication features of MySQL 5.5
Replication is big news and not just because of built-in replication from MySQL itself. Projects like Drizzle, Tungsten, and Galera are changing the replication game with higher speed and greater flexibility than ever before. Join Jay Pipes (Drizzle Replication API) and Robert Hodges (Tungsten Clustering) for a tour of the new replication solutions and how they can help your applications.
This panel discussion, moderated by Sumeet Bansal, principal solutions architect for Fusion-io, the solid-state storage provider, will make clear to audience members the benefits of solid-state technologies, both from a performance standpoint and in relation to TCO. Q&A with the panel to follow. Participants include panelists from Answers.com, Cloudmark, Percona and Craigslist.org.
Distributed processing doesn't always require learning completely new methods. See how we solved an intense computation problem which was over-taxing our main database by distributing it across multiple machines, using only MySQL built-in functionality. No Map/Reduce. No external frameworks. It's simple -- and it works well, too!
Are you interested in splitting your application's reads and writes, so writes go to the master and SELECT queries go to slaves? This session is for you. Learn common patterns, pitfalls, and solutions.
Running multiple MySQL instances on the same hardware can be useful in a number of cases.
Some uses are rapid creation of test instances, consolidating several instances, mixing masters and slaves for redundancy and to balance load, and better hardware utilization. We will cover aspects including load balancing, LVM, repeatable builds, ports, proxy, and problems we encountered along the way.
Learn about the features built into MySQL Connector/J targeted at scalability and reliability. The session will cover fault tolerant load balancing, replication-aware load balancing with slave fail-over, custom exception handling and some creative use of statement interceptors mixed with the built-in load balancer to implement sharding.
Scaling Applications with Caching, Sharding and Replication
Sharding is a hot topic. Every big web site is using some sharding
technique with home made solutions. The quest for the silver bullet goes on
without apparent good results. This session will present two MySQL storage
engines (Spider and Vertical partitioning) that implement transparent
sharding techniques.
Sphinx is a FOSS full-text search engine that makes searching MySQL databases a breeze. This talk covers a number of recently developed advanced Sphinx features (RT indexes, parallel searching improvements, subquery caches, hitless indexes, "blended" indexing, SphinxQL interface, etc) in detail, explaining how and why to use them in your application.
We’ve been talking about memcached for a while now. The topic is exciting, scary, and constantly evolving. You might have a hard time understanding how memcached works at all. You might be missing the “big picture”, or you might just need a handful of examples of how to properly handle common queries. Come here to learn more from the experts!
Although MySQL Replication is easy to set up and easy to configure, it is nevertheless a very versatile tool that can be used to solve many problems in modern applications.
MySQL 5.1 introduces partitioning – a very useful feature for databases with large tables. The latest release, MySQL 5.5 enhances partitioning to make them even more user friendly and powerful. This tutorial explains the basics, and some advanced techniques to get the most from partitioning.
Memcached has become the de facto standard for caching web applications. But, many users jump in feet first without understanding what it does or perhaps more importantly what it does not do. Once you understand memcached, you may come to realize that it is what it does not do that makes it so good.
Join the MySQL and InnoDB Engineering teams as they describe the
low-level technical details behind the MySQL 5.5 performance and
scalability gains.