At Twitter, we have gone through many iterations of storage systems as
we have coped with tremendous growth. In the process of building
these, we have been able to solve some of the core problems related to
distributed storage in a general fashion with a framework we wrote
named Gizzard. This has allowed us to quickly and safely develop
specialized components for the parts of Twitter that can no longer
handle the scale at which they are required to operate.
In this talk, I will go over how we got to where we are now, the core
principles we have followed to make writing and reasoning about these
systems as simple as possible, and how we have benefited from using
MySQL as a solid base.
A set of general distributed systems design principles
Application of said principles to a few concrete problems
Building off of established technology
Coordination among multiple readers/writers
Resilient error handling
At the end of this talk, whether or not they use Gizzard, participants
will be better equipped to develop their own scalable systems, having
seen how certain not necessarily intuitive solutions can help solve
many different distributed storage problems.
Matt is a software engineer on the Storage team at Twitter. The team is responsible for designing and building the fundamental storage services at the heart of Twitter’s platform. Before Twitter, Matt worked as full-stack developer at social gaming startup Serious Business. He lives in Berkeley, CA.
For information on exhibition and sponsorship opportunities at the conference, contact Yvonne Romaine at firstname.lastname@example.org
For media partnerships, contact mediapartners@ oreilly.com
For media-related inquiries, contact Maureen Jennings at email@example.com
To stay abreast of conference news and to receive email notification when registration opens, please sign up for the O'Reilly MySQL Conference Bulletin (login required).
View a complete list of O'Reilly MySQL Conference Contacts