As databases and applications mushroom in size and contention for resources brings the databases to its knees, implementing a caching layer using memcached (a.k.a memcache) can help to offload a significant amount of database traffic and greatly increase the scalability of the application. Percona's consulting experts have years of experience using memcached in a variety of different scenarios to offload traffic and increase overall application throughput.

If you don't implement your caching solution correctly, you can waste memory, increase network round-trip latency, and fail to divert as much traffic from the database as you could. Percona is the only company who has taken a quantitative, scientific approach to measuring memcached performance; we've built the tools to do it and released them as open source to the community.

Some of the ways in which we routinely help our clients with memcached include:

• Quantitatively measure how you currently use memcached and eliminate inefficiencies.
• Design a caching architecture best suited for the current environment.
• Scale session management far beyond the capabilities of a relational database.
• Offload often more than 90% of read traffic from the database.
• Evaluate non-standard use-cases such as using memcached replication.
• Work closely with memcached hardware appliance vendors on optimization for our clients' needs.
• Implement High Availability solutions with shared and dedicated memcached deployments.

Sometimes we help customers with cutting-edge atypical usage, such as implementing memcached as a Level 2 Cache for InnoDB (Percona employs Matt Yonkovit, who is the co-founder of the Waffle Grid Project).