Designing and Implementing

Scalable Applications with

Memcached and MySQL

A MySQL® White Paper

June, 2008

Copyright © 2008, MySQL AB

 Table of Contents
Designing and Implementing ........................................................................................................... 1

Scalable Applications with Memcached and MySQL ................................................................ 1

Table of Contents................................................................................................................................... 2

Introduction............................................................................................................................................. 3

Memcached Overview .......................................................................................................................... 3

How Memcached Works ...................................................................................................................... 4

Memcached Server ................................................................................................................................ 5

Memcached Clients ............................................................................................................................... 8

Example Architectures......................................................................................................................... 9

Solutions: Memcached for MySQL................................................................................................. 12

Conclusion .............................................................................................................................................. 13

Additional Resources.......................................................................................................................... 14

About Sun’s MySQL Portfolio .......................................................................................................... 15

Copyright © 2008, MySQL AB Page 2

Introduction
A “cache” duplicates, some or all the data stored on one or more databases for fast and
scalable access by applications. Leveraging a caching system is valuable when the original
data is expensive to fetch from the source database due to resource constraints. Once the
source data is populated in the cache, future requests for the same data make use of the
cached copy rather than fetching the data from the original source. Introducing a “caching-
layer” typically makes data access times faster and improves the ability for the underlying
database to scale in order to accommodate consistently higher loads or spikes in data
requests. Often, the cache is “distributed”, so that the contents of the cache can be spread
out over multiple systems to make more efficient use of available memory, network and
other computing resources.

Memached is an open-source, distributed memory caching system designed to tackle

today’s web-scale performance and scalability challenges. Many of the largest and most
heavily trafficked web properties on the Internet like Facebook, Fotolog, YouTube, Mixi.jp,
Yahoo, and Wikipedia deploy Memcached and MySQL to satisfy the demands of millions of
users and billions of page views every month. By integrating a caching tier into their web-
scale architectures, these organizations have improved their application performance while
minimizing database load. In doing so, they manage to increase their scalability to
accommodate more concurrent users and spikes in traffic while at the same time making
the most efficient use of their existing computing infrastructure.

Memcached Overview
Memcached is an actively maintained open source project distributed under the BSD license
with regular contributions from not only a community of developers, but also corporate
organizations like Facebook and Sun Microsystems. Danga Interactive originally developed
memcached to improve the performance and scalability characteristics of the blogging and
social networking site LiveJournal. At the time, the site was delivering over 20 million
dynamic page views per day to over 1 million users. With LiveJournal’s implementation of
memcached as a caching-tier, the existing load on the databases was significantly reduced.
This meant faster page loads for users, more efficient resource utilization and faster access
to the underlying databases in the event data could not be immediately fetched from
memcached.

Since memcached was implemented at LiveJournal, it has subsequently been deployed by

some of the largest web properties on the Internet. Memcached is most commonly
leveraged by organizations to increase the performance of their dynamic, database-driven
websites which handle large amounts of read requests by caching not only data and result
sets, but also, objects, like pre-compiled HTML.

Memcached is designed to take advantage of free memory on any system running Linux,
Open/Solaris, BSD or Windows with very low CPU overhead characteristics. It can be
installed on dedicated servers or co-located with web, application or database servers. Out-
of-the box memcached is designed to scale from a single server to dozens or even hundreds
of servers. Facebook currently has the largest known deployment of memcached servers in

Copyright © 2008, MySQL AB Page 3

 production, numbering well over 800.1 The ability to massively scale out is one of the key
advantages of deploying memcached as a caching-tier.

How Memcached Works

A hash is a procedure for turning data into a small integer that serves as an index into an
array. The net result is that it speeds up table lookup or data comparison tasks. Memcached
leverages a two-stage hash that acts as like a giant hash table looking up key = value pairs.
Memcached can be thought of as having two core components, a server and a client. In the
course of a memcached lookup, the client hashes the key against a list of servers. When the
server is identified, the client sends its request to the server who performs a hash key
lookup for the actual data. Because the client performs one stage of the hashing,
mamcached naturally lends itself towards the ability to easily add dozens of additional
nodes. An added benefit is because there is no interconnect or multicast protocol being
employed, the impact to the network is minimized.

Example Memcached Lookup

A basic memcached lookup can be illustrated in the example below with Clients X, Y, Z and
Servers A, B, C.

‘Set’ the Key and Value

• Client X wants to set the key “foo” with value “seattle”

• Client X takes the list of available memcached servers (A,B,C) and hashes the key
against them
• Server B is selected
• Client X directly connects to Server B, and sets key “foo” with the value “seattle”

‘Get’ the Key and Value

• Client Z wants to get key “foo”

• Client Z is able to use the same hashing process to determine that key “foo” is on
Server B
• Client Z directly requests key “foo” from Server B and gets back “seattle”
• Subsequent requests from Clients X, Y or Z for key “foo” will access the data directly
from Server B

1
http://highscalability.com/strategy-break-memcache-dog-pile

Copyright © 2008, MySQL AB Page 4

 Figure 1: Basic Memcached Lookup

Memcached Server
As previously mentioned, there are two core components to memcached, the server and the
client. In this section we cover some of the core capabilities of the memcached server,
including how the memcached server deals with memory allocation, the caching of data.
Overall, the memached server is implemented as a non-blocking event-based server with an
emphasis on scalability and low resource consumption.

Memory Allocation
By default, the memcached server allocates memory by leveraging what is internally
referred to as a “slab allocator”. The reason why this internal slab allocator is used over
malloc/free (a standard C/C++ library for performing dynamic memory allocation) is to
avoid fragmentation and the operating system having to spend cycles searching for
contiguous blocks of memory. These tasks overall, tend to consume more resources than
the memcached process itself. With the slab allocator, memory is allocated in chunks and in
turn, is constantly being reused. Because memory is allocated into different sized slabs, it
does open up the potential to waste memory if the data being cached does not fit perfectly
into the slab.

There are also some practical limits to be aware of concerning key and data size limits. For
example, keys are restricted to 250 characters and cached data cannot exceed the largest
slab size currently supported, 1 megabyte.

Copyright © 2008, MySQL AB Page 5

The hashing algorithm that memcached makes use of to determine which server a key is
cached on does not take into account the size of all the memory available across all the
servers participating in a memcached cluster. For those concerned about this issue, a
potential workaround is to run multiple memcached instances on servers with more memory
with each instance using the same size cache as all the other servers.

Caching Structure
When memcached’s distributed hash table becomes full, subsequent inserts force older
cached data to be cycled out in a least recently used (LRU) order with associated expiration
timeouts. How long data is “valid” within the cache is set via configuration options. This
“validity” time may be a short, long or permanent. As mentioned, when the memcached
server exhausts its memory allocation, the expired slabs are cycled out with the next oldest,
unused slabs queued up next for expiration.

Memcached makes use of lazy expiration. This means it does not make use of additional
CPU cycles to expire items. When data is requested via a ‘get’ request, memcached
references the expiration time to confirm if the data is valid before returning it to the client
requesting the data. When new data is being added to the cache via a ‘set’, and memcached
is unable to allocate an additional slab, expired data will be cycled out prior to any data that
qualifies for the LRU criteria.

Caching Database Queries

One of the most popular use cases for leveraging memcached is to cache the results of
database queries. Below is a simple outline of how the ‘add’ and ‘set’ memcached functions
can be leveraged in this respect.

• The server first checks whether a memcached value with a unique key exists, for
example “user:userid”, where userid is a number.

"SELECT * FROM users WHERE userid = ?"

• If the result is not cached, the request will issue a select on the database, and set
the unique key using the memcached ‘add’ function call.

• If this call was the only one being altered the server would eventually fetch incorrect
data, so in addition to using the ‘add’ function, an update is also required, using the
‘set’ function.

• This ‘set’ function call updates the currently cached data so that it is synchronized
with the new data in the database. (Another method for achieving a similar behavior
is to invalidate the cache using the ‘delete’ function so that additional fetches result
in a cache miss forcing an update to the data.)

At this point, whenever the database is updated, the cache also needs to be updated in
order to maintain the desired degree of consistency between the cache and the source
database. This can be achieved by tagging the cached data with a very low expiration time.
However, this does mean that there will be a delay between the update occurring on the
database and the expiration time on the cached data being reached. Once the expiration
time is reached, a subsequent request for the data will force an update to the cache.

Copyright © 2008, MySQL AB Page 6

You may discover during testing or benchmarking that memcached may not yield faster
performance than simply running queries on the source database. The perspective to keep
in mind is that memcached is designed to assist in scaling the database, so as connections
and requests to the source database ramp up, memcached will assist in alleviating load by
handling read requests.

Data Redundancy and Fail Over

By design, there are no data redundancy features built into memcached. Memcached is
designed to be a scalable and high performance caching-layer, including data redundancy
functionality would only add complexity and overhead to the system.

In the event one of the memcached servers does suffer a loss of data, under normal
circumstances it should still be able to retrieve its data from the original source database. A
prudent caching design involves ensuring that your application can continue to function
without the availability of one or more memcached nodes. Some precautions to take in
order not to suddenly overwhelm the database(s) in the event of memcached failures is to
add additional memcached nodes to minimize the impact of any individual node failure.
Another option is to leverage a “hot backup”, in other words a server that can take over the
IP address of the failed memcached server.

Also, by design, memcached does not have any built in fail over capabilities. However, there
are some strategies one can employ to help minimize the impact of failed memcached
nodes.

The first technique involves simply having an (over) abundance of nodes. Because
memcached is designed to scale straight out-of-the-box, this is a key characteristic to
exploit. Having plenty of memcached nodes minimizes the overall impact an outage of one
or more nodes will have on the system as a whole.

One can also remove failed nodes from the server list against which the memcached clients
hash against. A practical consideration here is that when clients add or remove servers from
the server list, they will invalidate the entire cache. The likely effect being that the majority
of the keys will in-turn hash to different servers. Essentially, this will force all the data to be
re-keyed into memcached from the source database(s). As mentioned previously,
leveraging a “hot backup” server which can take over the IP address of the failed
memcached server can go a long way to minimize the impact of having to reload an
invalidated cache.

Loading and Unloading Memcached

In general, “warming-up” memcached from a database dump is not the best course of
action to take. Several issues arise with this method. First, changes to data that have
occurred between the data dump and load will not be accounted for. Second, there must be
some strategy for dealing with data that may have expired prior to the dump being loaded.

In situations where there are large amounts of fairly static or permanent data to be cached,
using a data dump/load can be useful for warming up the cache quickly.

Copyright © 2008, MySQL AB Page 7

Memcached Threads
With the release of version 1.2, memcached can now utilize multiple CPUs and share the
cached between the CPUs engaged in processing. At its core is a simple locking mechanism
that is used when particular values or data needs to be updated. Memcached threads
presents considerable value to those making use of ‘multi-gets’, which in-turn become more
efficient and makes memcached overall easier to manage, as it avoids having to run
multiple nodes on the same server to achieve the desired level of performance.

Memcached Clients
In a typical memcached cluster, the application will access one or more memcached servers
via a memcached client library. Memcached currently has over a dozen client libraries
available including, Perl, PHP, Java, C#, C/C++, Lua and a native MySQL API. A complete
list of client APIs can be found at:

http://www.danga.com/memcached/apis.bml

Supported Client Libraries

It is important to note that there are different memcached client implementations. Not only
does the manner in which they store data into memcached vary between client libraries, but
also in how they implement the hashing algorithm. Because of this, the implications of
mixing and matching client libraries and versions should be carefully considered. However,
unlike the variances found between client versions, memcached servers always store data
and apply hashing algorithms consistently.

From a security standpoint, it does bear mentioning that memcached does not posses any
authentication or security features. Best practices in this area would dictate that
memcached should only be run on systems within a firewall. By default, memcached makes
use of port 11211.

PHP PECL
For information concerning the PHP extensions which allows for Object-Oriented and
procedural interfaces to work with memcached, please see:

http://pecl.php.net/package/memcache

Java Client Libraries

Two Java APIs for memcached are currently in development:

Memcached Client for Java

http://www.whalin.com/memcached/

Spymemcached – Java Client for Memcached

Copyright © 2008, MySQL AB Page 8

http://code.google.com/p/spymemcached/

Python Client Library

A 100% Python interface to memcached. For more information, please see:

ftp://ftp.tummy.com/pub/python-memcached/

Perl Client Library

A Perl client library for memcached. For more information, please see:

http://search.cpan.org/dist/Cache-Memcached/

C Client Library
There are currently three C libraries for memcached. For more information, please see:

apr_memcache

http://www.outoforder.cc/projects/libs/apr_memcache/

libmemcached

http://tangent.org/552/libmemcached.html

libmemcache

http://people.freebsd.org/~seanc/libmemcache/

MySQL API
The memcache_engine allows memcached to work as a storage engine to MySQL. This
allows SELECT/UPDATE/INSERTE/DELETE to be performed from it were a table in MySQL.

http://tangent.org/index.pl?node_id=506

Also, there is a set of MySQL UDFs (user defined functions) to work with memcached using
libmemcached.

http://tangent.org/586/Memcached_Functions_for_MySQL.html

Example Architectures
There are various ways to design scalable architectures using memcached and MySQL.
Below we illustrate several of these architectures.

Memcached & MySQL

Copyright © 2008, MySQL AB Page 9

In this architecture we combine several memcached servers and a stand-alone MySQL
server. This architecture allows for scaling a read intensive application. The memcached
clients are illustrated separately; however, they typically will be co-located with the
application server.

read write

Web Servers

Application
Servers

Memcached
Clients mc mc mc

Memcached
Servers ms ms ms

cache update
MySQL Server

Figure 2: Multiple Memcached Servers and a Stand-Alone MySQL Server

Scale-Out: Memcached & MySQL Replication

In this architecture we combine several memcached servers with a master MySQL server
and multiple slave servers. This architecture allows for scaling a read intensive application.
The memcached clients illustrated are co-located with the application servers. Read
requests can be satisfied from memcached servers or from MySQL slave servers. Writes are
directed to the MySQL master server.

Copyright © 2008, MySQL AB Page 10

 read from cache write read from slave

Application Servers
&
Memcached Clients mc mc mc

Memcached
Servers ms ms ms

MySQL
Master cache update

MySQL
Replication

MySQL
Slaves

Figure 3: Multiple Memcached Servers with a Master and multiple Slave MySQL Servers

Memcached, Sharding & MySQL Replication

With sharding (application partitioning) we partition data across multiple physical servers to
gain read and write scalability. In this example we have created two shards partitioned by
customer number. Read requests can be satisfied from memcached servers or from MySQL
slave servers. Writes are directed to the appropriate MySQL master server.

Copyright © 2008, MySQL AB Page 11

 read from cache write read from slave

mc mc mc

shard 1 (cust 1-999) shard 2 (cust 1000-1999)

read from cache read from cache

ms ms ms ms ms ms

cache update write cache update

write

MySQL MySQL
Replication Replication

read from slave read from slave

Figure 4: Sharding, multiple Memcached Servers with a Master and multiple Slave MySQL
Servers

Solutions: Memcached for MySQL

The MySQL Enterprise subscription now includes 24x7 production support for the
memcached distributed memory caching system. Support for memcached is available at the
Gold and Platinum subscription levels.

Memcached for MySQL enables organizations to:

• Implement a scalable, high performance data caching solution for their online
applications
• Reduce database Total Cost of Ownership (TCO) by eliminating licensing costs for
proprietary data caching software
• Reduce system TCO by making better use of resources like idle or spare RAM on
existing systems
• Incrementally add/remove data caching capacity, on-demand to quickly meet
changing requirements.

For more information about MySQL Enterprise, please visit:

http://www.mysql.com/products/enterprise/features.html

Copyright © 2008, MySQL AB Page 12

High Performance Data Caching
Memcached enables organizations to increase the scalability of new and existing MySQL
applications. Memcached combined with MySQL Replication, is an excellent solution for
improving application performance and leveraging scale out architectures at the same time.

On-Demand Scalability
Memcached can be incrementally scaled out in an on-demand fashion. Because Memcached
can scale to support dozens of nodes with minimal overhead, anywhere spare memory
resources exist is an opportunity to scale your application even further. Memcached is
designed as a non-blocking event-based server with no special networking or interconnect
requirements.

Lower TCO with Open Source

Delivering a scalable, high performance data caching solution based on open source
components is a cost-effective alternative to expensive proprietary software. Memcached is
a great way to protect yourself from proprietary vendor lock, which can result in
unpredictable costs as the size of your application grows and the number of users increases
exponentially. MySQL and Memcached can be implemented with confidence as they both
benefit from a large community of developers and users ensuring they stand up even under
the most extreme production loads. Further, memcached gives you the benefit of caching
more then just result sets. For example, preprocessed HTML can be cached on the same
Memcached nodes, further improving the performance of your web application.

24x7 Support from MySQL

Organizations can get the professional help they need from the Memcached and MySQL
experts at Sun Microsystems

Professional Consulting Services

With extensive memcached and MySQL knowledge plus the hands-on experience to
implement scalable, high performance solutions for MySQL customers, our consultants use
proven methodologies and expertise in Database Design, Architecture, Performance Tuning,
Replication, Fail-Over and Fault-Tolerance.

For more information about consulting services, please visit:

http://www.mysql.com/consulting/

Conclusion
In this whitepaper we have provided a basic introduction to memcached and how it can
improve application performance while minimizing database load. Memcached increases the
scalability of dynamic, data-driven applications by increasing the possibility of supporting
more concurrent users and spikes in traffic, while at the same time making the most
efficient use of existing computing infrastructure. We explored how the memcached server
allocates memory, implements hashing and interacts with memcached clients. We have also

Copyright © 2008, MySQL AB Page 13

illustrated several scalable architectures which combine memcached, MySQL and
Replication. Finally, we presented Memcached for MySQL which is a standard part of a
MySQL Enterprise Gold or Platinum subscription. Not only does a MySQL Enterprise
subscription get you access to 24x7 production support from MySQL and memcached
experts from Sun Microsystems, it also gives you access to the MySQL Enterprise Server
and MySQL Enterprise Monitor software.

Additional Resources

White Papers
http://www.mysql.com/why-mysql/white-papers/

Case Studies
http://www.mysql.com/why-mysql/case-studies/

Press Releases, News and Events

http://www.mysql.com/news-and-events/

Live Webinars
http://www.mysql.com/news-and-events/web-seminars/

Webinars on Demand
http://www.mysql.com/news-and-events/on-demand-webinars/

Download Memcached
http://www.danga.com/memcached/download.bml

Installation Tutorial
http://blog.ajohnstone.com/archives/installing-memcached/

Memcached for MySQL Reference Manual

http://dev.mysql.com/doc/refman/5.1/en/ha-memcached.html

Memcached FAQ
http://www.socialtext.net/memcached/index.cgi?faq

Memcached for MySQL Forums

http://forums.mysql.com/list.php?150

Copyright © 2008, MySQL AB Page 14

About Sun’s MySQL Portfolio
The MySQL product portfolio is the most popular open source database software in the
world. Many of the world's largest and fastest-growing organizations use MySQL to save
time and money powering their high-volume Web sites, critical business systems and
packaged software -- including industry leaders such as Yahoo!, Alcatel-Lucent, Google,
Nokia, YouTube and Zappos.com. At http://www.mysql.com, Sun provides corporate users
with commercial subscriptions and services, and actively supports the large MySQL open
source developer community.

To discover how Sun’s offerings can help you harness the power of next-generation Web
capabilities, please visit http://www.sun.com/web .

Copyright © 2008, MySQL AB Page 15