6

I have installed redis on my server and implemented object caching for data returned within a PHP based web application. The php model essentially executes a reasonably complex query and returns a detailed array of data. I have tested the caching and found everything to be working as expected. I first check to see if the key exists in redis. If it does, redis returns the data, the model unserializes and returns the previously cached data. If the cache has expired, the model executes the sql query, returns the data and sets the key and serialized value in redis.

So here are my questions.

  1. I'm not sure how to really benchmark this as it is all browser based. What tools are there out there that would allow me to get a reasonable benchmark to compare caching and not caching. I'm thinking of perhaps a php script that calls the api 1000 times via curl.

  2. I implemented this in redis because I once read that caching with redis will work across multiple sessions or ip addresses accessing the site. For example, if the api is accessed 1000 time an hour by multiple ip addresses/users, I am assuming this approach will reduce the load on the mysql server and let redis do the work of returning the cached data instead. Can anyone shed some light on this? Are my assumptions valid?

All comments are welcome!

Thanks!

Dave

Improve this question

2 Answers 2

Reset to default
8

To benchmark the web site, I would use something like Siege rather than writing a specific PHP script.

Regarding Redis usage, caching things in in-memory stores like memcached or Redis is now very common. Both memcached and Redis are suitable for this purpose, although for pure caching, memcached is arguably easier to setup. 1000 times an hour represents only 3.6 TPS - any data store (including MySQL) will support such traffic without any issue. Now, multiply this traffic by 100 or 1000, and the caching layer (memcached or Redis) will become mandatory to protect your database.

To use Redis for caching, you may want to check the EXPIRE command and have a look at the maxmemory-policy parameter in the configuration file.

Improve this answer
Sign up to request clarification or add additional context in comments.

1 Comment

Didier - Than you for your response! I will definitely take a look at Siege. You are correct about the TPS - I just wanted to set it at a reasonable amount for benchmark testing to see the effect of caching. In the meantime I also started trying out APC and am having good success with that as I see it can also be used for object/data caching along with opcode caching. I can experiment with both and try out Siege as the benchmarking too. On another note, I am new here and tried to give you an Up Vote but i am not permitted yet.
0

I have done extensive testing of cache backends for the Zend_Cache library. The tests were done using multiple php-cli processes and randomized data and considered read performance, write performance and cache tag cleaning performance. If testing just the cache backend the web server performance is not relevant so I recommend testing via CLI to simplify the testing. Also, testing with only one process will not give you an accurate picture of a backend's characteristics under heavy load.

MySQL is very fast itself and if you are doing single-record indexed queries then MySQL's own query cache will be very fast. I'd only recommend adding an additional caching layer for things that are slow (aggregated results of multiple queries or generating chunks of HTML). You can use Zend_Cache without including the entire Zend framework so I highly recommend you check out both Cm_Cache_Backend_Redis and Cm_Cache_Backend_File.

Improve this answer

Comments

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.