Skip to main content

Storing large values in Memcached with PHP

Memcached saved my users a minute per query
I'm working on a business intelligence tool that requires as an intermediary calculation a list of UK postcodes that are within a radius of a user supplied postcode.

It currently takes about 7 seconds to query my Postgres database to get this list out.  Unfortunately I need to do this several times as part of a goal seeking function so I need to greatly improve this lookup speed.

I'm already using the Postgres earthdistance module and have properly indexed my table so I realized that I needed to look for a caching solution.

Memcached places limits on the size of the value you can store.  The default setup is 1meg and I'm reluctant to change this because it adds to the deployment burden.  My result sets were sometimes up to 4 megs large - searching on a 20 mile radius in London yields a lot of postcodes!

My idea was to split the large piece of data into several smaller pieces and to place an index referencing the pieces as the value for the key we're trying to store.

I decided to make use of PHP's gzcompress() function to reduce the size of the element because I felt that the time I spend compressing the data is still going to be drastically less than running the query and I want to try my best to avoid cache evictions.

I'm currently using Laravel so the code snippets below use the facades made available by Laravel.  I think the code is readable enough to extend to other PHP environments and I think the approach could also be ported to other languages.




Labels:

Comments

Post a Comment

Popular posts from this blog

Using Fail2Ban to protect a Varnished site from scrapers

I'm using Varnish to cache a fairly busy property site.  Varnish works like a bomb for normal users and has greatly improved our page load speed. For bots that are scraping the site, presumably to add the property listings to their own site, though the cache is next to useless since the bots are sequentially trawling through the whole site. I decided to use fail2ban to block IP's who hit the site too often. The first step in doing so was to enable a disk based access log for Varnish so that fail2ban will have something to work with. This means setting up varnishncsa.  Add this to your /etc/rc.local file: varnishncsa -a -w /var/log/varnish/access.log -D -P /var/run/varnishncsa.pid This starts up varnishncsa in daemon mode and appends Varnish access attempts to /var/log/varnish/access.log Now edit or create /etc/logrotate.d/varnish and make an entry to rotate this access log: /var/log/varnish/*log { create 640 http log compress ...
Post a Comment
Read more

Translating a bit of the idea behind domain driven design into code architecture

I've often participated in arguments discussions about whether thin models or thin controllers should be preferred.  The wisdom of a thin controller is that if you need to test your controller in isolation then you need to stub the dependencies of your request and response. It also violates the single responsibility principal because the controller could have multiple reasons to change.   Seemingly, the alternative is to settle on having fat models. This results in having domain logic right next to your persistence logic. If you ever want to change your persistence layer you're going to be in for a painful time. That's a bit of a cargo cult argument because honestly who does that, but it's also a violation of the single responsibility principal.   One way to decouple your domain logic from both persistence and controller is to use the "repository pattern".   Here we encapsulate domain logic into a data service. This layer deals exclusively with imple...
Post a Comment
Read more

Grokking PHP monolog context into Elastic

An indexed and searchable centralized log is one of those tools that once you've had it you'll wonder how you managed without it.    I've experienced a couple of advantages to using a central log - debugging, monitoring performance, and catching unknown problems. Debugging Debugging becomes easier because instead of poking around grepping text logs on servers you're able to use a GUI to contrast and compare values between different time ranges. A ticket will often include sparse information about the problem and observed error, but if you know more or less when a problem occurred then you can check the logs of all your systems at that time. Problem behaviour in your application can occur as a result of the services you depend on.  A database fault will produce errors in your application, for example. If you log your database errors and your application errors in the same central platform then it's much more convenient to compare behaviour between...
Post a Comment
Read more