Category: System Architecture

Varnish Cache – You should seriously take a look

Varnish Cache

Over the last few weeks, we have been testing the use of Varnish  on one of the web sites I work on – and the results appear to be fantastic – with a few false starts in the configuration.  While I am not the one in the trenches configuring it – the potential performance and control we will have over our previous caching methods is quite frankly amazing.

If I could give you stats or more information I would, but at this time it is not a good idea.

You should really check it out – amazing results.

Video: Battle of the NoSQL Stars

Alright I will admit that I have yet to view this video, but it looks quite interesting.  The source is – one of my favorite sources and it is a subject I am starting to get fascinated with.  I think there is some real potential here for at least one of the systems I work with at my job.  The only problem is that there is a pretty significant cost to making the change.

I am curious if anyone out there has been using any of these technologies and what successes they may have had.

Surviving Apache Server Overload - Apache Server Overload

Apache Server overload – sometimes it feels like I spend half my working week dealing with this issue.  Mockyblog has written up a handy little summary of  how to approach this issue on small Linux servers.  I think following Alex’s steps as a first line of fire in these situations would be very good advice.  The article will also teach you some pretty basic lines of attack and methods of determining what exactly is going on with your server.  I have used most of his suggestions in the past, but the post did teach me a couple things that I did not know before.

I would say this is must reading for those just getting started in the hosting world.

MySQL Cluster – Time to test?

A friend just sent me a link to MySQL Cluster Sandbox and it looks pretty interesting.  I work on a site that is in the middle of needing bleeding edge database architecture and a good run-of-the-mill redundant system.  Historically we have simply thrown more hardware at the problem, but it seems like we have hit our limit.

While the system is far more read intensive than write – the writes are starting to grow.  Replication is fine and general caching is working to some degree, but it is not solving our problems.  Memcached is being considered, but I am not really sure that is going to solve all of our issues, especially with our projected growth.

I ramble – to tired and have a big presentation in the morning…