16 Mar

Zignal just scales to smithereens

Todays blog will be on scalability and the scaling of Zignal. In general there is always some talk about scalability and scaling of solutions from the non-Microsoft/IBM players. The solutions the big players most often come up with are to deploy solutions on mammoth application servers, databases and hardware.

We like to think about light app servers and flexible hardware as a positive solution for scaling. This thinking is more in line with how Google and Amazon are scaling their web platform. Also, you wake up one day to learn about successful platforms like Facebook, that are built on PHP and MySQL and serving more users than anyone could have ever dreamed of a few years back.

If I quote the Industria CTO on whether Zignal scales, his answer is simply, “yes, it scales to smithereens”! Furthermore, I can touch upon 4 key points on this as a good starter;

  1. We have lots of logic running on the actual end-user device, including data caches. This means that the client does not need to make too many requests on the server. In particular, EPG data is delivered in a compact form and up to two weeks worth of EPG data is cached by the client. The client then does not need to re-fetch the same data to be displayed again or in a different form.
  2. We have engineered the server using powerful OS Java tools like Spring and Hibernate that have been used for many large scale web- and transaction systems. By utilizing these tools we can scale the server to handle tens, even hundreds of thousands of users on single, powerful server hardware.
  3. The Zignal database structure is engineered to handle most data in local caches on the Zignal Server. These means we can scale horizontally on many servers without adding much load on the database or the database cluster.
  4. We are beginning to use the Terracotta clustering solution to distribute higher-level structures, rather than merely caching database records. This should lead both to higher throughput per Zignal Server in the cluster, and less load on the database per Zignal Server.

Yes, Zignal can scale for the millions of users and what we have now is a great start!

Related Content from the Web

One Comment

  1. 2
    margo
    May 23, 2008 at 05:46
    Permalink

    Impressive methods for scaling.

One Trackback

  1. May 15, 2008 at 20:21

    Restoration Hardware Supply…

    I found your site on technorati and read a few of your other posts. Keep up the good work. I just added your RSS feed to my Google News Reader. Looking forward to reading more from you….

Add Comment

Your email is never published nor shared. Required fields are marked *

*
*