Web Cache Service
Emencia deals with the performance of your web service
The web cache allows for optimal and more rational use of resources, and therefore an improvement in web server performance, in particular via a more rapid display of web pages.
For what kind of use?
- For any site that receives a large number of visits: the hardware resources and bandwidth will be used less, better optimized and the pages more rapidly displayed
- An event that causes the load on your web site to rise (radio ad, online poster campaign, TV ad, a fair, etc.)
- For any institutional or e-commerce site whose traffic volume is significant: you can avoid losing customers and prospects due to excessively long display response times
- For any service that consumes a large quantity of server resources (cpu, ram, I/O) to deliver ordinary content
- To improve your search engine visibility: Google penalizes your natural search engine optimization if the response time of your pages is too long.
A web service to better optimize your resources.
Operating Procedure
When a user makes a query using his web browser, the cache service will process it by checking its server storage space (or even other cache servers) to see whether this page indeed exists.
- If the server does not find the page requested or does consider it to be up to date, it will send the query back to the original server. This verification takes a few milliseconds at most.
- However, if the cache service finds the information requested, it will send the user directly to the page requested, while saving it in memory for a given period (the period can be configured depending on the page or object in the code).
The response to the next query related to this page will therefore take less time.
Advantages
- Rapid delivery of content to a very large number of users
- Limited, optimized used of the bandwidth
- Reduction in response time to a web query
- Reduction in the consumption of server resources
- Concealing of cuts on the master platform
- Less jams within the internal network
- Reduction and postponement of server investments
- Improvement in natural SEO thanks to a better response time
"Pimp my server"architecture
Cluster Mode
The Django and PostgreSQL hosting solution is deployed on a single server, but designed to distribute the load over a cluster if necessary.
All the components can be parallelized:
- The PostgreSQL database can use a master/slave replication with a number of slave servers in order to distribute the load for complex queries.
- The technical Django platform is perfectly suited to parallelization via an HTTP software load balancer, such as HAProxy or an A10, Brocade or Alteon hardware load balancer.
- The Django cache mechanism uses distributed systems, such as memcached.
The cluster mode is not necessary for the initial volume measurement, but is possible if the Customer wants to add a large number of users and guarantee the same quality of service (availability and response time).
If you are interested, please contact our technical sales department:
Roger Fernandez or David Robinet: +33 (0)1 47 20 23 01