Don't miss out Virtual Happy Hour this Friday (April 26).

Try our conversational search powered by Generative AI!

Per Ivansson
Oct 30, 2008
  9251
(0 votes)

Distributed Memory Caching in EPiServer Community using Velocity

It´s the last day of PDC2008 and I've been to a lot of sessions by now. The Cloud is everywhere... One session I really think stood out was the one about the Velocity project. We have looked on Velocity, Memcached and other similar distributed memory caching systems for a while now as one comparatively low investment to improve performance and scalability of EPiServer Community. One benefit of having a distributed "central" cache instead of, or in addition to, having local caches on each web server is that it reduces database calls when the cache is invalidated and needs to re-populate from persisted data.

For example: Say that you have five web servers in your cluster. When a cache is invalidated, it propagates throughout the cluster and one separate call from each web server has to made to the database to populate the local cache again. That is five roundtrips to the database to retrieve the exact same data. The more web servers in this setup, the higher the load on the database server. With a distributed cache, the distributed cache would populate itself from the persisted data only once, and concurrent requests from the web servers would use that cache instead of query the database themselves.

Velocity have some nice features beyond the basics. For example: It's fully scalable. You may just add machines as you go along as the sync between them is handled automatically. You can configure consistency vs availability of data by defining number of fail over nodes etc.. Another cool thing is that it supports tagging of cached data so that you retrieve collections of cached objects by their tagging. Even cooler is that is supports Linq, so you can actually write queries that filter and returns data directly from the cache. Everything can of course operate in that Cloud of theirs...

A fully scalable cache repository that you can query using Linq, opens quite interesting possibilities for high transactional applications such as EPiServer Community and we will surely look into this in the very near future. The current state of Velocity is CTP2. CTP3 will be released Q1 2009 and the first release is expected Q2/Q3 2009.

Oct 30, 2008

Comments

Please login to comment.
Latest blogs
Solving the mystery of high memory usage

Sometimes, my work is easy, the problem could be resolved with one look (when I’m lucky enough to look at where it needs to be looked, just like th...

Quan Mai | Apr 22, 2024 | Syndicated blog

Search & Navigation reporting improvements

From version 16.1.0 there are some updates on the statistics pages: Add pagination to search phrase list Allows choosing a custom date range to get...

Phong | Apr 22, 2024

Optimizely and the never-ending story of the missing globe!

I've worked with Optimizely CMS for 14 years, and there are two things I'm obsessed with: Link validation and the globe that keeps disappearing on...

Tomas Hensrud Gulla | Apr 18, 2024 | Syndicated blog

Visitor Groups Usage Report For Optimizely CMS 12

This add-on offers detailed information on how visitor groups are used and how effective they are within Optimizely CMS. Editors can monitor and...

Adnan Zameer | Apr 18, 2024 | Syndicated blog

Azure AI Language – Abstractive Summarisation in Optimizely CMS

In this article, I show how the abstraction summarisation feature provided by the Azure AI Language platform, can be used within Optimizely CMS to...

Anil Patel | Apr 18, 2024 | Syndicated blog

Fix your Search & Navigation (Find) indexing job, please

Once upon a time, a colleague asked me to look into a customer database with weird spikes in database log usage. (You might start to wonder why I a...

Quan Mai | Apr 17, 2024 | Syndicated blog