Concur's Journey to 1 Billion Couchbase Operations Per Day

Concur is a leading provider of integrated travel and expense management solutions for companies of all sizes. Initially, Concur utilized a homegrown caching system that turned out to be expensive and complicated to scale. They wanted to replace it with an open source solution that could help laterally scale their application, and provide the resilience and redundancy they needed – with simplicity and efficiency. They also wanted the new solution to support their aggressive plans for using NoSQL technology extensively throughout their stack in the future. 

Situation

Concur is a leading provider of integrated travel and expense management solutions for companies of all sizes. As a leading SaaS provider, Concur constantly seeks to improve scalability and performance across their architecture and infrastructure stack – especially the database and storage tiers.

Initially, Concur utilized a homegrown caching system that turned out to be expensive and complicated to scale. They wanted to replace it with an open source solution that could help laterally scale their application, and provide the resilience and redundancy they needed – with simplicity and efficiency. They also wanted the new solution to support their aggressive plans for using NoSQL technology extensively throughout their stack in the future.

Solution

Concur implemented a large Couchbase Server cluster, utilizing the Enyim (.NET) client for the web tier and the spymemcached (Java) client for the middle tier. They were impressed with the simplicity with which they could setup, configure and scale the cluster.

For example, the ability to use a single solution across multiple tiers and languages was a huge win for Concur and made Couchbase a more attractive option than other solutions, which required different implementations for each. Using the recommended reverse proxy setup, there’s just a single IP address for configuring REST connections and monitoring across multiple tiers.

Couchbase’s cluster management capabilities give Concur the ability to scale out their cluster with zero downtime. They also get the added benefit zero configuration management when adding or removing nodes from their cluster because both the .NET and Java clients are automatically updated when there is a topology change, ensuring smooth and consistent growth of the Couchbase cluster across the entire environment.

Results

The Concur team is very happy with the results of moving to Couchbase. Couchbase technology is incredibly resilient, cost effective, and fast – and the Couchbase team has provided excellent support for them when needed.

With the first phase of the Couchbase project complete, initial tests showed a 20 percent increase in application performance thanks to Couchbase’s low latency and sub-millisecond response times. Today, Concur processes more than a billion Couchbase operations per day (see slides). 

Deploying Couchbase solutions gives us greater scalability and blazing performance (sub-millisecond response times). We now have the ability to simply and fluidly increase capacity, enabling us to seamlessly respond to the needs of the application.”

– Drew Garner, Concur Director of Architecture Services

Concur is a leading provider of integrated travel and expense management solutions. Their cloud services process over $50 billion in travel and expense spend per year for global corporations.

 Concur was founded in 1993 on the premise of helping to drive costs out of business through innovation.