16 Jul 2012: Really Simple Systems, Europe’s largest provider of hosted CRM, has delivered 99.999% uptime over the last 12 months, avoiding the outages that have plagued rival vendors. To achieve 99.999% uptime means having no more than five minutes’ downtime a year, with 99.99% requiring no more than one hour downtime a year.

A year ago, the company implemented an innovative delivery platform across two data centres using ‘Master-Master’ database replication, with customers’ load balanced across both systems. The result is that if a data centre fails, customers on that system are seamlessly switched across to the other data centre, ensuring zero downtime and no data loss.

Data is replicated in real-time and is automatically exchanged between the two data centres, ensuring both systems are constantly up to date even if one fails or requires a restart.

Over 5,000 users from 2,000 organisations use the cloud-based CRM sales, support and marketing system supplied by Really Simple Systems, with the success of their own essential business processes relying heavily on the hosted services achieving maximum uptime and optimum efficiency.

Providing services that are business critical and playing a key role in its customers’ business continuity planning is what led Really Simple Systems to explore the measures it could put in place to substantially reduce the risk of unplanned downtime and ensure the smooth running of its customers’ CRM operations.

John Paterson, CEO of Really Simple Systems, explained: “All data centres go down sometimes, so the only way to achieve 99.999% uptime is to have two completely separate data centres, physically situated far away from each other with different electricity and internet suppliers, operating in Live-Live mode and replicating data in real-time. Having a standby system that can be switched over to in a few minutes is better than relying upon just one instance, but to get to 99.999% you need to be able to switch instantly.

“Most database management systems use database replication so that services can continue as usual in the event of a system failure; a back-up system will step in and pick up where the primary system left off with the aim of minimising downtime while repairs or restarts are taking place. Most commonly used in database replication is a ‘Master-Slave’ system structure as it is relatively simple to configure; however, it also has its disadvantages. Replicated data can only be automatically transferred in one direction, from Master to Slave, so any new data created while the Slave is acting as the primary system will not appear on the Master system when it operational, unless it is manually updated. In addition, it is usual for there to be a delay before the Slave steps in, resulting in business downtime for users of the system.”