Building a 24x7 non-stop service
As most of our customers will be aware, we have always delivered a highly available, multi-data centre, disaster-tolerant Secure Authentication Service. Over the past 4.5 years, we have achieved a remarkable 99.9999% availability of our RSA SecurID Authentication Service. In other words, our core service has been continuously available to your users for all but two of the 2.5 million minutes since October 2002!
But we want to do even better!
We are constantly striving to deliver genuine, 100%, 24x7 service availability: there are users of the service all over the world, and someone is logging in with their Signify credentials from somewhere, every minute of the day and night, every day of the year. The rapid growth in our international customer base also means that administrators are using the IMC around the clock to add users and manage tokens.
Last year we set an objective to enhance the availability and uptime of the IMC management portal to match the core authentication service, however we ran into constraints imposed by the architecture of the RSA SecurID servers which cannot accept administrative changes to user data when in 'fail-over' mode. So our infrastructure team, led by our CTO Liam Crilly, designed an entirely new service architecture which allows our Service Operations team to effect an immediate and complete systems switch-over to our replica authentication servers and IMC systems at our secondary data centre.
This system was built and rigorously tested during 2006 and went live at the end of the year. It allows us to continue to deliver a complete, fully manageable service even during a major data centre outage, and it also allows our engineers to perform all systems maintenance and upgrades without impacting any of our users or administrators.
So now, during scheduled maintenance windows, and the rare occasions when we have to invoke our Disaster Recovery procedures, the service still delivers full functionality and complete manageability. Your administrators can continue to add users, re-allocate tokens and perform all their day-to-day management tasks, while your users can change their PINs, invoke Emergency Access, and request the transmission of one-time passcodes, completely unaware that there may be anything amiss.
The nice thing is that delivering a non-stop, secure authentication service is one thing that you simply don't have to worry about, but we thought you'd like to know that we have been doing all the worrying on your behalf!