Update (14-Mar-2014) >
The recording and presentation is now available on Channel 9 - http://channel9.msdn.com/Events/SharePoint-Conference/2014/SPC312
< Update (14-Mar-2014)
I've been working on this topic for sometime now, and I'm happy that the solution is ready to be shared with you. I'll be presenting a session on this at SharePoint Conference 2014
Here's the abstract of the session
Public facing websites of large organizations generally have very strict Service Level Agreements (SLA) for parameters related to Business Continuity Management (BCM) such as High Availability (HA) and Disaster Recovery (DR). Such SLAs can’t be met with just with the HA capabilities of a SharePoint farm or 99.95% uptime offered by Azure Virtual Machines (VMs) using availability sets.
In this session we will discuss how to meet such demands by implementing two SharePoint farms on two different Azure IAAS data centres (DC) in hot stand-by DR mode which can failover within minutes. We’ll do a demo of the auto-failover feature on live SharePoint environments.
We will explain how we used a combination of technologies, SQL Server Always On, Log shipping, Azure Blob storage, and Azure Traffic Manager (ATM) to build the above solution.
Geo-redundancy in SharePoint is a relatively less explored area even in the on-premise world. Azure IaaS adds some more challenges such as non-availability of cross-datacenter VNET or AD domain. The session provides information on how to address these challenges and will lay out the available choices on techniques, consequences of the choices and hidden issues to watch out for in each of these aspects. The session is based on real life scenarios from two large customers with public web sites on SharePoint.