Virtual Service and Real Server availability explained.
One of the most fundamental components of Load Balancing or ‘Reverse-Proxy’ing is a Virtual Service. This may be sometimes referred to as a Virtual Server. The Virtual Service is an object created that accepts requests and balances these across configured Real Servers. This ensures an application can scale and grow through adding additional Real Servers and can have redundancy in case one server fails.
As well as Load Balancing the Load-balancer will typically be configured to perform Health Checking of Real Servers ensuring only Real Servers that are healthy receive requests.
If one of the Real Servers fails the health check this will result in the Real Server being marked as down and requests are then sent to the remaining available Real Servers.
If all Real Servers are down this means the entire Virtual Service will be marked as down and no new requests will be received.
It is important to understand the difference between these two scenarios.
Where a Real Server is down, users should still be able to access the Application, assuming there is enough capacity provided by the remaining Real Server. However when a Virtual Service is down this means no request will be accepted.
The key to ensuring 100% uptime is that any Real Server issues are dealt with quickly before any user impact is experienced.