Processing AAS Models Asynchronously

Analysis Services normally requires a live connection between the client and server. If the connection is interrupted, the server considers that the client has lost interest and cancels the running operation.Strictly speaking, there is  a ‘long running operations’ (LRO) protocol built over the HTTP XMLA messaging protocol which mitigates the issue of disconnects by attempting to keep alive the request/responses. This helps avoid the HTTP disconnection problem for the client libraries and improve the reliability of client applications. But in the end, your mileage will vary depending on the reliability of the networks in between the client machine and the service endpoint.

Read more...