Metering, billing, customer management, using the runtime API, extending SaaSGrid, and more. Written by our Client Services team.
In SaaSGrid 1.9, there are four tiers to consider in high availability scenarios:
(SQL Server)
High availability in the data tier is achieved using SQL Server Clustering, SQL Server Replication, and/or third-party (such as Xkoto) SQL Server clustering & replication tools.
(WCF)
In 1.9, this is SaaSGrid’s most robust, automated tier. SaaSGrid provides a high-performance WCF self-hosting service container. WCF services are deployed in availability groups according to deployment strategies, either out-of-the-box or developer-provided. WCF services may be deployed and balanced based on a variety of considerations, including server capacity and request load.
(IIS)
In 1.9 SaaSGrid offers front end load distribution out-of-the-box; greater detail is available in the Load Balancer Setup document. Front end load balancing on the other hand must be implemented using Microsoft or other third party IIS-compatible or -integrated (Citrix, F5, etc.) modules and tools. Purely within the Microsoft IIS stack, this can be achieved utilizing Application Request Routing 2.0 (ARR), the Web Farm Framework 2.0 and the Web Deploy 1.1 tool (formerly known as the Web Deployment Tool).
http://www.iis.net/download/applicationrequestrouting
http://www.iis.net/download/webfarmframework
http://www.iis.net/download/webdeploy
With SQL Server, the Web Deploy tool and MSMQ on the primary server, the WebServerWatcher class may be used to monitor changes and replicate them across a web farm.
http://blogs.iis.net/moshaikh/archive/2009/04/26/web-server-change-notification-or-web-server-compare-or-multi-server-sync-in-a-web-farm.aspx
(HAProxy and/or IIS and/or F5 or Citrix hardware)
SaaSGrid provides tenant and application load distribution utilizing HAProxy or IIS7, ARR and URL Rewrite modules. The machine hosting this functionality must be clustered for high availability. Additionally, whatever load balancing solution is implemented must itself also be clustered in order to satisfy high availability requirements. It is possible for a given node or cluster to function in both load distribution and load balancing roles at the same time.