Windows Azure for Function as a Service
June 24, 2014
Windows Azure for Function as a Service

Score 7 out of 10
Vetted Review
Verified User
Software Version
Windows Azure Public Cloud
Modules Used
- Worker Role
- Application Resource Routing
- Storage Queue
- Storage Tables
Overall Satisfaction with Windows Azure
Windows Azure is being used as a vehicle for implementing Function as a Service in various capacities. We find that this is an excellent vehicle to implement single purpose functions (both from a cost and simplicity standpoint) that can then be scaled up and out. We are currently using this model to integrate between an on premise PeopleSoft installation with a mobile platform for work order management and related activities. We find that this provides the field users better and accurate up to date information compared to a timed push model.
- Azure table storage has worked well for eventual consistency problems.
- Multi-tenant compute and data architecture was simple and intuitive
- Mechanism & Patterns to access Storage Queues have worked well for our multi-system (owned and operated by independent teams).
- Java support for worker role was satisfactory & helped us develop the overall architecture without needing to make adjustments.
- All the Java Application Servers that we analyzed for the solution were supported allowing us to test out models.
- Data management tools though third party were satisfactory
- Deployment was complicated with a lot of data (including SDK and Application Server) to be moved.
- Support for global distributed transactions through the platform for message oriented middleware was not found.
- Quick ROI given the low capital expense.
- Allowed us to be truly agile model given the flexibility of using all the features of the solution with low procurement cost and time.
- Helped us implement YAGNI and achieve design simplicity
- We found that this agile YAGNI model helped us keep all the teams up to speed with the solution as we can dissipate targeted message instead of a large complex soluition
We have reviewed an on premise and AWS model for this. The 2 main drivers for choosing Azure were the mobile platform being on Azure allowed for much better integration than a multi-cloud model. We also saw that Azure table storage handled the eventual consistency problem better than Amazon SimpleDB (at the time of our testing).