As the demo evolved, so did the project. The client’s needs were slightly different from what Hindsight Metrics does. So, a subproject arose: Poller. The name comes from the main activity the project does, polling for information from a given site. The type of information to be polled should be configurable by the client on the fly. This means that the client should have a portal to add, edit and remove information.
All this opened up new avenues of delivery: On-premises or Cloud! Besides the feeling of saying “I have an application on the cloud”, a client will have a better experience due to the scalability of the system and faster response time from my part. If the application were installed on-prem, the client would have to provide a server and network configurations would be needed, later if more resources are needed or a bug shows up, I would have to go in personally and apply the patch.
Thinking Poller can be sold to different clients, having more than two it becomes a hassle for myself. So I started to research the three biggest cloud providers: Amazon AWS, Windows Azure and Google Cloud Platform.
Overall, they all offer the same kinds of services: Compute engines, Storage solutions, Virtual Machines, and many many more. Because Poller services need to run on a windows machine, I chose Azure because it offers the lowest-cost options for virtual machines.
The migration to the VM was really smooth, aside from how the application runs, the object could just be copy/pasted on to the server (I have since installed Octopus Deploy, more on this later) and needed to open the necessary ports so it’s accessible from anywhere on the internet. Also, since I already had a google domain I set it up to redirect to the Azure VM, so it can be accessible with a friendly name.
Everything is set up for the demo with the client, which was postponed to the end of March, beginning of April.
Leave a Reply