This project aims to architect and implement a demo scenario of a System Design requirements.
Imagine you are one of our employees and you got assigned to develop a feature.
- The feature has a number of connected appliances that belongs to a number of customers.
- We have a need to be able to view the status of the connection among these appliances on a monitoring display
- The appliances send the status of the connection one time per minute
- The status can be compared with a ping (network trace). For example; no request from the appliance means no connection
- There is NO need to ping anything, it is just to understand the concept whether the Appliance is either Connected or Disconnected
- Your task will be to create a data store that keeps these appliances with their status and the respected customer
- An API that returns list of Appliaces with connected status and customer details
- Obviously, for this task, there are no real appliances available that can respond to your "ping" request. This can either be solved by using static values or by creating a separate machinery that returns random fake status
- Appliance could be separate service (Micro service)
- Can support high transactions (High number of appliances)
| Appliance Id | Factory Nr | Customer | Address | ||
|---|---|---|---|---|---|
| YS2R4X20005399401 | ABC123 | Kalles Grustransporter AB | Cementvägen 8, 111 11 Södertälje | ||
| VLUR4X20009093588 | DEF456 | ||||
| VLUR4X20009048066 | GHI789 | ||||
| YS2R4X20005388011 | JKL012 | Johans Bulk AB | Bulkvägen 12, 222 22 Stockholm | ||
| YS2R4X20005387949 | MN0345 | ||||
| YS2R4X20009048066 | PQR678 | Haralds Värdetransporter AB | Budgetvägen 1, 333 33 Uppsala | ||
| VLUR4X20005387055 | STU901 |
ApplianceStatusStore : Application that stores Appliance Connectivity status in the database.
Appliance : Application that represents an Appliance which sends ping requests to the ApplianceStatusStore ping REST EndPoint
To Support high transactions from high number of Appliances, a load balancer can be deployed to distribute the traffic between multiple ApplianceStatusStore applications.
Deploying a Key-value store like Etcd or Redis can improve db operation. Especially Redis could perform better since it is an In memory Database. All the ping requests from the Appliance can be stored in the Key-Value store and then the appliance status can be updated with a background job from the ApplianceStatusStore
An alternate solution could be to design the system using Publish Subscribe pattern where Appliance act as a publisher and publishes ping request to a specific topic in the Broker and then the Subscriber applications can read the request and update Appliance status. In this case ApplianceStatusStore will only need one REST Endpoint just to server the Appliance details to whoever request for them.