Dear ThingWorx community team, I am a new bee to this community group, first time working on IoT & also using platform and I have specific use case to work on using ThingWorx Platform and it is more related to graph database.
My use case details are as follows: We are working on Industry services connected with Sensors and actuators for around 30,000+ devices, has around 40 data points in each and it is mapped to customer assets/ports. Approximately need to get accommodate around 3 million nodes (vertex) and approximately 6 million edges, that is, each node will have a relationship (edge) to 2 other nodes and Azure Cosmos DB will be the backend database.
We need help to know whether ThingWorx platform can able to scale the load (since as per reading normally it scales vertically), please let us know how to consume the data into the ThingWorx platform, perform Assets hierarchy, functional calculations if any during the process (ETL operations in the platform) and finally store the data in graphical structure into Cosmos database. Few considerations are, can we able to perform all these operations through APIs to use in our applications and also, while performing query operations, it may go up to more than 100 levels down (decedents) to retrieve the results during any read operations and performance is the key factor to consider (acceptable result display < 2 seconds). Please let us know if it can be achievable, how to go with ThingWorx and believe it supports with this huge scale. Please suggest and share any useful information / documents for the same.
Thanks for your response Om Dukiya,
It really helped us to get valuable information around scalability concerns that we were looking in using ThingWorx platform.
Since I am new to ThingWorx platform and looking for implementing with the streaming data that we receive from IoT devices and transforming into Graph database format (that is, organizing the assets like customers, devices, controllers and so on with proper relationships in graph format to store in database like Cosmos DB to use for traversing the tree easily to fetch the required information within few milli-seconds that we are looking into it using this platform).
Need help to know whether the ThingWorx platform can be used purely as an Asset Tree or do we need to bring IoT data to make it function to get the desired output result in Graph database format, so that it can able to withstand to get validate with all possible tests (to hold more than 3 to 5 million relationships with multiple thousands around 30 to 40k assets with minimum 2 relationships from each vertex/node in this case).
Please help us to provide the clarification for the same.
Given the scope of your question, we would recommend contacting your Sales Rep for discussing with a member of the pre-sales team. Another possibility would be to arrange for consulting to assist with the planning and design of your solution. One question we had while reading this was why it would be necessary to store the information in ThingWorx if it's already stored externally. There may be many things to consider for designing a solution of this scale that performs well.
If you feel your questions have been answered, please mark the appropriate response as the Accepted Solution for the benefit of others with similar questions.