The Proof of Concept term has been all over the Web recently. It is often mentioned in connection with the Internet of Things – the global connected environment where gadgets collect data and exchange it over a network. What is PoC and does it exist outside the IoT realm? Read on to find out! The whats and whys of Proof of Concept in software development Proof of Concept (or PoC for short) is the evidence that a certain system or device can be implemented in the real world and become a commercial success. PoC is similar to prototypes (chunks of software that perform functions of the future system or gadgets built with off-the-shelf solutions like Raspberry PI and Adruino), although no deliverables are represented at the stage. The document often provides the description of business processes, their objectives and participants and demonstrates that the proposed solution meets certain business requirements. In other words, creating PoC is a great way to find a solution to a technical problem, make sure it really works and evaluate market demand for a product. With Proof of Concept, you can:
You want to retrieve data from enterprise apps written in different programming languages. Enterprise application integration (EAI) is often challenging and involves major recoding; does it mean you have to create Proof of Concept? It does – as long as you implement new APIs or libraries which have never been used for this purpose. In case similar EAI solutions exist on the market, Proof of Concept is replaced with thorough research. You address a custom software development company that analyzes the tech stack of your enterprise applications and finds the right middleware. Thus, there’s no coding done (not before you initiate the EAI process). Provided you have an in-house IT department, R&D facilities and relevant expertise, you can even conduct research on your own and simply pass the requirements to your vendor. It’s different with IoT. A couple of months ago we published a comprehensive article on IoT product development costs, citing novelty of an IoT solution as the key reason to prepare Proof of Concept. According to Pavel Shylenok, CTO at R-Style Lab, it makes sense to start with a POC when a market-ready product costs are considerably higher than the PoC itself. Some new ideas – like cutting-edge hardware sensors or trendy Home Automation systems –might just not be feasible due to technology limitations (it’s not there yet). Despite of the “generally it works” state, high latencies, low precision, high energy consumption (whichever applies) won’t allow POC to reach market-ready level. When it comes to IoT solutions, PoC results most often answer the following questions:
Home Automation systems often employ different sensors connected with some AI logic. While each sensor can work fine when used independently, combining them can be tricky due to mutual interference or high power consumption or low/unreliable performance. Before your vendor creates PoC, the questions we’ve mentioned above cannot be answered. Same with IoT healthcare solutions. Once we were approached by a US company that manufactures intervertebral prosthetic gadgets. Each of these devices has a unique identifier, so they wanted to build a custom machine vision system to automatically decode X-ray images. Although the R-Style Lab team implemented various APIs and technologies to reach the goal, the accuracy of the image recognition software staggered around 70%. In order to achieve better accuracy, we needed 4K resolution images which are not produced by modern X-ray machines, so the project was postponed for an indefinite period of time.
4 steps to PoC success
0 Comments
Your comment will be posted after it is approved.
Leave a Reply. |
Archives
February 2018
Categories
|