Challenges - kristiperreault/ece601-placement-exam GitHub Wiki

What challenges does this SaaS present? What will be difficult to work through?

• Storage: Most language translation occurs text-to-text, meaning that audio files will have to be saved somewhere, converted to text files that will also have to be stored, and translated to another text file. With multiple calls occurring, these files can quickly grow to be large and numerous, and we need a place to store and manage (retention, location, etc) all of them.
• Price: All cloud services will cost money, some more than others. Even in a "pay for what you use" model, we have a large number of files to store, and many processes utilizing several different services, which can quickly add up. How do we optimize cost for us and the user?
• Real-time translation: The process of audio to text, text translate, text to speech is complex, and may be fairly quick, but it may create too much latency in real-time video. How do we ensure this process works as smoothly and quickly as possible, without causing long waits and frustration for the user?
• Quality video: Basic video conferencing and streaming is prone to latency, video quality, and buffering as it is - adding on another layer of processing with the language translation may slow or pixelate the video feed even more. How do we provide all of this features without compromising quality?