Risk Mitigations
Following are the Risks involved and a short summary of how we mitigate those risks with our solution.
Last updated
Was this helpful?
Following are the Risks involved and a short summary of how we mitigate those risks with our solution.
Last updated
Was this helpful?
We have designed our solution with security in mind. We have used W3C technical standards and protocols that are highly secure and provide end-to-end encryption. We have taken extra care to make sure that all communication is secure and any user’s data is stored in a secure format in our databases. We also have a cyber security advisor Saurav Sengupta () who is helping us make sure that our infrastructure is secure. Apart from network security, we have added several application level security features like Password protected rooms, Role based access, Host controls like mute participant, pause video of a participant, remove a participant from a conference, put participants in a waiting room, and many others. Moreover, Unlike other video conferencing solutions, We default to the highest security setting for any meeting/webinar. Hosts can turn off some of the application level security features if they want to. We offer on-premise hosting solution which allows you to host all the backend infra and the database in Govt’s owned data center. This allows full-control over all the data and prevention from any unwanted access.
We have designed our backend infrastructure to scale to millions of concurrent connections. Our infrastructure is capable of supporting more than 500 participants per meeting room which can be scaled even further by adding more SFU servers. We use a concept called SFU cascades wherein a conference can span across multiple SFU servers. This is why we’re not limited by the CPU and bandwidth of a single SFU server. All our backend servers are horizontally scalable. To handle more load, we can add more servers and the infrastructure will be able to support additional load.
We also support multi-datacenter deployments by default. The users are connected to the datacenter closest to them to improve latency.
India, being a developing nation, doesn’t have high speed internet at all places. We believe that technology should empower everyone whether they are living in urban cities or remote villages with slow network connection. Our solution detects network conditions and automatically upscales or downscales video/resolutions and bitrates to make sure that we provide optimal experience to the users. Our media server (SFU) that handles video/audio communication supports Error correction, handles packet loss, and makes sure that it forwards the video/audio streams to the receivers depending on their bandwidth and display.
Some of the technical hurdles we would face will be in the field of AI/ML.
We are building closed captioning as part of our solution. Building this requires large training datasets in multiple Indian languages which is difficult to get since the area of AI based speech-to-text is very new. We plan to collaborate with Indian academic institutes for our data needs.
In order to develop our AI solutions, we would require powerful compute resources in the form of GPUs to perform parallelised large scale mathematical operations on our datasets.
Once we have developed the closed captioning solution, we need to make sure that it can perform speech-to-text tasks with reasonable latency. In this regard, we are evaluating multiple state of the art frameworks and open source solutions to make sure that all the performance requirements are met for a smooth user experience.
Apart from this, there are challenges involved in the cyber security field to make sure that we leave no stone unturned when it comes to security. We have taken extra measures to ensure that all the communication is end-to-end encrypted and have employed several application level security measures. We are also in touch with experts in the cyber security domain who are willing to guide us.