Vocational Training for UP State Govt (UPSDM)
AT A GLANCE
Uttar Pradesh, India’s most populous state, launched UPSDM (Uttar Pradesh Skills Development Mission). – an initiative to make its large, under-educated, largely unskilled youth, employable. By training the youth at zero fees in semi-skilled areas and vocational training.
Vocational Training for UP State Govt (UPSDM)
AT A GLANCE
Uttar Pradesh, India’s most populous state, launched UPSDM (Uttar Pradesh Skills Development Mission). – an initiative to make its large, under-educated, largely unskilled youth, employable. By training the youth at zero fees in semi-skilled areas and vocational training.
Challenges, not problems
Crazy timelines. An RFP which had half a page of details about the specs of the entire portal. Integration with some twenty-two different government assistance and aid schemes at Centre and State levels. Frenzied levels of access from cities, small towns, large villages, all with flaky Internet connections. And of course, timelines tighter than a debutante’s corset.
Solutions,not ideas
We responded to the extremely tight timelines and unprecedented scale-up challenge by setting up the system on AWS, with a hand-tuned Postgres on an EC2 instance and adding provisioned IOPS for throughput. We re-wrote the front-end code to deliver most master tables as static JS files as part of the code release, thus cutting down AJAX calls to the server to look up master data. This dramatically reduced the impact of unreliable Internet connections on application reliability. And because of our use of traditional EC2 foundation, we could move the application to the client’s on-premise data centre to comply with their govt regulations as soon as they set up their DC.
Challenges, not problems
Crazy timelines. An RFP which had half a page of details about the specs of the entire portal. Integration with some twenty-two different government assistance and aid schemes at Centre and State levels. Frenzied levels of access from cities, small towns, large villages, all with flaky Internet connections. And of course, timelines tighter than a debutante’s corset.
Solutions,not ideas
We responded to the extremely tight timelines and unprecedented scale-up challenge by setting up the system on AWS, with a hand-tuned Postgres on an EC2 instance and adding provisioned IOPS for throughput. We re-wrote the front-end code to deliver most master tables as static JS files as part of the code release, thus cutting down AJAX calls to the server to look up master data. This dramatically reduced the impact of unreliable Internet connections on application reliability. And because of our use of traditional EC2 foundation, we could move the application to the client’s on-premise data centre to comply with their govt regulations as soon as they set up their DC.