The client is a fast-growing startup that has expertise in helping its customers to provide exhilarating creative expressions to ideas, thoughts & memorable moments, turning boring PPTs to animated videos.
Customer requests were pushed via email and Google Docs and not centralized causing chaos & customer dissatisfaction. Frequent status updates are crucial for business success. Since client’s USP is based on response time, they wanted the response time to be minimized as much as possible with end-end workflow visibility including showing the wait time to process the video based on the current queue.
Based on the current technology constraints client had in terms of moving the entire video bot engine to cloud, OSoft Labs in turn proposed a hybrid cloud model. Created a phased approach for moving the entire video bot solution to cloud in a span of 3-6 months. Since there will be bandwidth constraints in a hybrid model based on download and upload speed, size of the request and response plays a key role. This was very carefully reviewed and taken care by minimizing round trips to on premise server. High quality images are stored in S3 and partial output streams are displayed to user from cache. The application maintained SQS queue along with priority to pick and process the requests. From earlier requests of one per day, the client is now able to handle more than 60 requests per day with just one video bot server. The entire solution is flexible to scale on premise video bot servers to handle more requests.
- Improved response time in generating the final output
- Scale environment to match high demand
- Queueing based on client’s profile
- Increased number of requests handled from one per day to 60 requests per day with just one video bot server
- Custom Security -Proprietary business logic and source code is running on premise
- Transparent and better user experience on overall status
JAVA, JQuery, HTML5/CSS, SQL Server, JSON, AWS Services – S3, EC2, Load Balancer, Route 53, Cloud Front, Cloud Formation, VPC, RDS, SES, SNS, SQS, CloudWatch, IAM