AWS-Based Financial Service Grid Computing System Architecture
The Financial Service Grid Computing Based on the cloud environment provides Dynamic Scaling and elasticity for computing jobs on demand. The aggregation service simplifies the development of grid software.
AWS provides a powerful platform for high-performance grid computing systems based on On-Demand hardware configuration and template-driven deployment, combined with low-latency access to existing internally deployed data sources.
1. market information, transaction data, and competitor data are installed in an enterprise's internal data center or AmazonSimple Storage Service (Amazon S3 ).
2. AWS DirectConnect establishes a low-latency and reliable connection between enterprise data and AWS from 1 to 10 Gb. For low bandwidth requirements, a can use VPC Gateway to establish a VPN connection.
3. Create private subnets for user data sources, grid computing clients, grid computing controllers, and engines.
4. Applications and enterprise data are securely stored in Amazon RDS ).
5. Amazon Machine Images (AMIs) contains operating systems and grid software, and grid controllers and engines run on EC2 that are enabled Based on AMI on demand.
6. static data, such as holiday calendar, QA library, and additional gridlib boot data, can be downloaded by the Grid Engine from Amazon S3 at startup.
7. The computing results of the Grid Engine are stored in Amazon DynamoDB. This NoSQL database provides configurable read/write throughput and allows on-demand scaling.
8. Amazon DynamoDB uses Amazon Elastic MapReduce (Amazon EMR) map/reduce jobs for aggregation. The final results are stored in Amazon S3.
9. The Grid Computing client collects aggregation results from Amazon S3.
10. Amazon Glacier is used to archive aggregation results. This is a low-cost, secure, and durable storage service.
This article permanently updates the link address: