AWS Serverless Solutions

Health Cloud Center

Case Study: Administrative processes automation system.

About Health Cloud Center

HCC is a leading company in the Labor Medicine Service with more than 15 years of experience in the professional field. Their goal is to become the ideal tool to resolve licensing situations that arise in the dynamics of Human Resources.

Our client has placed special emphasis on developing a management platform that facilitates communication between the parties involved with the consequent saving of time and effort.

The application handles processes in forms of procedures and tasks that are generated to the different actors. The procedures and tasks are displayed in the generic worktray depending on the permissions and roles.

There are 2 types of procedures at this time: requests for vacation licenses and medical licenses and subsequently after the procedure are started the corresponding tasks are generated for the corresponding actors to complete them.

Features:

• Different types of procedures can be managed: Orders for Medical licenses, Holidays, Paternity / Maternity, as well as order of documents to employees, etc.
• The login is configured for each client depending on their requeriments, at this moment Cognito is being used with username and password
• We have a generic worktray that shows the content depending on the role of the users where their procedures and tasks are displayed.
• The Roles are defined by the client.
• The procedures go through an evaluation / audit process for approval / rejection.
• The information can be analyzed in a Dashboard and the users with the corresponding roles can access it

The Challenge

Implement a solution that allows to reduce the time of request and response in the generation of procedures that is low cost and flexible to adapt it to different types of procedures and end users, elastic so that it adapts to the demands and dimensions of the company.

Initially the project focuses specifically on reducing the management time and the digitization of the documents required for the employees sick leave and vacation leave requests, also improve the response time for the approval or rejection of the requests.

Problem Statement/Definition:

Reuse our workflow engine to add extra functionality and thus generate new solutions for different types of clients. It was required to have access to specific employee data via mobile and web applications and create a flexible and simple online paperwork management product for companies.

The customer's employee information was accessed through the web service provided by the customer. Only the minimum information needed to use the application (ID Number) was stored and the web service was called whenever additional information was needed. The rest of the information is stored in DynamoDB and Elastic Search and was always encrypted in transit and at rest.

We integrate the triggers that were used to execute the lambdas:

Api Gateway: They triggered lambda using Proxy Integration or Custom Authorizer. This integration was used to resolve access to data via all external customers (web/mobile).

SQS and Cloudwatch events: When the product was developed, Step Functions did not support any method of integration other than lambda functions (it was not enough for the case we had), so a CloudWatch events was integrated to process certain pending tasks in Step Functions executions.

Dynamo triggers and Elastic Search Service(Lambda Triggers): They were used to synchronize the data stored by transactional processes in DynamoDB with Elastic search. This allowed us to have statistical data in real time.

Cognito: was integrated to validate OAuth tokens from a lambda used in a custom Authorizer.

Proposed Solution & Architecture:

A module for the management of employee files was developed and 2 additional modules were developed on our workflow engine:

1. Analytics Module that is responsible for obtaining Elastic Search reports, need for advanced reports, were obtained via lambda that spoke to Elastic Search. The lambda used information added by the customauth to validate the access permissions to the information that the user had. The info in Elastic Search was sent via Dynamo trigger.
2. Notifications Module that manages user subscriptions and interacts with SNS to send push notifications and other types. Solved the problem of the need to communicate with users by different means and devices (Android, IOS, SMS). It was used to establish the relationship between users and phone numbers or device ids and to separate from the application the logic of adapting to changes (e.g., phone change). This module was also integrated with API Gateway and DynamoDB.

Operational Excellence: Cloudwatch was used to centralize logs and metrics and X-Ray to perform tracing. And the team was instructed in the use of X-Ray and Cloudwatch Insights for tracking and troubleshooting.

Security: We used the Droptek security and preferences application that makes use of Api Gateway and Authorizer functionality to allow invocations using API Keys and OAUTH Token. For Lambdas were defined roles and policies through the Cloud Formation scripts that installed them.

Reliability: The entire installation of the stack was automated with Cloud Formation to have the possibility of reconstructing the environment if necessary. The deployment is automated using Continuous Integration, with the possibility of performing rollbacks if necessary.

Native AWS backup services are used to safeguard data from all databases. All services are stateless and the configuration is applied from the continuous integration environment.

Performance Efficiency: VPCs and subnets were used with enough IP to support the assumed load peaks. At the data level, was used autoscale (Dynamo feature) from dynamodb. A sizing of memory and execution time of the Lambdas was performed according to the estimates delivered by the client.

Cost Optimization: Lambdas were designed thinking of the minimum possible use according to the requirements estimated by the customer. Cost task was applied in all the lambdas and the client was given an estimated monthly cost according to the load estimates delivered.

Outcomes and Achievements.

A simple tool was developed at low cost that can be adapted to different needs. Promptly our client received a positive impact with a low cost investment, little development time and low resources, measured in the following benefits:

• Integration of the management of procedures in a single platform. Reduction of paperwork
• Increase in the efficiency of Human Resources personnel to manage employee requests because they now have an established process to which they must adhere and follow.
• Reduction in search times by 70%: all the information is consolidated in a single place. Previously, information was stored in folders, making a simple search process a complex task that could lead to unwanted or inconsistent results.
• Reduction in the time of the staff in their tasks of approval or rejection of requests, because each area communicated by mail which always causes delays in responses.
• Availability of automatic information for immediate decision making.
• Control of resources and reduction of absences. Having a platform where employees are required to present a medical certificate as an obligatory form, reduces absences in cases of unjustified medical absence.
• The required documentation is digitized, so the request for the procedure is linear and transparent.
• Ease of compliance with documentation delivery and approval requirements.
• The process was manual and not digitized and involved employees having to take the paper to a physical location. So the reduction of paper was 100%.

Having an order in the tasks, a defined process to follow in each step brings the benefits of improving the efficiency of who executes it and reduces completion times.

After this solution was implemented the HCC client made a new implementation using the same workflow engine architecture to solve another different business topic.

FD goes by file exchange via S3 with M-control-managed transfer

Prisma goes for integration via RPA

The integration with On Premise is done through the connection: Direct Connect that the bank has with AWS and connecting to the services exposed by IBM Message Broker.

Step Functions is used to coordinate the sequence of executions and connects with the lambdas via API both to launch new state machines to interact with them or to take pending tasks using CloudWatch Event Rules (when the project was developed there was no possibility to push SQS directly from Step Functions).