Shubhra Prakash Nandi

I developed a CardDAV server using LDAP for authentication and contact back end. The server offers the following features:

• Authenticate CardDAV users against LDAP
• Multiple address book support
• Global, shared, and private address book support
• Full read/write support between CardDAV clients and LDAP with contact renaming
• Bidirectional sync between LDAP and CardDAV clients for new contacts, modified contacts, and contacts deleted via the CardDAV protocol (contacts deleted directly in LDAP are deleted in CardDAV clients asynchronously)
• Extensive and customizable configuration options to map vCard properties to LDAP attributes, including multi-value and composite value properties and attributes
• Media-like image storage support
• Contact group support

I created a MySQL 5-node (five instances of the MySQL server) InnoDB HA cluster with a multi-master mode of operation (write any instance and read from any). The cluster was configured with multiple features.

CLUSTER FEATURES:
• Automated failover to a healthy instance when an instance goes down.
• Automated recovery when an instance comes back up.
• Strict data consistency across cluster instances.
• Automated load balancing of reading and writing queries over all cluster instances, achieving better throughput than a standalone instance.

The cluster also had no single point of failure. Hot incremental backup was done using a replication instance connected to the cluster using snapshots. The incremental backups were used for the point-in-time recovery of data. There was an offsite replication instance doing chained replication from the onsite replication instance for disaster recovery.

I created a solution architecture using RESTful services and microservices for a SaaS product supporting millions of users and thousands of vendors. REST services were written using the PHP Phalcon C extension framework with all microservices being flexible enough to be clustered for high availability and redundancy. The primary objective of this project was to able to scale (or auto scale) out without hampering the availability of the services.

I developed the integration gateway platform for one of the world's biggest automobile manufacturers for their vehicle telematics, onboarding problem diagnostics, remote assistance, crash reporting, turn-by-turn navigation, and infotainment system.

This integration platform uses the latest service-oriented architecture (SOA) suite from Oracle Fusion Middleware. It connects all customers' systems to provide a uniform interface for all end users to consume IT services.

The SOA suite is hosted on a Linux platform. It uses the Agile development methodology and continuous integration to achieve high productivity while keeping new functionality and features flowing for the customer, giving it an edge over its competitors in this fast-growing and competitive IT space in North America and China.

This ADS-layering project was about building an Oracle OLAP database to provide data to external vendors.

The client's operational database stores all customer and vendor transactions for its multi-channel, integrated marketing platform. Currently, the client has batch feeds to provide this transactional data to external vendors, which is time-consuming, complex, and resource-intensive.

To overcome these challenges, I needed to build the ADS-database layer (also known as the provisioning layer by the customer), which can expose the same data to the vendors through database views. Vendors will be able to connect to this database in a restricted manner and query their data.

The project aimed to integrate a manufacturing execution system (MES) suite called PSI Metals with the client's legacy systems. This project was part of the client's initiative to upgrade their system to an industry-wide standard MES product and SAP compliance.

The integration of PSI Metals into the existing system would decommission many existing applications and background processes. It would require new changes for the existing applications to work with PSI Metals, especially with their message-based interaction system. Thus this project would carry out the necessary remediation activities to the existing applications and integrate PSI Metals with the legacy systems.

The project had a high degree of complexity and risk as almost all critical business processes were highly or substantially impacted.