İsmail Taha AYKAÇ

Internet service providers and mobile operators supply various vertical services over an underlying network infrastructure. Any outage and quality problems directly or indirectly affect their services and customers dramatically. Therefore correctly managing this infrastructure is hugely significant. However, the infrastructure for a big telco company is complex as it involves multiple layers of technologies with various vendors, many pieces of equipment, diverse use cases, etc.

Our platform provides multi-technology and multi-vendor observability by providing the following:

• Auto-discovery of the physical and logical equipment, circuits, and services.
• Correlation of different vendors and different technologies.
• Live event collection and correlation with services, teams, and customers.
• Computation of SLA.
• Live and simulated outage impact analyses.
• Monitoring utilities, reports, and over 100 integrations with other systems.

A large internet service provider that has 35 million customers has used it for more than ten years. I have been the project manager of this constantly evolving platform for six years.

This pharmaceutical track-and-trace system has been developed, managed, and maintained by TechNarts. The system tracks and traces all the drugs and controls every movement in their supply chain. Therefore, the sale of counterfeit medications, reimbursement fraud, drug theft, and barcode scams are made more difficult.

The system annually tracks around three billion products with an average of 15 notifications for each, 45 billion requests in total. It has a very low response rate, which was about ten milliseconds. I oversaw the core system team.

This app is developed for a large communications service provider (CSP) with over 10,000 corporate customers. Both customers and the CSP need to monitor and log the traffic of the MPLS services. The app collects data for 350,000+ services every five minutes and stores it inside a big data store. The collection is done with SNMP polling, resulting in around 2.5 million SNMP GET requests done in five minutes.

The app is divided into five layers:
• Front end: Single-page application developed with TypeScript and React.
• Back end: Application developed with Python 3, Django, and the Django REST framework.
• Data Layer: PostgreSQL for application data and Cassandra for measurement data.
• Measurement Microservices: Microservices for scheduling, collecting, processing, and other operations.
• Message Broker: Apache Kafka for messages between layers and microservices.

The microservices are designed to be stateless and communicate with the message broker. This leads all components to be managed and maintained separately.

As the project architect, I designed the architecture and contributed to the different stages and layers. I primarily handled CI/CD pipelines and app containerization.