Specific problems.
Measured improvements.
Selected projects from Boris Pilka’s previous roles and engagements. Each result belongs to the system and conditions described.
Aviation networks
Less time recovering. More time connected.
An event-driven redesign of network monitoring and traffic steering.
<0.5sbest-case failover
Less time recovering. More time connected.
An event-driven redesign of network monitoring and traffic steering.
The context
Network monitoring and traffic steering for air traffic management needed faster failover and a more modular architecture.
Boris’s role
Software architecture and implementation.
The approach
Proposed and implemented an event-based architecture. The redesign allowed monitoring and steering to be offered independently and introduced an improved graphical monitoring interface.
The result
Worst-case failover fell from 17 seconds to under 5 seconds. Best-case failover fell from 7 seconds to under 0.5 seconds.
C · Java · Python · Prometheus · Grafana · RHEL · Docker
Reported results from Boris Pilka’s prior project work. The best- and worst-case figures describe that implementation’s measured behaviour.
Network monitoring
More throughput. Less overhead.
A probing-engine redesign improved efficiency across three measured dimensions.
−70%required bandwidth
More throughput. Less overhead.
A probing-engine redesign improved efficiency across three measured dimensions.
The context
A network probing solution in an air traffic management product needed to use network and processor capacity more efficiently.
Boris’s role
Performance optimisation and redesign leadership.
The approach
Led a comprehensive redesign of the probing solution to improve its resource efficiency and supported throughput.
The result
Required bandwidth decreased by 70%, CPU load by 30%, and maximum packets-per-second throughput increased by 35%.
C++ · Linux · Performance engineering
Relative improvements reported for the redesigned implementation in Boris Pilka’s prior engineering work.
European ATM research
From network architecture to flight trials.
Multi-homing and mobility research in aviation networks.
SESARresearch & trials
From network architecture to flight trials.
Multi-homing and mobility research in aviation networks.
The context
SESAR research addressing multi-homing and mobility across changing connectivity conditions in aviation networks.
Boris’s role
Software architecture and team leadership.
The approach
Guided development from proof of concept to field validation, using networking and mobility technologies including LISP and NETCONF.
The result
The research progressed from an initial prototype to successful flight trials.
Python · IOS-XE · LISP · NETCONF · SNMP · RHEL · Docker
Prior research and engineering experience within SESAR projects. Programme names provide context for the work.
High-load ticketing
Ready for the moment demand arrives.
Autoscaling cloud infrastructure for major ticketing peaks.
30ktickets in minutes
Ready for the moment demand arrives.
Autoscaling cloud infrastructure for major ticketing peaks.
The context
A Slovak event-ticketing platform faced concentrated demand during major event releases.
Boris’s role
Infrastructure development and operations.
The approach
Developed and operated autoscaling cloud infrastructure, supported by configuration management, deployment tooling and load testing.
The result
The platform supported the sale of 30,000 tickets within minutes during major event releases.
Google Cloud · Python · Ansible · Terraform · JMeter · Jenkins · Nginx · Docker
An observed event peak from Boris Pilka’s prior ticketing-platform work.
Distributed systems
Put workloads where resources make sense.
Energy-aware orchestration across distributed data centres.
Energyaware placement
Put workloads where resources make sense.
Energy-aware orchestration across distributed data centres.
The context
Orchestration across distributed data centres needed to account for energy availability, heat demand and latency.
Boris’s role
Core developer on the first generation of the orchestration software.
The approach
Implemented microservices that selected a site using those operational metrics, then migrated jobs when conditions changed.
The result
A working first-generation orchestration system with metric-based placement and automatic job migration.
Python · Flask · OpenStack · Kubernetes · Docker · RabbitMQ · Prometheus
Architecture and implementation experience from Boris Pilka’s prior project work.
Telecommunications
High-throughput service provisioning.
Protocol implementation, provisioning middleware and custom testing tools.
500+requests per second
High-throughput service provisioning.
Protocol implementation, provisioning middleware and custom testing tools.
The context
Core middleware provisioning services for telecommunications customers.
Boris’s role
Protocol implementation and test-tool development. A later engagement covered a migration to Java.
The approach
Implemented a new generation of a proprietary protocol in an existing Tibco-based platform and built supporting test tooling. In the later engagement, migrated provisioning middleware to a pure Java application designed for high throughput.
The result
The protocol implementation supported provisioning at more than 500 requests per second. The Java migration was a separate engagement.
Java · Python · Tibco · Solaris · Spring Boot
The reported throughput belongs to the protocol implementation, rather than the later Java migration.
Have a similar constraint? Bring the performance target, failure mode or delivery problem.
Discuss your project