Robert Krien Email & Phone Number

Tel Aviv District, Israel

My group owned multiple projects in the Uplynk video streaming solution over the years:Cloud Slicer LiveCloud-based solution for live video ingest.Design challenges include:AWS resource sizing and reservationResilience and inter-regional failoverSecurity in a multi-tenant cloud environment: Secure and stable ingest routing for RTMPS, SRT and RISTOur solution usesAWS EKSCustom Kubernetes ERDs and OperatorsNginx and Mist Server as ingest proxiesDynamic modifications of AWS Networking Security and Firewall rulesAPI GW and Lambda for exposing a REST API using Spring integration CI/CD using Gitlab Terraform and HelmSplunk, Observe, Prometheus, Grafana …VibeAdvanced video ingest, routing and processing mesh to leverage Verizon’s backbone network for reliable video-routing and processing. Implemented using globally dispersed interconnected K8s clustersVideo Policy APICDN selection for millions of concurrent streaming sessions based on static and real-time data. We designed the solution in two layers: A Backend that processes live stats and static configurations into a set of easy-to-infer Decision Instructions.A Frontend that uses these Decision Instructions to serve large numbers of individual selection requests efficiently.Toolset: Java, Jetty, Python, various AWS technologies, CI/CD based on Screwdriver (a Yahoo-internal variant of Jenkins), CloudFormation.Research on proprietary UDP-based protocols for Video StreamingVarious research and reporting projects on CDN usage and reliability using Scala on Apache Spark, R.Tools used: Java, Python, Scala, C++, Kubernetes, AWS(DynamoDB, S3, network routing and security...), Apache Spark, Nginx, Mist-Server, Android, Chromium, Quick,SRT, RIST, RTMPS, Lightning (a proprietary UDP-based Streaming Protocol)Things learned: CI/CD for owned and operated global systems in a public cloud, Video protocols, Video streaming, CDNs.

Herzelya And Haifa

ScaleIO advanced management system for large-scale installations:Designed and implemented the system with the teamCentral management server deployed as a HA cluster using Java NIO and Quasar Fibers to meet massive scale and QOS requirementsScale-out cluster of worker nodes for monitoring and aggregating real-time data over large scale installations (10 servers, hundreds of thousands of storage devices, millions of volumes)ScaleIO appliance:Defined the data model for modelling the hardware components of the various vendors EMC partnered with in a unified mannerCo-designed the on-boarding process for clients to plan and execute and monitor the set-up of large installationsImplemented the Netty infrastructure for the management server.ScaleIO management infrastructure:Lead the design of the management solutionCo-designed the protobuf network protocol with the driver group.Implemented networking and model infrastructure.Lead the design and implementation of ssl/tls adoption including certificate/ca handling, threat modelling and exception handling.Implemented alert detection infrastructure.Tools used: Java, Spring JNI, Quasar Fibers, NIO, Netty...Things learned: Storage paradigms like volumes and snapshots etc, work in highly distributed systems with massive scale requirements and more

Herzelya

Architected the central management infrastructure and implemented core elements of it.Designed and implemented integration with leading clustering technologies like MSCS, Steeleye and various Linux native clustering solutions.Defined REST API for monitoring and managing large installations.Designed and helped to implement VSphere Plug-in for seamless integration with VMWare environments.Helped implement C SPI for CIM integration using EMC's ECOM Broker.Implemented standalone Java management clientTools used: Java, Hibernate, JNI, C, C# and various scripting languages. Low level integration with Windows/COM, various Linux flavors, VMware ESX OS.Things learned: Architecture of modern storage controllers and servers, standard practices for setting up large scale data-centers, OS device management

Raanana, Center District, Israel

Filled multiple positions in the SAP Portal core group:Design & development of key design-time applications and frameworks for content developers.Transformation of the Portal Runtime infrastructure from a standalone servlet running on Tomcat to a J2EE extension in SAP's own J2EE server. Migration of SAP's proprietary portal application deployment format to standard J2EE files and descriptors.Adaption of the custom class-loading concepts of our original Portal implementation to a standard methodology.Dependency management for applications and content.Maintenance of our SOAP implementation that features automatic creation of stubs and proxies for existing services. Re-design and implementation of key-parts of SAP's complex and feature-rich data repository ( Portal Content Directory) to solve memory and performance issues with the previous implementation that resulted in a 40% reduction in the memory footprint of standard installation.Frequent service as emissary for the Israel lab in different locations including a 4-month stint as the permanent integration representative in SAP's headquarters in Germany.

Jerusalem, Israel

We developed a mapping application aimed at mobile devices using J2ME.Due to the very limited resources available on the mobile devices of the time this required a very resource-aware approach.My role included the implementation of the mobile client in various flavors of Java as well as the design, development and execution of the batch processes that transform commercially available mapping data to a format that can be easily consumed by a client with limited resources.