WP1 - Placement: This work package studies methods to determine optimal placement of anycast nodes.
  • T1.1 - Literature study of (DNS) anycast operational practices, performance parameters, and related work on self-managing (DNS) anycast networks.
  • T1.2 - Define framework for anycast node evaluation. That is how are performance metrics used in node placement evaluation (which weight factor and their relation to each other, etc.).
  • T1.3 - Simulation study with real-world Internet topology (e.g., from CAIDA or UCLA) and DNS TLD query traces. With the evaluation framework from T1.2, a set of strategic optimal locations should be derived.
WP2 - Monitoring and Execution: This work package is a first step towards self-managing anycast infrastructure, and includes the basic monitoring plumbing at the anycast nodes and at a global coordinator for overview. The execution part allows series of steps to be executed at the anycast node to start/stop, reconfigure, etc., a service. For some of these parts, modeling of knowledge is already required.
  • T2.1 - Study current practices in monitoring anycast services, e.g., RIPE DNSMON, RIPE RIS, or Oregon Route Views. Devise breakdown of local and global performance parameters for monitoring. Think of framework bringing everything together.
  • T2.2 - Initial design and implementation of anycast infrastructure monitoring subsystem, including local components at anycast node as well as global monitoring infrastructure combining local information.
  • T2.3 - Similar to T2.2, an accompanying execution subsystem is designed and implemented to remotely manage parts of the anycast infrastructure, including the (dynamic) discovery of virtualization capabilities at different sites.
  • T2.4 - Evaluation of monitoring subsystem, e.g. scalability and timeliness of results, and execution subsystem combined.
WP3 - Analysis and Planning: This work package adds the analysis, planning, and knowledge components to the management system. And with a feedback loop, this all makes the system a fully self-managing infrastructure.
  • T3.1 - Literature study autonomic computing and communication, knowledge modeling, root cause analysis, and planning.
  • T3.2 - Knowledge model for DNS anycast self-management infrastructure (both local and global knowledge db).
  • T3.3 - Design and implementation of analysis module for local and global root cause.
  • T3.4 - Design and implementation of planning module for local and global planning.
  • T3.5 - Evaluation of all components of the MAPE-K feedback loop on correctness and effectiveness.
  • T3.6 - Final report.