Research Problem

Calypso IP project was launched on basis that there was need for own router software that can be directed by Calypso services (these services were meant to be created by Calypso SP project that is the other part of whole Calypso project). Goal of Calypso SP was changed and at this moment Calypso SP doesn't have any needs for router that can be directed by services.

Because of this change of goal it was noticed that there wasn't so much need for service guided router control. This was one reason why Calypso IP will focus more and more to routing related technologies and that how they could be improved. Calypso IP will implement basic routing protocol (OSPF) and all necessary modules in IP router. This includes researching how QoS could be implemented in future routers and how IPv6 is related to this matter and so on. Calypso IP will do basic work for other TKK's projects like improving OVOPS++ coding environment and creating basic steps for IP router protocol that is suitable for needs of other TKK's project as well. Project will work closely with Scoms project in the area of research and implementation of IP routing related to MultiProtocol Label Switching (MPLS) development by Scoms project. IP-routing in hierarchical Mobile IP environment is researched together with GO/Core. Towards the end of project, Calypso-IP will concentrate more and more on IP-routing in Mobile IP environment.

Research Plan

Project plans made on Calypso IP project:

Overview of the project plan for 1999-2000. (in Finnish)

Project plan for 1999-2000. Version 1.2. (in Finnish)

Project plan, timeline 1.9.1999-31.8.2000. (in Finnish)

Project plan for 1998-1999. Version 0.4 (in Finnish)

Project plan for 1998-1999. Version 1.0 (in Finnish)

Project Description

Calypso IP project was launched on the basis that there was need for own router software that can be directed by Calypso services (these services were meant to be created by Calypso SP project that is the other part of the complete Calypso project). Goal of Calypso SP was changed and at this moment Calypso SP doesn't have any needs for router that can be directed by services. Because of this change in the goal it was noticed that there was not so much need for service guided router control. This was one reason why Calypso IP will focus more and more on routing and related technologies and how they could be improved.

Calypso IP project started at the beginning of summer 1999. Since that the project has researched different aspects related to internet network routing technologies. Based on this research it was decided that project will implement routing protocol that can handle the ever growing need for fast and reliable internet network routing. Open Shortest Path First (OSPF) was chosen to be the implemented protocol. Decision was based on the fact that internet networks will grow dramatically in the near future and OSPF is a protocol that offers multiple features and has capacity to handle this situation. One more fact was that OSPF protocol is "open" and extremely popular. First step in the implementation process is to provide basic IPv4 implementation of the protocol and after that implement IPv6 of OSPF.

OSPF implementation will be tested and used on the VTT supplied hardware. This hardware is equipped with several Ethernet interfaces and it performs IP forwarding between these interfaces. Calypso-IP project can therefore test and study different IP-routing scenarious on a real hardware IP-router. We are especially interested with routing in Mobile IP environment because this hardware is used also in GO/Core project. The requirements of IP-router in Mobile IP environment will be studied.

It might be possible that some configuration tools has to be written to configure the VTT hardware but this depends hardware vendor.

Ovops++ framework and C++ are used in development work of OSPF. This is quite new aspect in OSPF (and in any routing protocol) implementation. Earlier OSPF implementations have been done almost every time in C -language and utilizing frameworks that are built only for routing protocols. Our approach is to use generic object oriented framework and object oriented technologies and collect valuable experience in this field. This kind of highly object oriented implementation offers some great benefits: (1) code reuse, (2) simple code, (3) using existing framework (Ovops++), (4) implement the protocol exactly like it is described in standards etc. There are also some negative points in object oriented implementation: (1) the lack of performance compared to pure C implementation, (2) additional system libraries, and (3) overhead from the framework. Even implementation of OSPF with C++ will require more CPU than ordinary C implementation; use of Ovops++ and C++ is one of the research results. It gives the idea of how object oriented languages and frameworks suite in this kind of protocol development.

Besides the implementation of OSPF, the project will research QoS aspects in routing protocol. This kind of intelligence in routing protocol leads to better utilization of network and smarter routing. The network bandwith usage in Mobile IP environment, where several mobile users consume network resources, combined with QoS characteristics is definitely important research area.

Project will research how ISO/IEC 9646 testing methodology could be applied to OSPF protocol testing. Results will be testing model, some test cases and Gateway implementation between OSPF protocol under test and OpenTTCN tester software by Open Environment Software Ltd. And hopefully OSPF that is more bug free than before the testing.

Calypso IP will do close co-operation with other TKK's (Helsinki University of Technology) projects. This means shared code and research of IP routing in Mobile IP networks with GO/Core project and some consulting in implementation of various IP-protocols and IP-routing that are offered to Scoms project.

Calypso IP is also dealing with another research area than routing technology research. One of the project researchers is forming idea how ordinary protocol tester environment can be used for conformance testing of CORBA objects. There will be model of this idea and some example implementations. This kind of research has been done in only one or two other projects in the whole world, and that is why this topic has been taken under study in this project.