More than 30 years have been past since the current Internet architecture was designed. Since then, several functions that were not taken into consideration have been added to the original Internet architecture like “patchwork.” Such an approach may lead the Internet to failure in the near future. Therefore, “clean slate” approach that aims at redesigning the Internet from scratch has attracted attention.
There are several proposals of new network architectures not based on the layering model. However, Teraoka Lab. concluded that the layering model is appropriate for New Generation Network and proposes a six-layer network architecture called ZNA (Z Network Architecture). Some of the features of ZNA are as follows:
- Insert the session layer between the application layer and the transport layer to provide the application layer with more sophisticated communication schemes.
- Separate identifiers and locators of nodes to support mobility, multihoming, and security.
- Support heterogeneity of network layer protocols to interconnect networks that use the new network layer protocol and those that use legacy network layer protocols such as IPv4/v6.
- Separate the data plane and the control plane to reliably transmit control packets regardless of data traffic.
- Introduce Abstract Entities and Inter-Layer System in the control plane to support cross-layer collaboration.
- Support scalable QoS-guaranteed communication.
We are designing and implementing the protocol in each layer.
On the other hand, we are focusing on problems that should be solved in the current Internet. One of them is AAA (Authentication, Authorization, and Accounting), which is indispensable for service providers to provide services to users. As the first step to design the AAA architecture, we employed the Diameter architecture standardized in IETF. We are designing and implementing a universal AAA infrastructure based on Diameter. As the second step, we will evaluate the universal AAA infrastructure. If we conclude that the Diameter architecture is not suitable for New Generation Network, we will design a new AAA architecture as the third step.
Mobility support and QoS-guaranteed communication are also problems that should be solved in the current Internet. One of our goals of mobility support is to provide a high-speed train with a broad band Internet connectivity. We are proposing a fast handover mechanism using cross-layer collaboration. We have implemented our fast handover mechanism and had an experiment using an automobile and a train in service.
To provide QoS-guaranteed communication, a lot of mechanisms have been proposed and standardized such as path computation, signaling protocols, admission control, and queue scheduling. However, there is no real system that provides QoS-guaranteed communication in which such various mechanisms are well combined. We are designing an architecture for a system that provides scalable QoS-guaranteed communication.
From the viewpoint of data handling in the Internet, the data volume we are dealing with has been increasing year by year. Such large volume data must be able to be saved and retrieved smoothly to/from the network and must be persistent even if there is hardware failure in the network. We are designing a distributed file storage system that can smoothly save and retrieve large volume data to/from the network with keeping persistency of the data.
March 2011, Fumio Teraoka