ATLANTIS: NEMO Basic Support Implementation

$Id: atlantis.html,v 1.6 2005/01/26 09:39:07 mitsuya Exp $

Introduction

As shown in development and spread of cellular phones, there are big needs for communications in the mobile environments. People primitively want to communicate at anytime and anywhere. Especially when people take vehicles, it becomes more likely because we don't have nothing to do else.

A vehicle consists of a group of nodes such as sensor nodes and devices held by passengers is connected to the Internet through several network devices. It is typical for buses, trains and airplanes to have such Mobile Internet environment. The nodes in such vehicles have to change the attachment point of the Internet frequently. But the nodes would then not be able to maintain transports and higher-layer connections if it changes the attachment point. Thus, it is important to provide mobility support like Mobile IP. These nodes include low cost network applications with only limited space for extra functions. It is preferred that a solution has no impact on these low cost nodes. Therefore, existing Host Mobility Protocol is not suitable for this situation.

The IETF NEMO (NEtwork MObility) working group has been created in fall 2002 to standardize a solution for this problem. The NEMO WG has decided to first specify an easy, straight forward, but not optimal solution for this problem referred to as NEMO Basic Support. The solution is to set up a bidirectional tunnel between the home agent (HA) and a mobile router much like it is done in Mobile IPv6. The first version of this solution has been issued in June 2003.

ATLANTIS is an implementation of NEMO Basic Support by Nautilus6 Project which is a working group in WIDE.

News

29 July 2004: Update Introduction.
24 July 2004: This page has been created.

Overview of Nautilus6 NEMO Implementation

Based on draft-03
NetBSD 1.6.1 w/ KAME MIPv6
- Currently, snap 20040126
Support both MR and HA
Support Explicit and Implicit mode

HA supports both mode at once
MR switches their mode by configuration

Support DHAAD
IPSec between HA-MR

Current Status

MR-MR communication works
Nested mobility works
- VMN, VMR
- Just tried 1st level of nested
Support Simple Multiple I/F switching
- Put preference on each I/F
- Just using KAME MIPv6 function
Interoperability
- Connectathon 2004
- With Nokia and another from Keio Univ.

Evaluation of NEMO Basic Support Protocol

We performed evaluation of NEMO Basic Support Protocol using our implementation.

Scenarios

LFN: simple NEMO (1 level)
LFN: netsted NEMO (> 1 levels)
LFN: simple NEMO with some of the 8 multihoming cases
LFN: nested NEMO with some of the 8 multihoming cases

Metrics/Parameters

Topology
packet overhead (=bandwidth used for signaling (% signal/data)
End to End performance (Round Trip Time, Packet Loss and Bandwidth)
Handover Overhead (Latency and Packet Loss)
Amount of Encapsulation

In the conclusion, a NEMO approach is better than everyone in a bus using its own MIP6.

Downloading and Installation

We cannot keep the source code of our implementation open because the IPR issues are still unclear. Now, we are asking our lawyer. So please wait a month for the release as open source code. For the details of the IPR issues, you can refer to the KAME Newsletter.

Anyway, you can get our code right now with some agreements. For the detail of this, please ask us via e-mail, root @ nautilus6.org.

For Installation, please refer README.nemo on the first directory in the source code.

Interoperability Testing

We are always positive to have a interoperability testing with your NEMO implementation. Actually, we already confirmed our implementation has the interoperability with others.

For the details of this, please ask also us via e-mail, root @ nautilus6.org.

Developer Information

Koshiro MITSUYA
A doctor student at Jun Murai Laboratory, Graduate School of Media and Governance, Keio University, Japan. For the details, please refer to his home page.