[net2-wg] Description of Zigbee Core tasks

[Joe Polastre]

Documents
=======================
Zigbee 1.0 Specification
Available at www.zigbee.org

IEEE 802.15.4 Specification
Available at standards.ieee.org


Summary Zigbee Services
=======================

Network Routing
  - Cost based (very similar to MultiHopLQI)
  - Route discovery and table (TinyAODV)
  There are minor semantics to Zigbee routing that differs from AODV,
such as restrictions on which nodes may be routers and which nodes may
only be endpoints

Communication Scheduling
  - Leaves are battery powered and act as children in a star
  - Routers are externally powered and act as router in the mesh and
parents in each star

Security
  - MAC, Network, and Application security available

Packet Formats
  - PHY, MAC, Network, and Application headers

Initial Work to be Done
=======================

1) Define net2 architecture, document, and start decomposing Zigbee
into its components
2) Port TinyAODV to net2 architecture
3) Build support around TinyAODV for Zigbee routing, including
specification of nodes as routers and endpoints and Zigbee-defined
interfaces
4) TinyOS support for Zigbee "Service Access Points (SAP)" composition
primitives
5) Develop sufficient 802.15.4 MAC layer, specifically beacon frames
and coordinator modes (association)
6) Composition of 802.15.4, SP, and TinyAODV/Zigbee in a decoupled manner
7) Security architecture, implementation, and coordination between layers
8) Passing of packets, packet headers, and encapsulation across layers

I'm not specifying the Zigbee Application Presentation part as a goal
of the initial work, because I believe it is more important to get the
functionality and architecture in place before customizing for
particular Zigbee application profiles.

Items 1-4 should be straight forward.  5 may be trickier, but I think
this group has enough knowledge of 15.4 that we could do it.  Items
6-8 are more challenging and interesting because they define how well
the net2 architecture works, and how developers should think about
composing protocols in the future.