Victor Wen , Adrian Perrig, Robert Szewczyk
SPINS: Security Suite for Sensor Networks
http://citeseer.nj.nec.com/perrig01spins.html

Summary: (contributed by Suvda Myagmar)

First some background information. This paper is a part of embedded
systems project conducted at UC-Berkeley. Earlier this semester we also
discussed "The BiBi- one time signature and broadcast authentication
protocol" paper by Adrian Perrig. Some of the authors published paper on
TESLA: Timed, Efficient, Streaming, Loss-tolerant Authentication
protocol. This paper utilizes the micro version of this protocol.

Sensor networks is a set of small devices that measure environmental
parameters. When turned on, these devices automatically initiate route
discovery to the base station. In certain applications like tracing enimy
movement in Afganistan, it's critical that the base station receives
untampered information from the sensor nodes. Thus, we need security for
sensor networks.

The authors devided the security problem into 2 parts:
1. SNEP-data confidentiality, two-party data authentication, data
freshness
2. muTESLA- authenticated broadcast

Sensor nodes are very power devices. They have very small memory,
processing power, and low communication bandwidth. Because of these
limited resources regular security protocols can't be used in sensor
networks. Public key cryptography has too much data overhead. So the
authors use secret key encryption. Freshness of data is guaranteed by
using nonce (unpredictable bitstring). Two-way authentication is done by
attaching the encryption counter. For authenticated broadcast, authors
use delayed  key disclosure scheme to achieve asymmetric authentication.

Assumptions we made during the discussion: The secret key is hardwired in
each node, thus each node can talk to only one particular base station.
The base station keeps an encryption counter for each node. Each node can
tell from the packet whether this node is the intended recipient of this
packet.

We concluded that the paper introduces a very innovative, robust idea. We
wished to see some numbers on how long does it take to brake the
encryption key using very powerful machine. After all the encryption key
is never changed.

Strong accept: 5 out of 5