Question

Show how to construct a send-constrained channel from a receive-constrained channel, and vice versa. Hint: use a trusted node connected to the given channel

Answer 1

Figures & Tables

• 

  Figure 1

• 

  Figure 2

• 

  Figure 3

• 

  Figure 4

2 Related work

Context-awareness has been identified as a key issue in nomadic computing with location being the most prominent contextual parameter . Contextual data are usually collected via sensing technologies, e.g., GPS and the Active Badge

However, to our knowledge, very little has been published on authenticating contextual data. Many sensing-based approaches are intrinsically difficult for use in authentication

Any one-way channel c has message send and receive operations as follows: m = c.receive() c.send(m). We denote the principals that perform those operations for a uniquely identified message m as receiver(m) and sender(m), respectively.

A constrained channel is a one-way communication channel that is either send-constrained or receiveconstrained or both (Figure 1):

send-constrained channel scf on the predicate f: if m = scf.receive() then f (sender(m)).

receive-constrained channel rcf on the predicate f: f(receiver(m)) for any message m appearing in an operation rcf.send(m).

1 Implementing constrained channels

The telephone system shows how we can exploit a channel with appropriate mechanical characteristics. Our telephone examples assume that the series of links and logic circuits connecting one telephone to another is proof against tampering, and that telephone circuits are not physically moved. That type of assumption is more likely to hold reliably in the case of short-range wired links in physically secured environments

Constrained propagation

One way of establishing location information is to use network time-of-flight. In principle, with sufficiently accurate instrumentation and knowledge of real-time system parameters, we can use round-trip times to bound the location of a network node. To gauge a bound on the distance to node M, node N sends a 1-bit message to it, which M is to return to N immediately. If the speed of signal propagation is c then, if M can return the message to node N in time t < (l + 2d/c), it is within a distance d of N, where l is the total communication latency imposed by software and hardware

2 Constrained channels as building blocks We can use constrained channels as building blocks for making further constrained channels. We do so by inserting a proxy between two constrained channels; and by running a protocol that turns a send-constrained channel into a receive-constrained channel, or vice versa.

The self-verification protocol

Interestingly, a constrained channel can be used to verify a principal’s own location. In the case of a receive-constrained channel, a principal can send a message to a receive-constrained channel rclL for location L and check if he can receive the same message from this channel. The principal has to know how to send a message to rclL but the self-verification protocol itself is trivial once the receive-constrained channel is available. The difficulty lies in how to construct rclL so that its guarantees are securely implemented

5 Conclusion

We have described a model of send- and receive constrained channels for context authentication. We have shown how to construct simple examples of constrained channels on location predicates. For this purpose, we use physical communication channels that are subject to mechanical or signal propagation constraints.

Applicability

In Section 1, we listed six problems that exemplify the types of context authentication in which we are interested. The first four are examples of location authentication; the third and fourth involve selfverification of location (a computer inside a building, a kiosk near to a human carrying a short-range radio transceiver). We have shown how to achieve those types of authentication