Embedded Systems November 2000 Vol13_12

Issue link:

Contents of this Issue


Page 131 of 189

However, even if the transmission delay is constant, the problem of out-of-date information still exists. Two di ffere nt types of problems are caused by message delays. One type of problem resul ts from variable delays Uitter). That is, the time taken for messages to reach the destination may vary significantly. The delays depend on a numbe r of facto rs, such as the route taken through the communica- tion medium, congestion in the medi- um, congestion at the processing sites (for example, a busy receiver ), inter- mitten t hardware failures, and so on. If the transmission delay were con- stant, we could better assess when a message has been lost. For this reason, some communication networks are designed as synchronous networks, so that delay values are fixed and known in advance. Howevet~ even if the transmission delay is constan t, the problem of out- ofdate information still exists . Since messages convey information about state changes between components of tl1e distributed system, tl1e informa- tion in tl1ese messages may be out of date if tl1e delays experienced are greater than the time required to change from one state to the next. This can lead to unstable systems. Imagi ne trying to drive a car in a situ- ation where the visual input to your eyes is delayed by several seconds. Transmissio n delays also lead to a edgement protocol. In such protocols, the receiver notifies the sendet· when it receives a message. Of course, there is the possibili ty that the acknowledge- ment message itself will be lost, so that such protocols are merely an opti- mizatio n and not a solution. The most common technique for detecting lost messages is based on time-outs. Namely, if we do not get a positive acknowledgement that our message was received within some rea- sonable time in terval, we conclude that it was dropped somewhere along the way. The difficul ty of this approach is distinguishing whether a message (or its acknowledgement) is simply slow or actually lost. If we make the time-out interval too short, we risk duplicating messages and, in some cases, reordering. If we make th e inte rval too lo ng, the system may become unresponsive. Transmission delays While transmission delays are not nec- essarily failures, they can certainly lead to failures. We've already discussed the situation where a delay can be miscon- strued as a lost message. 130 NOVEMBER 2000 Embedded Systems Programming complex situation that we will refer to as the relativistic effect. This is because transmission delays between different processing sites in a distributed system may be different. As a result, diffe rent sites could see the same set of mes- sages but in a diffe rent order. In Figure 2 case, distributed sites NotifierP and NotifierQ each send out a notification about an event to the two clients (ClientA and ClientS). Due to tl1e different routes taken by the individual messages and the differ- ent delays along those ro utes, we see that C L i entB sees o ne sequence ( event1 followed by eventZ), whereas C L i entA sees anothe r. As a conse- quence, th. e two cl ien ts may reach dif- fe rent conclusio ns about the state of the system.

Articles in this issue

Archives of this issue

view archives of EETimes - Embedded Systems November 2000 Vol13_12