The first program, RUNDQ, was written by Wenxin Xu, modified by Byung Lee, and does a great job of demonstrating the heart of DQRAP, a pair of queues that dance across the screen as well as Astaire and Rogers. RUNDQ also provides throughput and delay. The default offered traffic is equal to 90% of capacity of a link and follows a Poisson distribution. Click here to download RUNDQ.
The second program, RUNFREE, is a variation on RUNDQ in that it demonstrates DQRAP operating in the free access mode, i.e., no apriori knowledge of the network by a station before it transmits. This is accomplished by operating the queues as LIFO. The throughput and average delay of free access DQRAP are the same as with DQRAP but operating LIFO means that at times the poor soul at the head of the queue sometimes waits a long time for the queue to run down. This is graphically demonstrated by RUNFREE which can be downloaded by clicking here. This program was also written by Wenxin Xu and modified by Byung Lee.
The third program, DQLAN, demonstrates DQRAP in a LAN supporting transmission of variable length frames. Whereas the previous demos demonstrated the operation of DQRAP by moving numbers that represent packets across the screen this program pushes the envelope with graphics to demonstrate the actual flow of frames. The offered traffic follows a Poisson distribution and is equal to approximately 90% of the capacity. This program was written by Chien-Ting Wu. Click here to retrieve DQLAN.
The fourth program demonstrates a remarkable variation of a remarkable protocol: XDQRAP (Extended DQRAP). DQRAP utilizes control minislots that carry an implied request for a single slot -- XDQRAP minislots carry requests for a multiplicity of slots. The upshot is that a station requests the number of slots required for an upper layer PDU, segments the PDU, and then transmits the segments without any encapsulation of the individual segments. Thus variable length frames are transmitted as a series of segments thus permitting higher priority transmissions to preempt the line and not wait till the end of transmission of the lower priority frame. The simulation demonstrates this feature -- the minislots carry one, two or three bars indicating requests for one, two, or three slots. Priority messages are shown preempting ongoing transmission of a lower priority message. Again this program was written by Chien-Ting Wu. Click here to retrieve XDQRAP.
The fifth program, WDQLAN, demonstrates DQRAP in a wireless half-duplex environment. This program was used to both demonstrate how DQRAP works and to obtain performance data. The scenario consists of eight remote stations and a central station. The original program permitted all the parameters shown in the upper left corner to be varied to reflect a bimodal traffic pattern. The conditions for the demo are fixed at 75% short packets and 25% long packets with a 50/50 split of traffic between inbound and outbound channels. The half-duplex operation means that there is a delay between the inbound requests by remote stations and the resulting feedback sent by the central unit. This does provide more time for the viewer to figure out what is going on and thus is a favorite of mine in explaining DQRAP. This program was also written by Chien-Ting Wu. Click here to obtain WDQLAN.