Low delay or low loss (LDOLL) queuing policies for asynchronous transfer model (ATM) fast packet switching allow either retrieval priority or storage priority to be given to the ATM cells of traffic streams with different performance requirements. The improvements in performance that can be achieved by using one bit in the ATM cell header to make the binary LDOLL distinction are analyzed. In the analysis, simulation of the level, where the cell transmission rate of a source may vary between zero and the peak rate and computation at the level, where queuing of individual cells is considered, are combined. It is shown that LDOLL queuing, compared with FIFO queuing in ATM switches, can reduce cell loss probability by many orders of magnitude for low loss traffic. Low delay traffic incurs a slightly higher cell loss which is compensated with a reduced cell delay mean and variance.>

The power of ATM (asynchronous transfer mode) is its ability to provide bandwidth on demand, different sources can have different bandwidth requirements. Sources also differ in performance requirements, some ask for minimal delay variations, whereas others must have extremely low cell loss probabilities. It is shown how these complementary performance requirements can be explained with an LDOLL (low delay or low loss) queue, where sources get either service priority or storage priority. The space of possible LDOLL queuing policies is very large, even after a justified reduction, the size is still O(2/sup Q2/), Q being the maximum number of ATM cells in the LDOLL queue. Using Markov decision theory and concepts of linear programming, only Q so-called efficient solutions are achieved. These are the LDOLL threshold policies, which are conceptually appealing, robust in performance, and practical from the implementation viewpoint.>

Adding frame structure to slotted ALOHA makes it very convenient to control the ALOHA channel and eliminate instability. The frame length is adjusted dynamically according to the number of garbled, successful, and empty timeslots in the past. Each terminal that has a packet to transmit selects at random one of thentimeslots of a frame. Dynamic frame length ALOHA achieves a throughput (expected number of successful packets per timeslot) of 0.426 which compares favorably with the1/e (\approx0.368)upper bound of ordinary slotted ALOHA.

The problem of routing in a computer networks is dealt with by separating the information problem and the control problem. Several classes of information policies are examined, and numerical comparisons are made of their performance, using representative values of network parameters. Routing based on the expected values of delay, rather than actual values of delay, is shown to be superior to the other algorithms considered.