Abstract: This paper introduces a new aspect of queueing theory, the study of systems that service customers with specific timing requirements (e.g., due dates or deadlines). Unlike standard queueing theory in which common performance measures are customer delay, queue length and server utilization, real-time queueing theory focuses on the ability of a queue discipline to meet customer timing requirements, for example, the fraction of customers who meet their requirements and the distribution of customer lateness. It also focuses on queue control policies to reduce or minimize lateness, although these control aspects are not explicitly addressed in this paper. To study these measures, we must keep track of the lead times (dead-line minus current time) of each customer; hence, the system state is of unbounded dimension. A heavy traffic analysis is presented for the earliest-deadline-first scheduling policy. This analysis decomposes the behavior of the real-time queue into two parts: the number in the system (which converges weakly to a re flected Brownian motion with drift) and the set of lead times given the queue length. The lead-time profile has a limit that is a nonrandom function of the limit of the scaled queue length process. Hence, in heavy traffic, the system can be characterized as a diffusion evolving on a one-dimensional manifold of lead-time profiles. Simulation results are presented that indicate that this characterization is surprisingly accurate. A discussion of open research questions is also presented.

Abstract: A distributed computing system having (host and I/O) end nodes, switches, routers, and links interconnecting these components is provided. The end nodes use send and receive queue pairs to transmit and receive messages. The end nodes use completion queues to inform the end user when a message has been completely sent or received and whether an error occurred during the message transmission or reception process. A mechanism implements these queue pairs and completion queues in hardware. A mechanism for controlling the transfer of work requests from the consumer to the CA hardware and work completions from the CA hardware to the consumer using head and tail pointers that reference circular buffers is also provided. The QPs and CQs do not contain Work Queue Entries and Completion Queue Entries respectively, but instead contain references to these entries. This allows them to be efficient and constant in size, while the Work Queue Entries and Completion Queue Entries themselves can vary in size, for example to include a variable number of data segments. Additionally, several mechanisms are provided to improve the overall efficiency of this process under different memory configurations.

Abstract: A system for allocating bandwidth resources among various mobile stations which are wirelessly connected to a base station. The length of the data queue in each mobile station is determined and information regarding that length is placed in a field in the outgoing data packet. When it is received in the base station, this field is decoded and the queue length information used to allocate bandwidth resources among the mobile station connections. This allows a very quick response to data queue lengths and accordingly better service.

Abstract: A wireless phone includes a memory, a processor, and a call queue function, the call queue function to enable the configuration and storage in the memory of a call queue, the call queue comprising an ordered list of entries to dial, wherein the processor and the memory cooperate to enable the call queue function.

Abstract: In a multi-QOS level queuing structure, packet payload pointers are stored in multiple queues and packet payloads in a common memory pool. Algorithms control the drop probability of packets entering the queuing structure. Instantaneous drop probabilities are obtained by comparing measured instantaneous queue size with calculated minimum and maximum queue sizes. Non-utilized common memory space is allocated simultaneously to all queues. Time averaged drop probabilities follow a traditional Weighted Random Early Discard mechanism. Algorithms are adapted to a multi-level QOS structure, floating point format, and hardware implementation. Packet flow from a router egress queuing structure into a single egress port tributary is controlled by an arbitration algorithm using a rate metering mechanism. The queuing structure is replicated for each egress tributary in the router system.

Abstract: This research extends the dynamic user optimal assignment under the point queue concept so as to deal with physical queues. Given time-dependent many-to-many OD volumes, this paper first shows the formulation of the assignment subject to the flow conservation and the first-in-first-out (FIFO) queue discipline. The optimal condition is then defined and the physical queue propagation based on the kinematic wave theory is discussed. Finally, a solution algorithm is proposed and typical differences between point and physical queue analyses are presented through an example calculation.

Abstract: A scheduler allocates service to enqueued cells of connections provisioned for guaranteed service levels employing a structure with one or more of the following features to achieve efficient high-speed packet switching (cell relay). For very high-speed switching of connections, such switches operating up to 10, or even 100, Tbps, burst scheduling of cells is employed in which a number of cells, termed a burst, are serviced when a queue is eligible for service. When a queue has less than the number of cells in the burst (termed a short burst), the scheduler still schedules service, but accounts for saved service time (or bandwidth) of the short burst via queue length when the queue's eligibility is next considered for service. In addition, high and low bandwidth connections of a queue may be allocated into two sub-queues, with priority assigned to the two queues and delay-sensitive traffic (high bandwidth connections) assigned to the higher priority sub-queue. High-bandwidth queues may be assigned high and super-high priority to account for ties in scheduling decisions for eligible queues. The scheduler may tentatively schedule a queue as eligible for service, but employ information generated from output ports indicating an output port's congestion to decide whether a bid for service of the eligible queue should be generated. The scheduler may allocate service to at least two different traffic types separately, e.g., standard traffic queues and guaranteed bandwidth traffic queues, based on traffic type specific algorithms that determine an eligibility for a queue associated with each type of traffic. When eligibility of queues for each type of traffic is determined, then another scheduling algorithm allocates service to each eligible queue of the traffic types based on, e.g., a rate proportional server algorithm.

Abstract: In a video-on-demand environment, batching of video requests is often used to reduce I/O demand and improve throughput. Since viewers may defect if they experience long waits, a good video scheduling policy needs to consider not only the batch size but also the viewer defection probabilities and wait times. Two conventional scheduling policies for batching are the first-come-first-served (FCFS) policy, which schedules the video with the longest waiting request, and the maximum queue length-(MQL) policy, which selects the video with the maximum number of waiting requests. Neither of these policies leads to entirely satisfactory results. MQL tends to be too aggressive in scheduling popular videos by considering only the queue length to maximize batch size, while FCFS has the opposite effect by completely ignoring the queue length and focusing on arrival time to reduce defection. In this paper, we introduce the notion of factored queue length and propose a batching policy that schedules the video with the maximum factored queue length. We refer to this as the MFQL policy. The factored queue length is obtained by weighting each video queue length with a factor which is biased against the more popular videos. An optimization problem is formulated to solve for the best weighting factors for the various videos. We also consider MFQL implementation issues. A simulation is developed to compare the proposed MFQL variants with FCFS and MQL. Our study shows that MFQL yields excellent empirical results in terms of standard performance measures such as average latency time, defection rates, and fairness.

Abstract: A call center includes functionality for dynamically positioning newly received calls within an established call queue. The calls are positioned within the queue based on predefined service objectives for call types supported by the call center and a length of time that other calls have already been waiting in the queue. In a preferred embodiment, when a new call is received by the call center, an analysis is performed for individual queue positions within the queue until a queue position is identified that meets a predetermined selection criterion. The new call is then placed within this queue position and subsequent calls within the queue are moved accordingly.

Abstract: Managing memory access to random access memory includes fetching a read lock memory reference request and placing the read lock memory reference request at the end of a read lock miss queue if the read lock memory reference request is requesting access to an unlocked memory location and the read lock miss queue contains at least one read lock memory reference request.

Abstract: An input/output bus bridge and command queuing system includes an external interrupt router for receiving interrupt commands from bus unit controllers (BUCs) and responds with end of interrupt (EOI), interrupt return (INR) and interrupt reissue (IRR) commands. The interrupt router includes a first command queue for ordering EOI commands and a second command queue for ordering INR and IRR commands. A first in first out (FIFO) command queue orders bus memory mapped input output (MMIO) commands. The EOI commands are directed from the first command queue to the input of the FIFO command queue. The EOI commands and the MMIO commands are directed from the command queue to an input output bus and the INR and IRR commands are directed from the second command queue to the input output bus. In this way, strict ordering of EOI commands relative to MMIO accesses is maintained while simultaneously allowing INR and IRR commands to bypass enqueued MMIO accesses.

Abstract: A distributed computing environment and associated method for monitoring a queue-based messaging system. The queue-based messaging system controls the exchange of messages between a server process and client process applications. A messaging application residing at the server computer platform and forming part of the queue-based messaging system manages a plurality of trigger-initiated local queues, each associated with one of the plurality of client process applications and having queue depth, trigger enable, get enable and put enable attributes. A monitoring tool residing on the server computer platform acquires a value for the queue depth, trigger enable, put enable and get enable attributes for a plurality of trigger-initiated local queues and generally simultaneously displays, on a user interface coupled to the server computer platform, the value for the queue depth, trigger enable, put enable and get enable attributes for each one of the plurality of trigger-initiated local queues.

Abstract: Methods, apparatus, and articles of manufacture for retaining packet order in multiprocessor systems utilizing multiple transmit queues are disclosed herein. Embodiments of the present invention define multiple transmit queues for a given priority level of packets to enable the multiprocessor system to process and queue packets of equal priority in different transmit queues. Queuing packets of equal priority in different transmit queues minimizes processor time spent attempting to acquire queue-specific resources associated with one particular transmit queue. In addition, embodiments of the present invention provide an assignment mechanism to ensure that packets corresponding to a common flow are queued in the same transmit queue in order to eliminate, to the extent possible, out-or-order packets, which many times results in lost packets and a reduction in realized network throughput.

Abstract: A communications network is described having a class-based queuing architecture. Shared class queues receive packet flows from different customers. In one embodiment, there are eight classes and thus eight shared queues, one for each class. A scheduler schedules the output of packets by the various queues based on priority. Each customer (or other aggregate of packet flows) is allocated a certain space in a class queue based on the customers' Service Level Agreement (SLA) with the service provider. A queue input circuit detects bits in the packet header identifying the customer (or other criteria) and makes selections to drop or pass packets destined for a shared queue based on the customers' (or other aggregates') allocated space in the queue. In another embodiment, the relative positions of the nodes in the network are taken into account by each node when dropping packets forwarded by other nodes by detecting a node label (or other ID code) so that packets from the various nodes are dropped in a more fair way when there is congestion in the network, irrespective of the “passing ” node's position relative to the other nodes.

Abstract: A queuing system, method and computer program product is provided for transferring electronic information over a data network, such as the Internet. A host or system administrator monitors host resources by preventing one or more host resource limits from being exceeded. Where a resource limit is reached, a queuing protocol is initiated with a client requesting data. The queuing protocol permits the client to enter a queue and wait until host resources may be allocated to service the data request. The queue is maintained by the host, which may provide periodic updates to the client regarding queue length, position and estimated wait time. Clients may advance through the queue in accordance with a first-in-first-out algorithm, or some alternative priority scheme. When a client reaches the head of the queue, host resources are checked. If sufficient resources are available to service the client request, they are allocated to the client at the head of the queue.

Abstract: Quality-of-service (QoS) guarantees in networks are increasingly based on per-flow queueing and sophisticated scheduling. Most advanced scheduling algorithms rely on a common computational primitive: priority queues. Large priority queues are built using calendar queue or heap data structures. To support advanced scheduling at OC-192 (10 Gbps) rates and above, pipelined management of the priority queue is needed. We present a pipelined heap manager that we have designed as a core integratable into ASICs, in synthesizable Verilog form. We discuss how to use it in switches and routers, its advantages over calendar queues, and we present cost-performance tradeoffs. Our design can be configured to any heap size. We have verified and synthesized our design and present cost and performance analysis information.

Abstract: An apparatus and method for virtualizing a queue pair space to minimize time-wait impacts. The apparatus and method allocate virtual queue pairs from a virtual queue pair pool of a node to connections between the node and other nodes. The connection is established between a physical queue pair of the node and physical queue pairs of other nodes. However, from the viewpoint of the other nodes, they are communicating with the present node using the virtual queue pair and not the physical queue pair for the present node. By using the virtual queue pairs, the same physical queue pair may accommodate multiple connections with other nodes simultaneously. Moreover, by using a virtual queue pair rather than a physical queue pair, when a connection is torn down, the virtual queue pair is placed in a time-wait state rather than the physical queue pair. As a result, the physical queue pair may continue to function while the virtual queue pair is in the time-wait state.

Abstract: A traffic manager for a network switch port stores incoming cells in a cell memory and later forwards them out of the cell memory and the switch port. Each cell is assigned to one of several flow queues and each flow queue has an assigned minimum forwarding bandwidth with which cells of that flow queue must be forwarded from the cell memory and has an assigned maximum bandwidth with which cells of that flow queue may be forwarded. When any flow queue is active (i.e., when it has cells currently stored in the cell memory), the traffic manager allocates a sufficient amount of the switch port's available cell forwarding bandwidth to each active flow queue so that cells of that flow queue are forwarded with at least the flow queue's assigned minimum bandwidth. Each flow queue also has an assigned forwarding weight, and the traffic manager also dynamically allocates a portion of the switch port's excess forwarding bandwidth, above that needed to accommodate each active flow queue's minimum bandwidth, among all active flow queues in relative proportion to each active flow queue's assigned forwarding weight. Thus the actual forwarding bandwidth allocated to each active flow queue is the sum of its assigned minimum forwarding bandwidth and its allocated portion of excess bandwidth. However the traffic manager limits the actual forwarding bandwidth allocated to any one flow queue so that it does not exceed the flow queue's assigned maximum forwarding bandwidth.

Abstract: A dispatcher (130) in a multiprogramming or multitasking operating system in a data processing system selects the next tasks to be executed by an available processor. Access to shared resources are controlled by locks and queues, where tasks are queued when they find the shared resource locked, and dequeued one by one as the lock is unlocked. When a lock is unlocked, the first task in a FIFO queue is dispatched with a temporary priority (140) at least as high as any in the queue. This first task must retain this temporary urgency until it releases the resource or until its urgency is further increased due to the addition of a higher priority task (144) to the resource queue or a dependent resource queue. This prevents starvation of higher priority tasks waiting in the FIFO queue.

Abstract: This paper considers the queue length distribution in a class of FIFO single-server queues with (possibly correlated) multiple arrival streams, where the service time distribution of customers may be different for different streams. It is widely recognized that the queue length distribution in a FIFO queue with multiple non-Poissonian arrival streams having different service time distributions is very hard to analyze, since we have to keep track of the complete order of customers in the queue to describe the queue length dynamics. In this paper, we provide an alternative way to solve the problem for a class of such queues, where arrival streams are governed by a finite-state Markov chain. We characterize the joint probability generating function of the stationary queue length distribution, by considering the joint distribution of the number of customers arriving from each stream during the stationary attained waiting time. Further we provide recursion formulas to compute the stationary joint queue length distribution and the stationary distribution representing from which stream each customer in the queue arrived.

Abstract: We propose an adaptive bandwidth allocation algorithm for reservation protocols to support quality-of-service networking on the future multimedia Internet. QoS design is the fundamental functionality of the next generation IP router to enable differentiated delivery and to guarantee the delivery quality for diverse service traffic. The proposed algorithm called adaptive weighted fair queue (AWFQ) scheduling employs the queue status and priority assignment to determine the bandwidth sharing of various Internet services and to ensure the defined QoS policy is obtained. We propose a smart methodology of distributing the outlink bandwidth in a flexible, fair and prioritized manner to utilize network bandwidth more effectively and efficiently. In simulation, we compare the proposed AWFQ algorithm with priority queue (PQ) and weighted fair queue (WFQ) that are two famous scheduling algorithms in practice. In comparison, we examine their performances in terms of packet loss, bandwidth utilization and queue dynamics to verify our algorithm. We find that the proposed algorithm has the following advantages: 1) improves the bandwidth starvation of low-priority traffic occurring in PQ, 2) has a high bandwidth utilization close to PQ, 3) has a low queue congestion and 4) maintains the fairness of WFQ. Overall, the proposed algorithm not only has an excellent performance close to PQ but also improve the disadvantage of PQ.

Abstract: The present invention provides a method and apparatus for scheduling message processing. The present invention provides a scheduling mechanism, or scheduler (602), that receives messages (608) and stores the messages (610 or 614) in a first queue (604) or a second queue (606) based, in part, on various criteria associated with the messages. The criteria include message attributes, such as message priority, virtual private network ('VPN') classification and destination software function. The first queue (604) can be a first-in-first-out queue, and the to second queue (606) can be a multi-dimensional queue. The scheduler (602) then schedules the queued messages (612 or 616) for processing based, in part, on various operating criteria (618), such as historical operating data, current operating data and anti-starvation criteria. In addition, the scheduler (602) can be programmed to function in a variety of operating modes.

Abstract: The Internet research community is promoting active queue management in routers as a proactive means of addressing congestion in the Internet. Active queue management mechanisms such as Random Early Detection (RED) work well for TCP flows but can fail in the presence of unresponsive UDP flows. Recent proposals extend RED to strongly favor TCP and TCP-like flows and to actively penalize `misbehaving' flows. This is problematic for multimedia flows that, although potentially well-behaved, do not, or can not, satisfy the definition of a TCP-like flow. In this paper we investigate an extension to RED active queue management called Class-Based Thresholds (CBT). The goal of CBT is to reduce congestion in routers and to protect TCP from all UDP flows while also ensuring acceptable throughput and latency for well-behaved UDP flows. CBT attempts to realize a `better than best effort' service for well-behaved multimedia flows that is comparable to that achieved by a packet or link scheduling discipline, however, CBT does this by queue management rather than by scheduling. We present results of experiments comparing our mechanisms to plain RED and to FRED, a variant of RED designed to ensure fair allocation of bandwidth amongst flows. We also compare CBT to a packet scheduling scheme. The experiments show that CBT (1) realizes protection for TCP, and (2) provides throughput and end-to-end latency for tagged UDP flows, that is better than that under FRED and RED and comparable to that achieved by packet scheduling. Moreover CBT is a lighter-weight mechanism than FRED in terms of its state requirements and implementation complexity.

Abstract: This paper exposes an inherent weakness in current active queue manage ment techniques such asRedin that they rely on queue lengths to indicate the severity of congestion. In light of this observation, a fundamentally different active queue management algorithm calledBlueis proposed.Blueuses packet loss and link utilization to manage congestion. Using simulation and controlled experiments,Blueis shown to significantly outperformRedin providing lower packet loss rates and smaller queuing delays to networked applications such as interactive audio and video.

Abstract: A message queue server emulates a computer peripheral that not only supports communication between two mainframes, but also provides a gateway to open systems computers, networks, and other similar message queue servers. The message queue server provides protocol-to-protocol conversion from mainframes to today's computing systems in a manner that does not require businesses that own the mainframes to rewrite legacy applications to share data with other mainframes and open systems. The message queue server emulates a mainframe peripheral coupled to a first mainframe having a first protocol. The system includes digital storage to temporarily store information from the first mainframe. The system includes at least one manager that (i) coordinates the transfer of the information of the first protocol between the mainframe peripheral emulator and the digital storage and (ii) coordinates transfer of the information between the digital storage and (a) a second mainframe having a second protocol or (b) a computer network having a third protocol. Preferably, the message queue server emulates a tape drive and arranges the stored messages in a queue. Optionally, the message queue server manages the message queues as a function of information usually found in a standard tape label.

Abstract: A queue manager monitors status of a server queue in a network server and status of a device queue in a peripheral device at the same time. A user interface displays the status of jobs in the server queue and device queue on the same display and allows a user to manipulate any of the jobs on either queue using the same user interface.

Abstract: Disclosed is a packet transmitting apparatus for transmitting packets belonging to a plurality of groups having priorities that differ from one another. The apparatus includes a queue controller, which is provided in a scheduling unit, for generating queues on a per-group basis and giving a packet transmit privilege in order, in accordance with the round-robin method, to elements constituting each of the queues, queue by queue; and a packet-transmit group decision unit, which is provided in the scheduling unit, for deciding that a packet transmit group is a group having the highest priority among groups in which a packet corresponding to at least one queue element is awaiting to be transmitted. The queue controller gives the packet transmit privilege to the highest-priority queue element of a queue of the packet transmit group that has been decided and transmits the packet that corresponds to this queue element.

Abstract: Disclosed is a client/server-based transaction system which comprises a client workstation (300) and a transaction server (310). The client workstation (300) has an application program (320) not having any transaction management resources. The client comprises a transaction interface (325). The transaction server provides an input queue (330), an output queue (340) and a control queue (350). The control queue (350) resides on the transaction server. Start of a transaction (355), depicted as the dotted line, activates the transaction interface which initiates a data flow to the control queue and stores a unique session information. The data flow to the input queue stores a request to the transaction server. The completion of the transaction is signaled by a response in the output queue. The client retrieves this response, and deletes the control queue information in a same Unit of Work. In case the application program crashes before the transaction is completed, the session information in the control queue is used for recovery processing.