Abstract: Methods and apparatus are provided for self-organized caching in a content delivery network. One or more requested content items are cached at a node in a content delivery network. The disclosed method determines if the content item is stored in a local cache; and, if the content item is not stored in a local cache, the following steps are performed: requesting the content item from another node; and storing the content item in the local cache it one or more predefined capacity criteria are satisfied. The content delivery network can be a hierarchical network or a non-hierarchical network. The predefined capacity criteria can evaluate a popularity index of the content item relative to other items stored in the local cache.

Abstract: Systems for switching traffic include a physical machine running source and destination virtual machines (VMs). The source VM issues a data unit addressed to the destination VM. The physical machine has a physical network interface in communication with the VMs. The physical network interface transmits a sub-packet, which includes a partial portion of the data unit, over a network while a majority portion of the data unit remains at the physical machine. A network switch on the network receives the sub-packet transmitted by the physical network interface. The network switch performs one or more OSI Layer 2 through Layer 7 switching functions on the sub-packet and returns that sub-packet to the physical network interface. The physical network interface identifies the data unit stored in the memory in response to the sub-packet returned from the network switch and forwards the identified data unit to the destination VM.

Abstract: The prevent invention provides a method for performing channel estimation and detection in a Multiple-Input-Multiple-Output (MIMO) system, including: generating a demodulation reference signal, DM-RS, pattern for an extended cyclic prefix (CP); and estimating and detecting channel performance based on the generated DM-RS signal pattern. The DM-RS pattern supports from rank 1 to rank 8 patterns for 8 layer transmission. A transmitter, a receiver and a system thereof are also provided. The solution of the present invention improves channel estimation accuracy, saves channel estimation implementation and implementation complexity at terminals.

Abstract: Apparatus, systems and methods for receiving one or more input signals and providing output signals in various video, audio, data and mixed formats are described. One or more input processors receive the input signals. Each of the input processors provides one or more packetized signals corresponding to one or more of the input signals received at the input processor. Each output processor can receive one or more packetized signals and generate or more output signals. The output signals correspond to one or more of the input signals, additional locally generated signals or data relating to the signals or any combination of such signals. Use of a packet router according to the invention allows input signals encoded as one set of packetized signals to be recombined to provide additional packetized signals incorporating the same or different combinations of the packetized signals.

Abstract: Disclosed herein is a transmission apparatus including a receiving unit configured to receive ACK packets transmitted from a reception apparatus in response to transmission data items, an ACK storing unit configured to store a latest ACK packet which is last received among the ACK packets, and a TCP/IP processing unit configured to perform processing of receiving an ACK packet in accordance with the latest ACK packet supplied from the ACK storing means at a predetermined timing.

Abstract: A communication apparatus includes a plurality of first queues to which priorities or predetermined bandwidths are assigned respectively, a plurality of second queues which are corresponded to any of the plurality of first queues and to which the priorities are assigned respectively, a first writing portion which writes data blocks into any of the plurality of second queues or any of the plurality of first queues which is not corresponded to the second queues in accordance with a rule, a second writing portion which reads the data blocks stored in the plurality of second queues in order of the priorities assigned to the second queues respectively, and writes the data blocks into at least one of the first queues which is corresponded to the second queues, and a transmitting portion which reads the data blocks stored in the plurality of first queues in order of the priorities assigned to the first queues respectively or at time intervals corresponding to the bandwidths assigned to the first queues respectively,

Abstract: Methods and systems for using counters to monitor port buffers are described. In one embodiment, the method includes receiving data packets into a system port buffer having different queues. Also, counters are associated with the queues. Further, the method includes scanning the counters, in which the counters count the number of dropped data packets associated with an individual queue of the different queues. Moreover, the method includes determining if the number of dropped data packets for an individual queue is greater than a predetermined number. If the number of dropped data packets is greater than the predetermined number, then a message is generated.

Abstract: An overhead processor for data transmission in digital communications is disclosed. Incoming data is transmitted along a datapath. If there are two or more groups of incoming data, arriving separately, the initial group(s) of received data can be held in an elastic store until the arrival of additional group(s) of data, and upon the arrival of additional group(s) of data, all received data are combined and transmitted into flip-flop(s). The data is transmitted from said flip-flop(s) to a logic element to determine the new data context of imminent incoming data prior to any additional incoming bytes arriving along the datapath. Therefore, the number of overhead processors required for multi-byte data transmission is reduced, potentially reducing the number of required overhead processors in digital communications to 1.

Abstract: Systems, methods, devices, and processors are described for quality of service (QoS) packet scheduling in satellite communications systems. A packet received at the QoS packet scheduler may be assigned a virtual departure time utilizing novel self-clocked fair queuing techniques. The virtual departure time for a packet assigned to a queue may depend on a different weight assigned to the queues. Queues may be treated as low latency queues in some cases and may be provided with committed information rates in other cases. Low latency queues may be assigned weights equal to infinity, or the reciprocal of the weight equal to zero. Queues with committed information rates may assign different weights to packets depending on whether the rate that packets are received exceeds the committed information rate. Packets may then be scheduled based on their virtual departure time order.

Abstract: A telecommunication and multimedia management apparatus and method that supports voice and other media communications and that enables users to: (i) participate in multiple conversation modes, including live phone calls, conference calls, instant voice messaging or tactical communications; (ii) review the messages of conversations in either a live mode or a time-shifted mode and to seamlessly transition back and forth between the two modes; (iii) participate in multiple conversations either concurrently or simultaneously; (iv) archive the messages of conversations for later review or processing; and (v) persistently store media either created or received on the communication devices of users. The latter feature enables users to generate or review media when either disconnected from the network or network conditions are poor and to optimize the delivery of media over the network based on network conditions and the intention of the users participating in conversations.

Abstract: Methods for switching traffic include a physical machine running source and destination virtual machines (VMs). The source VM issues a data unit addressed to the destination VM. The physical machine has a physical network interface in communication with the VMs. The physical network interface transmits a sub-packet, which includes a partial portion of the data unit, over a network while a majority portion of the data unit remains at the physical machine. A network switch on the network receives the sub-packet transmitted by the physical network interface. The network switch performs one or more OSI Layer 2 through Layer 7 switching functions on the sub-packet and returns that sub-packet to the physical network interface. The physical network interface identifies the data unit stored in the memory in response to the sub-packet returned from the network switch and forwards the identified data unit to the destination VM.

Abstract: The architecture and techniques described herein can improve system performance with respect to the following. Communication between two interdependent hardware engines, that are part of pipeline, such that the engines are synchronized to consume resources when the engines are done with the work. Reduction of the role of software/firmware from feeding each stage of the hardware pipeline when the previous stage of the pipeline has completed. Reduction in the memory allocation for software-initialized hardware descriptors to improve performance by reducing pipeline stalls due to software interaction.

Abstract: In one embodiment, a method is provided. The method of this embodiment provides storing a packet header at a set of at least one page of memory allocated to storing packet headers, and storing the packet header and a packet payload at a location not in the set of at least one page of memory allocated to storing packet headers.

Abstract: An embodiment of the present invention provides a method of requesting bandwidth allocation in a wireless network, comprising using a partial bandwidth request by a mobile station (MS) operable in the wireless network to a base station (BS) operable in the wireless network, wherein the partial bandwidth request requests bandwidth for only a portion of all packets in an uplink (UL) queue.

Abstract: A routing method performed by node equipment includes: receiving a first frame including a wait number, incrementing the wait number, and storing the incremented wait number as a local wait number; receiving a second frame including a wait number of a destination node equipment, and comparing the wait number in the second frame and the local wait number; transmitting the second frame to an adjacent node equipment having a larger wait number than the local wait number, when the wait number in the second frame is larger than the local wait number; and returning the second frame to a source node equipment of the second frame, when the wait number in the second frame is larger than the local wait number but there is no adjacent node equipment having a larger wait number than the local wait number.

Abstract: A method, system, and apparatus are directed towards selectively concatenating data into a packet to modify a number of packets transmitted over a network based on a combination of network and/or send-queue metrics. In one embodiment, Nagle's algorithm is used for concatenating data into a packet. The concatenation may be selectively enabled based on heuristics applied to the combination of metrics. In one embodiment, the result may indicate that there should be a concatenation, or that data should be sent immediately, or that a current state for whether to concatenate or not should be maintained. The heuristics may include an expert system, decision tree, truth table, function, or the like. The heuristics may be provided by a user, or another computing device. In another embodiment, the concatenation may be enabled based on a conditional probability determined from the combination of metrics.

Abstract: A transmission device (100) outputs, to a reception device (200), a stored amount of packets in a first transmission buffer (105) just before a certain packet has been written to the first transmission buffer. A first reception buffer amount read unit (206) in the reception device reads a stored amount of packets in a first reception buffer just before the certain packet has been read from the first reception buffer. A storage amount addition unit (207) adds the stored amount of packets in the first transmission buffer and the stored amount of packets in the first reception buffer, and a correction unit (208) adjusts the frequency of a variable frequency oscillator in a reception timestamp timer 209 so that the resulting added value is a constant value.

Abstract: A packet scheduler has input connections, and data packets received on the input connections can be placed in queues. The packet scheduler includes a first scheduler, for identifying a first of said queues according to a first queue scheduling algorithm, and a second scheduler, for identifying a second of said queues according to a second queue scheduling algorithm. The packet scheduler determines whether the first of said queues contains a packet of data to be sent and, if so, it selects the first of said queues as a selected queue. If the first of said queues does not contain a packet of data to be sent, it selects the second of said queues as a selected queue. The packet scheduler then determines whether the respective packet of data can be sent from the selected queue, by maintaining a deficit counter indicating a current data quota for the respective queue, and also by maintaining a global deficit counter.

Abstract: A network device receives traffic associated with a network of intermediate network devices and user devices, classifies the received traffic, and allocates the classified traffic to traffic queues. The network device also schedules particular queued traffic, provided in the traffic queues and bound for particular intermediate network devices, using a hybrid weighted round robin (WRR) scheduler where the hybrid WRR scheduler schedules the particular queued traffic according to one of a 1-level WRR schedule, a 1.5 level WRR schedule, or a 2-level WRR schedule. The network device further provides the particular queued traffic to the particular intermediate network devices based on the scheduling of the hybrid WRR scheduler.

Abstract: A method for data transmission via data networks, in particular via ethernet-based data networks, having at least two participants; the data networks have at least one topological real-time field and one topological non-real-time field, and data telegrams sent via the data networks have at least one real-time part and at least one non-real-time part. By a coupling unit, data telegrams from the non-real-time field which arrive inside the real-time part are delayed and transmitted to the non-real-time part.

Abstract: A network device for processing data on a data network includes a port interface configured to receive a data packet from a data network and to send a processed data packet to an egress port of the plurality of ports, a packet evaluation module configured to parse the received data packet and modify the received data packet to form the processed data packet and a search engine configured to perform searches of lookup tables using parsed data packet values and to return search results to the packet evaluation module to assist in modifying the received data packet. At least one lookup table shares at least two different types of entries in that same at least one lookup table, where the search engine is configured to distinguish between the at least two different types of entries in that same at least one lookup table.

Abstract: A system and method for scheduling delivery of packetized data traffic at an output port of a network node featuring a hybrid queue with a unicast subqueue and a multicast subqueue. An interleave indicator for each unicast and multicast entry in the queue is used to toggle the header type field of the hybrid queue. Upon departure, a scheduler selects between the unicast subqueue and multicast subqueue as dictated by the state of the header type field. The disclosed interleave indicator techniques add minimal overhead to the system, allowing ready scaling with increasing port densities. Further, the data structure of the multicast subqueue can be selected to optimize the system and achieve desired performance characteristics.

Abstract: A node in a mobile ad-hoc network or other network classifies packets (a) in accordance with a first set of priority levels based on urgency and (b) within each priority level of the first set, in accordance with a second set of priority levels based on importance. The node: (a) queues packets classified at highest priority levels of the first and/or second sets in high-priority output queues; (b) queues packets classified at medium priority levels of the first set in medium-priority output queue(s); and (3) queues packets classified at low priority levels of the first and/or second set in low-priority output queue(s). Using an output priority scheduler, the node serves the packets in order of the priorities of the output queues. In such manner, orthogonal aspects of DiffServ and MLPP can be resolved in a MANET or other network.

Abstract: A method for providing ordered data packets to a processor is provided. The method comprises receiving a plurality of data packets, each comprising an order identifier. At least one out-of-order data packet of the plurality of data packets is placed in a priority queue. One or more correction packets are received. One or more missing data packets or corrupted data packets of the plurality of data packets are detected based on the one or more correction packets. The one or more missing data packets or corrupted data packets of the plurality of data packets are corrected based on the one or more correction packets. The one or more corrected data packets are provided to the processor. The at least one out-of-order data packet in the priority queue is provided to the processor.

Abstract: Techniques for using target issue intervals are provided. Request messages may identify the size of a data packet. A target issue interval may be determined based on the request messages. The target issue interval may be used to insert a delay between sending subsequent request messages.

Abstract: A system using cooperative communication with rateless codes is presented that uses communication transmission aspects of cooperative communication with rateless codes over Rayleigh fading channels and queuing aspects for buffering messages at intermediate relays. The system transmits a subsequent message while a current message is en route to the destination by receiving and buffering the current message in queues at intermediate relays. A relay with a best instantaneous communication link to the source may receive a message first and forward the received message to the destination. Alternatively, if inter-relay communication links are strong, all relays may cooperate simultaneously.

Abstract: The present invention provides a preemptive priority scheduling system and method for optimal load balancing of messages and preserving the lightweight allocation resources in an intersystem communication. The invention also provides a system and method for scheduling of messages of a plurality of classes in an intersystem communication.

Abstract: A transmit packet generated by a CPU 1 is held in a buffer 100a (100b). From among packets received from Ethernet 820a (820b), a packet, a destination of which is a communication device 800, is held in the buffer 100a (100b). A packet which should be transmitted is transmitted from a transfer judging circuit 200 to Ethernet 820a or 820b through a MAC unit 300a or 300b. If a transfer judging circuit 200 judges a packet from the Ethernet 820a to be a packet, a destination of which is another communication device, with reference to a destination MAC address, this packet is transferred to the Ethernet 820b through MAC 300b. If a usage rate of a transferring FIFO buffer 130a (130b) exceeds a threshold value in the process of transmitting a packet held in a transmitting FIFO buffer 120a (130b) on a priority basis, the priority order of a transfer packet is made higher than that of a transmit packet so that the transfer packet is transferred to the Ethernet 820a or 820b in preference to the transmit packet.

Abstract: A shared memory packet switching device includes: a shared memory providing a shared memory space; an input logic unit associated with at least one receive port, and being operative to determine whether the associated receive port is saturated by determining whether a number of packets received via the associated receive port and currently stored in the shared memory exceeds a drop threshold value; a packet routing control unit operative to determine a destination one of the transmit ports for each of the received data packets; and an output logic unit associated with at least one of the transmit ports, the output logic unit being communicatively coupled with the packet routing control unit, and being operative to determine whether the associated transmit port is congested by determining whether a number of packets currently stored in the shared memory that are to be transmitted via the associated transit port exceeds a congestion threshold value, and also being operative to generate an associated output full

Abstract: In a method of improved media frame transmission in a communication network. Initially a plurality of “original” or regular media frames are provided for transmission. According to the invention, robust representations of the provided regular media frames are generated and stored locally. Subsequently, one or more of the regular media frames is/are transmitted. The invention detects an indication of a loss of a transmitted media frame, and the idea is to transmit, in response to a detected frame loss, a stored robust representation of the lost media frame and/or a stored robust representation of a subsequent, not yet transmitted, media frame to increase the media quality.

Abstract: A client receives streaming data from a server at a first data rate and buffers the streaming data in a buffer having a finite size. The client consumes the streaming data from the buffer at a second data rate. The second data rate is variable. Based on a data depth of the buffer relative to a reference data depth, the client generates a sequence of signals and sends the sequence of signals to the server to dynamically adjust the first data rate. The sequence of signals influences the data depth of the buffer toward the reference data depth.

Abstract: A packet switch includes a pipeline scheduler for scheduling packets according to a credit-based flow control protocol. A credit update pipeline stage initializes available credits for egress ports of the packet switch. A request pipeline stage generates packet requests for packets based on the available credits. A grant pipeline stage selects packets based on the ports requests and the available credits, and generates port grants for the selected packets. Additionally, the credit update stage updates the available credits based on the port grants. The packet switch routes the selected packets from ingress ports of the packet switch to the egress ports based on the port grants. In some embodiments, ingress ports generate enqueue requests based on the packets, an enqueue pipeline stage generates enqueue states based on the enqueue requests, and the request pipeline stage selects packets for routing based on the enqueue states and the available credits.

Abstract: Techniques that assist in processing of failure detection protocol (FDP) packets. Techniques are provided that assist a CPU of a network device in processing incoming FDP packets. In one embodiment, only a subset of FDP packets received by the network device is forwarded to the CPU for processing, the other FDP packets are dropped and not forwarded to the CPU. In this manner, the amount of processing that a CPU of the network device has to perform for incoming FDP packets is reduced. This enables the network device to support newer FDPs with shorter periodic interval requirements.

Abstract: The invention provides a method and apparatus for network congestion management. The method includes inserting a probe frame into data traffic in the network from a first endpoint in the network, reflecting the probe frame from a second endpoint in the network back to the first endpoint, receiving the reflected probe frame at the first endpoint via the network, determining a round-trip delay based on flow of the probe frame through the network, and throttling network traffic according to the determined delay to manage network traffic congestion.

Abstract: A computing environment containing a mesh network that is adapted to provide a reliable transport mechanism over which services may be delivered. Nodes of the mesh can automatically select routable addresses without conflicts, which allows nodes of the mesh to be accessed, even as the mesh changes through the addition or deletion of nodes. Also, nodes communicate with a protocol that supports service advertisements. These advertisements can identify mesh nodes that supply services, such as file or print servers, for which devices that have not yet connected to the network may be searching. Advertisements can also identify services to be used by nodes in the network, allowing, for example, a node to select a gateway providing a reliable connection to an external network. The mesh network can be used as a transport for communication using protocols, such as TCP/IP, that generally exhibit poor performance when using unreliable transports.

Abstract: A connection setting apparatus includes: a device-registered section to register first and second device information specifying a first and a second information-communication device, respectively; a physical-connection-port-information-acquiring section to acquire a physical connection port information from a first information output port of the first information communication device; a logical-connection-port-information-acquiring-section to acquire a logical connection port information from the second information output port of the second information communication device; a connection-port-display section to display a mark representing the first physical connection port specified by the physical connection port information and a mark representing the second information communication device; a selection section to select the first physical connection port and the second information communication device to be communication parties in information communication by a selection operation to the marks; and a corre

Abstract: In one embodiment, a method includes comparing an occupancy level of a buffer of a port of a router to a threshold, and controlling the port to operate at a first voltage and frequency based at least in part on the comparison, and at least one other port of the router is controlled to operate at a second voltage and frequency. Other embodiments are described and claimed.

Abstract: The transmission delay of a voice frame can be reduced by performing internal collision resolution and frame aggregation according to the presence or absence of a voice frame awaiting transmission in a MAC layer, thereby reducing an end-to-end voice transmission delay time for a VoIP service.

Abstract: A system for managing bandwidth use in a device. In a specific embodiment, the device is a network device that includes a first data scheduler that is adapted to initially share available device bandwidth among a first type of traffic and a second type of traffic on an as-needed basis. A traffic monitor communicates with the first scheduler and causes the first data scheduler to guarantee predetermined transmission characteristics for the second type of traffic. The first data scheduler includes one or more routines for prioritizing first type of traffic above the second type of traffic when the network device is in a first operational mode, and prioritizing the second type of traffic above the first type of traffic when the network device is in a second operation al mode. The minimum transmission characteristics include a minimum service rate and a minimum latency for the second type of traffic.

Abstract: A method for reducing power consumption in a mobile station is provided. The method includes holding data associated with a first polling message until a first scheduled RF wake-up of the mobile station. The method also includes sending the data associated with the first polling message at substantially the same time as the first scheduled RF wake-up of the mobile station. The method further includes holding a second polling message until a second scheduled RF wake-up of the mobile station. The scheduled RF wake-ups of the mobile station are determined according to a schedule established by a service provider of the wireless communication network. In certain embodiments, the polling messages are associated with an ActiveSync® Ping command.

Abstract: There are provided a CPU connection circuit and a method wherein the CPU connection circuit is a circuit to be employed by two CPUs by alternately conducting a changeover between two buffers disposed therebetween to prevent an event that data processing cannot be fully executed by the CPU on the receiving side. There is included a memory controller 303 which monitors whether or not an amount of data stored by a CCPU 1 in either one of buffers 301 and 302 reaches a predetermined threshold value; when the amount of data stored by the CCPU 1 in the buffer 301, 302 reaches the threshold value, the memory controller 303 requests an ACPU 2 to acquire the data stored in the buffer and changes the storage destination of data from the CCPU to the other one of the buffers; the threshold value is a value more than a unit quantity of data which the CCPU 1 sends to the buffer 301, 302.

Abstract: An embodiment of the invention involves a method for selecting and maintaining wireless communications for wireless mesh networks between tier-2 and tier-3 nodes. The method comprises a first operation of receiving channel information from a first wireless node operating as an access point. The channel information includes each channel number used by one or more wireless nodes that are detected by the first wireless node to be operating within a signal coverage area of the first wireless node. Then, an active scan is conducted on a frequency spectrum for wireless signals based on the channel information. Such scanning is accomplished by initially scanning frequencies of each channel number used by the one or more wireless nodes. After the scanning, a determination is made whether to (i) maintain wireless communications with the first wireless node or (ii) establish new wireless communications with a new wireless node (AP). Other embodiments are described and claimed.

Abstract: A method and network node for distributing media across an IP network, the network node having multicast and peer-to-peer ingress interfaces, and unicast and peer-to-peer egress interfaces. For each of a plurality of contents, the network node selects an ingress interface and receives media packets representing the content at the selected interface. For each media packet received, the network node determines from information conveyed in the media packet whether that media packet can be cached at the network node and caches the media packet if permitted. When the network node receives a request from a further network node for a given content, the network node selects an egress interface and forwards media packets representing the given content to the further network node over that selected interface. If media packets representing the given content were cached, the cached media packets are forwarded to the further network node.

Abstract: A router may be tested using a packet-based testing technique in which the test packets are generated by the router. In one implementation, a forwarding plane in a router may include a first component to process header information of packets to determine forwarding information, and a memory component to store payload data for the packets. A control plane of the router may generate test packets, insert the test packets into the forwarding plane, receive a second set of packets from the forwarding plane, analyze the second set of packets to determine whether the second set of packets correspond to the inserted plurality of test packets, and output, based on the analysis, test results, relating to the operation of the routing device.

Abstract: The present invention relates to a node and a method for allocation of transmission rates to a mobile terminal for sending packet data over an air interface in a radio telecommunications network by sending messages to the mobile terminal using at least a first radio channel. The invention comprises: providing at least a first grant queue for queuing messages to mobile terminals, the messages indicating allocated transmission rates to be sent over the at least first radio channel; receiving a rate request from a mobile terminal; allocating a transmission rate to the mobile terminal in dependence of the current air interface interference and the assigned transmission rates in the grant queue; and queuing a message, to be sent to the mobile terminal, indicating the allocated transmission rate in the grant queue.

Abstract: Adaptive De-Jitter Buffer for Voice over IP (VoIP) for packet switch communications. The de-jitter buffer methods and apparatus presented avoid playback of underflows while balancing end-to-end delay. In one example, the de-jitter buffer is recalculated at the beginning of each talkspurt. In another example, talkspurt packets are compressed upon receipt of all remaining packets.

Abstract: Embodiments of the present invention provide a method comprising determining, by a processor, at least an indication of a size of a de-interleaved symbol from a frame within a frame buffer, estimating, by the processor, a target transport channel buffer in a channel buffer for the de-interleaved symbol, and writing, by the processor, the de-interleaved symbol in the channel buffer based on the estimated target transport channel buffer.

Abstract: A plurality of frame buffers of a communication device store input frames for respective flows, which are units of managing communication. A sequential scheduler and an adjustment scheduler cyclically visit the plurality of frame buffers to read a frame for external output from each frame buffer. The sequential scheduler reads one frame per a visit to each frame buffer at a speed lower than a communication speed of the communication device. The adjustment scheduler reads one or more frames per a visit to each frame buffer such that a restriction on read quantity defined by a reference value greater than the shortest frame size is imposed.

Abstract: A machine implemented method and system for communication between a computing system and an adapter is provided. An application from among a plurality of applications sends a message to the adapter with a value V. The adapter queues the message at the first storage location and writes the value V at a second storage location after the message is successfully queued at the first storage location. To determine if the message was successfully queued, the computing system reads the written value at the second storage location and compares it to the value V that was sent with the message.

Abstract: A task obtained by a communications processor is decomposed into one or more requests that form a request group. The requests of the request group are sent to main memory and responses to those requests are expected. There may be requests for a plurality of request groups being processed concurrently. However, responses to the request groups are to be returned to the communications processor in the order in which the request groups were sent from the communications processor. To ensure this ordering, dependencies between the request groups are tracked by hardware coupled to the communications processor.