31 results about "Packet scheduling algorithm" patented technology

Packet communication networks for transmission to wireless subscriber devices utilize both wireline and wireless packet routing components. The routing elements of these two different types often implement different packet scheduling algorithms, typically a form of Weighted Fair Queuing (WFQ) in the wireline portion of the network and Proportional Fairness (PF) queuing in the wireless domain. To improve resource allocation and thus end to end quality of service for time sensitive communications, such as integrated multimedia services, the present disclosure suggests adding the notion of slack time into either one or both of the packet scheduling algorithms. By modifying one or more of these algorithms, e.g. to reorder or shuffle packets based on slack times, global optimal resource allocations are possible, at least in certain cases.

The present packet scheduling algorithm gives cellular network operators greater flexibility in adjusting the way resources are allocated among interactive best-effort data users. Best effort data users with different radio link qualities may have different amounts of data delivered to them using the same amount of radio resource. In the context of link adaptation, this characteristic complicates the fairness issue in cellular environments and has a profound impact on the overall system performance. As a result, the present packet scheduling algorithm is capable of allocating radio resource dynamically, not only based on channel conditions, but also to achieve different performance trade-offs among users with different link qualities. According to the algorithm, channel quality is determined for each user, channel efficiency is calculated and the channel efficiency value is used as the primary factor in weighting the delivery of packets to (or from) a given user. In a packet schedule weighting equation, a value of exponent may be varied from negative to positive to give good (or bad) users better service. However, performance of users with bad channel qualities degrades the performance of good channel users in a disproportionate manner. It is shown that it is frequently preferable to favor users with good channel qualities.

The invention discloses a unicast and multicast service packet scheduling algorithm for a satellite-borne CICQ structural switch under a GEO channel environment, belonging to the technical field of satellite network and high-performance packet switching. The algorithm aims at relieving the multicast HOL Blocking problem in CICQ switch packet scheduling from input / output scheduling algorithm, and compensating the packet which fails in transmission and requires retransmission due to the GEO channel problem. The method takes relieving blockage of head packet on next packet of a multicast queue as a target on the premise that a CICQ switch works in work-conserving as far as possible in input scheduling, takes the number of packet retransmission times in a channel as the principle concerning factor in unicast and multicast packet decision, and takes relieving blockage of head packet for next packet of a cross buffer corresponding multicast queue in output scheduling. The method sufficiently takes GEO satellite network characteristics and requirement into consideration, and has an excellent practical application value.

A system comprises a base station, a plurality of mobile stations, and a packet scheduling algorithm that operates on the base station. The packet scheduling algorithm groups the plurality of mobile stations into a plurality of groups based on average throughput with the base station. The packet scheduling algorithm selects one of the plurality of groups for applying a scheduling metric.

The invention discloses a CICQ structure switch packet scheduling algorithm based on virtual queue length coordination single multicast competition. The algorithm provided by the invention includes two parts: input scheduling and output scheduling. In the input scheduling, first find out the output port with the minimum sum of columns and cross-node buffer groups, and then select the destination port of the unicast header cell and the multicast header cell among the input ports that have cells going to the output port, and Set the input port with the fewest elements, calculate the weight of the unicast header cell and the multicast header cell, select the header cell with the largest weight, and transmit it to the corresponding cross-node buffer. In the output scheduling, the weight of the cross-node cache is equal to the sum of the corresponding unicast queue captain and the virtual captains of all the multicast queues whose header cells go to its corresponding output port, and the packet in the cross-node cache with the largest weight leaves the output port. Compared with the typical method, the algorithm provided by the invention has better throughput rate and packet average delay performance.

The invention belongs to the technical field of network, in particular relates to a data concurrency transmission method of a multi-network interface device. The method comprises the following steps: (1) a transmission end and a reception end detect the amount of network interfaces for communication and label the network interfaces so that each network interface for communication has a serial number; (2) the transmission end divides the data to be transmitted into a plurality of message segments and adds transmission serial number for each message segment; (3) the transmission end remainders the transmission serial number of the message segment to the amount of the network interfaces and transmits the message segment according to the obtained remainder by selecting the corresponding network interface; and (4) the reception end orders the received message segments. The invention can fully utilize the traditional network to solve the problems of unreasonable data distributing and dispatching, disordered reception, low retransmission efficiency, and the like by an improved grouping dispatching algorithm.

The invention discloses a CICQ structure crossbuffer queue balance packet scheduling algorithm. According to the CICQ structure crossbuffer queue balance packet scheduling algorithm, making a switch work in a work-conserving state as a core is realized or is approached to a maximum degree, output ports are taken as matching reference, output ports having smallest crossbuffer queue length are selected for priority matching, crossbuffer packet occupation of all output ports is balanced, so working in the work-conserving state is made to realize or approach to a maximum degree, the passing percentage is improved, and average packet time delay is further reduced. As is shown in simulation result comparison, for the CICQ switch with crossbuffer as one packet, the algorithm is better than known mainstream algorithms in the average packet time delay, and excellent practical values in large-scale high performance CICQ switches are realized.

The invention discloses a GEO satellite-borne CICQ structure switch single-multicast mixed service packet scheduling algorithm, which belongs to the technical field of satellite network and high-performance packet switching. The purpose of the present invention is to alleviate the problem of multicast HOL Blocking in the grouping scheduling of CICQ switch in the input-output scheduling algorithm, and at the same time compensate the grouping that needs to be retransmitted due to the transmission failure caused by the GEO channel problem. In the input scheduling, on the premise of making the CICQ switch work as work-conserving as possible, the goal is to alleviate the blocking of the multicast queue head group to the secondary group. In single-multicast packet arbitration, the number of times the packet is retransmitted in the channel is considered as the primary factor. In the output scheduling, the goal is to alleviate the blockage of the sub-packet by the cross-buffer corresponding to the head of the multicast queue. The invention fully considers the characteristics and requirements of the GEO satellite network, and has good practical application value.

The invention discloses a GEO satellite-borne CICQ structure exchange group scheduling algorithm considering the channel state. The algorithm of the present invention takes realizing or maximally approaching the switch to work in the work-conserving state as the core, with the goal of improving the channel utilization rate, and under the condition that the maximum matching number remains unchanged, the output port with a normal channel state can obtain the largest matching number , to reduce the average delay of packets while increasing the pass rate. The comparison of the simulation results shows that for the GEO satellite on-board CICQ switch with a cross-buffer capacity of one packet, under various common satellite channel states, the algorithm of the present invention is superior to the existing mainstream in terms of pass rate and average packet delay The algorithm has good practical value in GEO high-performance spaceborne CICQ switches.

The invention provides the technical scheme of an on-line certificate state query method for a dynamic packet scheduling algorithm on the basis of a limited priority level. The scheme is based on the limited priority level dynamic packet scheduling algorithm; when a CA (certificate authority) center issues an updated-state digital certificate, the algorithm is utilized to dynamically group and store the updated digital certificate into a CA database, and a characteristic identifier is correspondingly added; and when the digital certificate state is queried, the characteristic identifier is matched to look up the priority level group corresponding to the digital certificate to obtain the real-time state of the digital certificate. According to the method, the query time of the data base is reduced, the query time is shortened, the query efficiency is improved, and the on-line certificate state query method has a good applicability for the financial stock exchange and the bank in which the effectiveness of the digital certificate is frequently verified and the state of the digital certificate changes frequently.

The embodiment of the invention discloses a packet scheduling algorithm and a device thereof, belonging to communication field. The packet scheduling algorithm includes the following steps: partitioning the service packets into real-time service packets and non-real-time service packets according to headers of the service packets; judging whether legal real-time service packets exist in the real-time service packets according to shaping virtual clock algorithm, if yes, sending out the legal real-time service packets, and if no, sending out the non-real-time service packets; if new legal real-time service packets appear during the process of sending out the non-real-time service packets, suspending the sending of the non-real-time service packets and sending out the new legal real-time service packets. The device comprises a sort module, a judge module, a real-time service sending module, a non-real-time sending module and a notification module. The packet scheduling algorithm and the device can ensure that the sending of the real-time service packets is free from the influence of the non-real-time service packets and can avoid the occasion that the non-real-time service packets wait for too long. Therefore, the packet scheduling algorithm and the device have strong practicability, reliability and effectiveness.

An IP layer data packet scheduling algorithm oriented to multi-service coexistence and deterministic network requirements in the Industrial Internet of Things belongs to the technical field of real-time data transmission and multi-network integration of the Industrial Internet of Things. Current packet scheduling algorithms lack the analysis of the coexistence of continuous periodic data and discrete control data in the Industrial Internet of Things. The algorithm considers the priority and transmission delay of the data packets at the same time, and obtains the partial switching conditions of the hybrid switching system through the calculation formula of the maximum transmission delay of the data packets of each priority, and uses the hybrid switching system model to carry out the packet scheduling process Based on the analysis, the packet scheduling algorithm of IP layer is designed. The algorithm of the invention can be used to schedule data packets from different types of networks, realize multi-network integration, meet the real-time requirements of data packets in the industrial internet of things, and reduce the packet loss rate.