CN107135056B - Anycast service resource allocation method for reducing frequency spectrum fragments and time delay - Google Patents

Anycast service resource allocation method for reducing frequency spectrum fragments and time delay Download PDF

Info

Publication number
CN107135056B
CN107135056B CN201710297444.XA CN201710297444A CN107135056B CN 107135056 B CN107135056 B CN 107135056B CN 201710297444 A CN201710297444 A CN 201710297444A CN 107135056 B CN107135056 B CN 107135056B
Authority
CN
China
Prior art keywords
service
spectrum
path
frequency spectrum
fragmentation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201710297444.XA
Other languages
Chinese (zh)
Other versions
CN107135056A (en
Inventor
刘焕淋
杨玉明
杜君丹
陈浩楠
吴兰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing University of Post and Telecommunications
Original Assignee
Chongqing University of Post and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chongqing University of Post and Telecommunications filed Critical Chongqing University of Post and Telecommunications
Priority to CN201710297444.XA priority Critical patent/CN107135056B/en
Publication of CN107135056A publication Critical patent/CN107135056A/en
Application granted granted Critical
Publication of CN107135056B publication Critical patent/CN107135056B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/0001—Arrangements for dividing the transmission path
    • H04L5/0003—Two-dimensional division
    • H04L5/0005—Time-frequency
    • H04L5/0007—Time-frequency the frequencies being orthogonal, e.g. OFDM(A), DMT
    • H04L5/001—Time-frequency the frequencies being orthogonal, e.g. OFDM(A), DMT the frequencies being arranged in component carriers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/12—Avoiding congestion; Recovering from congestion
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
    • H04L47/2425—Traffic characterised by specific attributes, e.g. priority or QoS for supporting services specification, e.g. SLA
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/70—Admission control; Resource allocation
    • H04L47/72—Admission control; Resource allocation using reservation actions during connection setup
    • H04L47/726—Reserving resources in multiple paths to be used simultaneously
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0058—Allocation criteria
    • H04L5/006—Quality of the received signal, e.g. BER, SNR, water filling
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04Q—SELECTING
    • H04Q11/00—Selecting arrangements for multiplex systems
    • H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
    • H04Q11/0062—Network aspects
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04Q—SELECTING
    • H04Q11/00—Selecting arrangements for multiplex systems
    • H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
    • H04Q11/0062—Network aspects
    • H04Q2011/0086—Network resource allocation, dimensioning or optimisation

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Quality & Reliability (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The invention provides an anycast service resource allocation method for reducing frequency spectrum fragments and service transmission time delay, belonging to the technical field of optical fiber communication. The invention provides a route and spectrum allocation strategy for avoiding fragments, which comprises the following specific steps: and selecting a frequency spectrum block with the size equal to the service request bandwidth to transmit the service during frequency spectrum allocation, if no frequency spectrum block exists, pre-allocating the service to all available frequency spectrum blocks on the path, calculating the corresponding frequency spectrum fragment degree, and selecting the frequency spectrum block with the lowest path fragment degree to transmit the service. The invention also provides a specific method of the service segmentation strategy for reducing the time delay, which comprises the following steps: and (3) the service with the failed spectrum allocation is transmitted in a single-path service division mode through fragmentation avoidance, and if the service fails, the service is transmitted in a multi-path service division mode with the minimum time delay difference. The invention can obviously reduce the transmission time delay and the frequency spectrum fragments of the anycast service in the elastic optical interconnection data center network, reduce the network bandwidth blocking rate and improve the network frequency spectrum utilization rate.

Description

Anycast service resource allocation method for reducing frequency spectrum fragments and time delay
Technical Field
The invention belongs to the technical field of optical fiber communication, and relates to an anycast service resource allocation method for reducing frequency spectrum fragments and time delay.
Background
In recent years, in a data center internet, with the popularization of new low-latency Network applications such as cloud computing, mobile applications, Network videos, and Content Delivery Networks (CDNs), the bandwidth demand of a Network has geometrically increased. The research reports by Cisco indicate that: over the past five years, global IP (Internet Protocol) requests have grown more than fourfold, and in the future. To date there are over 500000 data centers worldwide and are growing at 32% per year. By 2018, global IP requests will exceed 1.4ZB (1ZB 1024B, ZB: Zettabyte, Zettabyte); and, services such as network video and CDNs sensitive to time delay account for nearly 85% of them. When a user sends a request, the data center network selects a proper data center to provide service according to the network resource condition, and can return the processed delay-sensitive data to the user with low delay, and if the delay is too large, the network performance and the user experience are reduced. In order to ensure the transmission performance of such services and improve the utilization rate of resources, the anycast service has only one source node, and can select the most appropriate service model for providing services for the source node from a plurality of candidate destination nodes, thereby attracting wide attention of people. The anycast service is a very effective service transmission mode in the data center network, one service can be provided by a plurality of data centers distributed in the network, the transmission of the service among the data centers is well supported, and the development of network applications such as network videos, CDNs and the like is promoted.
With the continuous enlargement of the network scale and the continuous increase of the flow of the data center, a serious challenge is caused to the limited bandwidth resource, so that the transmission performance requirement of the big data service of the data center is met, and the problem which needs to be solved urgently is formed. The conventional optical network is constructed based on a DWDM (Dense Wavelength Division Multiplexing) technology, and the bandwidth allocation granularity of the conventional optical network is coarse, usually 50GHz or 100GHz, which causes a large amount of bandwidth resource waste, and the conventional DWDM network is not suitable for the development of a data center network. Jinno et al, m.jinno. of NTT corporation, proposed in 2009 a dynamic optical network-elastic optical network based on OFDM (Orthogonal Frequency Division Multiplexing) technology, with more flexible bandwidth allocation. The transmission capacity of the elastic optical network can reach Tb/s, the bandwidth allocation granularity is 12.5GHz, a proper modulation level can be automatically selected according to the transmission distance and the transmission rate of the service, and a proper frequency slot resource can be flexibly allocated. The elastic optical network has the advantages of ultra-large capacity, high bandwidth, high energy efficiency, flexibility and the like, becomes a powerful physical layer technology to effectively support interconnection among data center networks, and can well meet the transmission of anycast services among data centers. However, due to the limitation of the continuity and consistency of the spectrum of the elastic optical network, anycast services dynamically arrive and leave, and the spectrum resources of the path are fragmented in the process of continuously establishing and removing the optical path. The spectrum fragments refer to small, isolated and discontinuous spectrum blocks, and the spectrum fragments cannot be utilized by subsequent services, so that the spectrum utilization rate of the network is reduced, and the bandwidth blocking rate of the network is increased. How to effectively reduce the spectrum fragments of the network and reduce the bandwidth blocking rate of the network becomes an urgent problem to be solved in the elastic optical interconnection data center network.
In view of the problem of spectrum fragmentation of the elastic optical network, researchers have made many studies, mainly to arrange the generated fragments in the network and periodically reroute and allocate the spectrum to the traffic being transmitted, so as to connect smaller, isolated and discontinuous spectrum fragments into a larger spectrum block, thereby reducing the fragmentation degree and bandwidth blocking rate of the network. Since the defragmentation periodically re-routes and allocates spectrum to the traffic being transmitted, the traffic being transmitted in the network is interrupted, and there is no exact basis for the defragmentation period and how much traffic is defragmented each time, which can only be determined empirically.
Another method for reducing spectrum fragmentation is service segmentation, that is, when the network load is heavy and there is no available spectrum resource in the network, the service is segmented into several smaller sub-services for independent transmission, thereby improving the probability of successful transmission of the service and reducing the bandwidth blocking rate of the network. However, when there are not enough resources on a single path, each sub-traffic will be transmitted through multiple paths. However, due to different distances from the service source node to the data center, the sub-services arrive at the data center and are processed at different times, which may generate a serious delay difference, which may reduce the performance of the new network applications with low delay, such as network videos, CDNs, and the like.
Therefore, a new routing and spectrum allocation strategy is designed, spectrum fragments and service delay of the network are reduced, and bandwidth blocking rate of the network is reduced, so that the method has very important significance for development of the elastic optical interconnection data center network and the whole information and communication industry.
Disclosure of Invention
In view of this, the present invention provides an anycast service resource allocation method for reducing spectrum fragmentation and time delay, which is used to reduce spectrum fragmentation and service time delay, reduce bandwidth blocking rate, and improve spectrum utilization rate.
In order to achieve the purpose, the invention provides the following technical scheme:
an anycast service resource allocation method for reducing spectrum fragmentation and time delay comprises the following steps:
s1, judging the nature of the new event; if the new event is that a new service R arrives, the next step is carried out, otherwise, if the new event is that the service R ' is out of service, the split services which have the same source node with the R ' and the resource request is smaller than the resource released by the R ' are merged and transmitted;
s2, executing a route spectrum allocation strategy of fragmentation avoidance; if the distribution is successful, the step S1 is carried out, otherwise, the next step is carried out;
s3, executing the single-path service segmentation strategy based on fragment avoidance; if the distribution is successful, storing the segmentation services into a container VS, and performing a step S1, otherwise, performing the next step;
s4, executing a multi-path service division strategy based on the time delay limit; if the distribution is successful, storing the segmentation service into a container VS, and performing step S1, otherwise, blocking the service, and performing the next step;
s5, judging whether the service is transmitted; if the transmission is not completed, the step S1 is performed, otherwise, the algorithm ends.
Further, the fragmentation avoidance routing spectrum allocation strategy comprises the following steps:
s201: determining the frequency spectrum resource required by the service according to the path length and the size of the service bandwidth request;
s202: checking whether a spectrum block with the size exactly equal to the spectrum resource required by the service exists in the path residual spectrum resource; if so, allocating the service to the frequency spectrum block; if not, the service is pre-distributed to all available spectrum blocks, the corresponding path spectrum fragmentation degree is calculated, and the service is distributed to the spectrum block with the lowest path fragmentation degree.
Further, the single-path service segmentation strategy based on fragmentation avoidance comprises the following steps:
s301: selecting a single path with the most residual resources, sequencing all available spectrum blocks from large to small, dividing services according to the sizes of the first n-1 larger spectrum blocks, and distributing the first n-1 sub-services to corresponding spectrum blocks;
s302: and pre-distributing the remaining nth sub-service to all the remaining available spectrum blocks, calculating the corresponding path spectrum fragmentation degree, and distributing the nth sub-service to the spectrum block with the lowest path fragmentation degree.
Further, the delay-constraint-based multi-path traffic splitting policy includes the following steps:
s401: arranging, combining and grouping all available paths according to the service division number n, then calculating the path difference of the maximum length and the minimum length in each group of paths, and selecting the multipath with the minimum path difference to carry out spectrum allocation through combination;
s402: in the process of spectrum allocation, n multipaths are firstly sequenced from large to small according to the maximum residual spectrum block, then services are divided according to the maximum spectrum block size in the first n-1 paths, sub-services are allocated to the corresponding maximum spectrum block, for the nth sub-service, the nth sub-service is pre-allocated to all available spectrum blocks on the nth path, the corresponding path spectrum fragmentation degree is calculated, and the nth sub-service is allocated to the spectrum block which enables the path fragmentation degree to be the lowest.
The invention has the beneficial effects that:
the invention can obviously reduce the time delay and the spectrum fragment of the anycast service in the elastic optical interconnection data center network, reduce the bandwidth blocking rate of the network and improve the spectrum utilization rate of the network. The method has very important significance for the development of the elastic optical interconnection data center network and the whole information and communication industry.
Drawings
In order to make the object, technical scheme and beneficial effect of the invention more clear, the invention provides the following drawings for explanation:
FIG. 1 is a diagram of a flexible optical interconnect data center network;
fig. 2 is a schematic diagram illustrating changes in the level of spectrum fragmentation of a path after different spectrum blocks are allocated to a service;
FIG. 3 is a schematic view of traffic segmentation;
FIG. 4 is a graph of multipath division delay differences;
FIG. 5 is a schematic diagram of a single-path traffic segmentation strategy based on fragmentation avoidance;
FIG. 6 is a schematic diagram of traffic splitting multi-path transmission;
FIG. 7 is a general algorithm flow chart.
Detailed Description
Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
Fig. 1 shows an elastic optical interconnection data center network, which is an application scenario of the present invention. The physical topology of the elastic optical interconnection data center network is G (V, E), wherein V represents a node set, each node in the network is provided with a BV-WXC (Bandwidth-Variable Wavelength-Cross-Connect) which can enable signals to be added and dropped and can exchange and forward the signals, E represents a link set, and the maximum Bandwidth capacity of any one link is represented by B. There are several fixed data centers in the network for providing service, each data center has local connection with the exchange node in the network, the exchange node connected to the data center is VDC。
The anycast service request is expressed as R (s, b, c), the source node is s e.g. V \ VDCB is a bandwidth request, the unit is a frequency slot, that is, FS, c is a computing resource request, the unit is the number of servers, and the relationship between the bandwidth request and the computing resource request is c ═ α × b, where α is a constant. After an anycast service is generated from a source node, all available paths from the source node to each data center are calculated firstly, and then according to the residual calculation resources and paths in the pathsAnd selecting the optimal path and the data center for the service according to the condition of the residual computing resources of the data centers connected with the path to serve the service. According to m (p), computing resources and data center computing resources can be considered jointly, and the optimal path and the residual resource situation of the data center are evaluatedk)＝BW(pk)×CkMeasured by parameters, wherein BW (p)k) For the spectrum resources available on the kth candidate path, CkAnd calculating the residual computing resources of the data center connected with the kth candidate path. m (p)k) The larger the value, the more resources are left in the current path and data center, and the better the path and data center.
The inventive fragmentation avoidance routing and spectrum allocation strategy will be described with reference to fig. 2, using parameterssum(pk) < B assess the extent of spectral fragmentation on the path, where MaxBlock (p)k) Is a path pkThe upper maximum available spectrum block, B is the total number of frequency slots on the path, sum (p)k) The number of occupied frequency slots on a path,representing the percentage of the maximum spectrum blocks that are free on the path to the remaining available spectrum resources of the path. A larger EF value indicates a greater degree of spectral fragmentation of the path. In spectrum allocation, when a service is allocated to different spectrum blocks, the spectrum fragmentation degrees of paths are different, as shown in fig. 2, which is a schematic diagram illustrating the change of the spectrum fragmentation degrees of paths after the service is allocated to different spectrum blocks, and if a service request requires spectrum resources of 2FSs, a path has 3 available spectrum blocks. As shown in fig. 2(a), after the service is allocated to the spectrum block 1, the spectrum fragmentation degree of the path at this time is calculated to be EF equal to 0.625; as shown in fig. 2(b), after the traffic is allocated to the spectrum block 2, the spectrum fragmentation degree of the path is reduced to EF of 0.375, and as compared with fig. 2(a), the traffic is allocated to the spectrum block 2, so that the spectrum fragmentation degree of the path can be reduced, and the remaining spectrum blocks are larger, so that the subsequent traffic can be better served; if the service is distributed, as shown in FIG. 2(c)After the spectrum block 3, the spectrum fragmentation degree of the path is EF 0.375, and as compared with fig. 2(b), although the service is allocated to the spectrum block 3, the spectrum fragmentation degree of the path is not further reduced, but no fragmentation occurs, and the spectrum block 2 in fig. 2(b) becomes a spectrum fragmentation and cannot be used by subsequent services. Comparing the three spectrum allocation methods, it is clear that the method shown in fig. 2(c) not only can reduce the spectrum fragmentation degree of the path, but also can avoid the generation of new fragments.
Therefore, in the spectrum allocation stage, firstly, a spectrum block with the size exactly equal to the service request bandwidth is selected, and after the service is allocated to the spectrum block, not only the spectrum fragment degree of a path is reduced, but also new spectrum fragments are not generated; if no spectrum block with the size exactly equal to the bandwidth requested by the service exists on the path, the service is allocated to the spectrum block with the lowest fragmentation degree of the path, so that the fragmentation degree of the path can be effectively reduced.
The specific flow of the routing and spectrum allocation strategy for debris avoidance can be divided into the following steps:
s101, calculating K candidate paths from a source node to each data center by using Dijkstra algorithm, namely pk(K ═ 1, 2.., K), and a parameter value m (p) of each candidate path is calculatedk)；
S102, pressing K candidate paths into m (p)k) Sorting the values from large to small;
s103, selecting m (p)k) Candidate path p with the largest valuekLet k equal to 1, check the optimal path pkThe size of all spectral blocks above, and check pkConnected data center residual computing resource Ck；
S104 if pkThere is a single available spectrum block, and pkIf enough computing resources remain in the connected data center, jumping to the step S106, otherwise, jumping to the step S105 by k + +;
s105, jumping to the step S103 if K is less than or equal to K, otherwise jumping to the step S3;
s106 if pkIf the frequency spectrum block with the size being exactly equal to the service request bandwidth exists, the service is distributed into the frequency spectrum block, and the data center is distributed for calculationResource, and jumping to step S1, otherwise jumping to step S107;
s107, service is pre-distributed to p respectivelykIn all available spectrum blocks, and calculating corresponding path spectrum fragmentation degree
S108, the step ofSorting according to the sequence from small to large;
s109, distributing the service to the clientIn the smallest spectrum block, allocating computing resources in the data center, and jumping to step S1;
as new services continue to arrive and network load continues to increase, it is difficult to find a single available spectrum block that satisfies service transmission in the network, and another method of reducing spectrum fragmentation, service segmentation, has emerged. Traffic splitting refers to splitting traffic into several smaller sub-services when there is no single available spectrum block in the network that satisfies the traffic transmission. The method comprises the steps of dividing the service, changing the unavailable frequency spectrum block of the original service into available resources, and selecting a proper path and a proper frequency spectrum block for each sub-service respectively, so that the service is transmitted successfully, the bandwidth blocking rate of the network is reduced, and the frequency spectrum utilization rate of the network is improved. After the services are divided, each sub-service can be transmitted through different spectrum blocks on a single path, and when available resources on the single path are insufficient, each sub-service can be transmitted through different spectrum blocks on multiple paths. As shown in fig. 3, which is a schematic diagram of traffic division, as can be seen from fig. 3(a), for a traffic request 1, there is no single spectrum block available on a path 1, and the traffic is blocked, at this time, a traffic division strategy is adopted to divide the traffic into two smaller spectrum blocks for transmission in the spectrum block 1 and the spectrum block 2 on the path 1, respectively. As can be seen from fig. 3(b), for the service request 2, there is no single spectrum block available on both the path 2 and the path 3, and at this time, the service is divided and transmitted on different spectrum blocks on the two paths, so that the probability of successful transmission of the service is increased.
However, when the sub-traffic reaches different data centers through multi-path transmission after traffic segmentation, the paths from the source node to the different data centers are different, and thus a large delay difference is generated. As shown in fig. 4, which is a diagram of multipath division delay difference, when an anycast service cannot be processed by a single data center and is divided into 2 sub-services to be processed by two data centers, there are 3 available paths in the network. As can be seen from fig. 4, if two sub-services are transmitted to the data center 1 and the data center 2 for processing through the path 1 and the path 2, the path difference between the two sub-services is 900 km; if the two sub services are transmitted to the data center 2 and the data center 3 for processing through the path 1 and the path 3, the path difference between the two sub services is 0 km; if two sub-services are transmitted to the data center 2 and the data center 3 for processing through the path 2 and the path 3, the path difference between the two sub-services is 900 km. Therefore, when the anycast service is processed by a plurality of data centers through multi-path transmission after being divided, each sub-service selects different paths, the path difference is large, and the time delay is different.
Fig. 5 is a schematic diagram of the fragmentation avoidance based single-path service segmentation strategy, and it is known from fig. 5 that the required bandwidth of a service is 7FSs, the computation resource of a data center is 15, the computation resource is sufficient, there is no single spectrum block capable of transmitting the service on the path, but the idle spectrum resource on the path is greater than the service request bandwidth, so that the service is transmitted in a single-path segmentation manner. Since the spectrum resources of the spectrum block 2, the spectrum block 3 and the spectrum block 4 are greater than the service request bandwidth, the service is divided into 3 sub-service transmissions. As shown in fig. 5(a) and 5(b), after the sub-service 1 and the sub-service 2 are transmitted through a larger spectrum block, the spectrum block 1 and the spectrum block 4 remain in the network, and the remaining service request bandwidth is 2 FSs. Spectrum fragmentation occurs if the sub-service 3 is allocated to the spectrum block 4, so the sub-service 3 is allocated to the spectrum block 1 with the lowest path spectrum fragmentation, and then calculation resources are allocated to the service, and the path resources and the data center calculation after service division are as shown in fig. 5 (d).
The specific flow of the single-path service segmentation strategy based on fragment avoidance can be divided into the following steps:
s201: let k equal to 1, select m (p)k) Candidate path p with the largest valuekSorting the available spectrum blocks on the path from large to small;
s202: judging whether the resources are enough, if the sum of the available frequency spectrum blocks on the path is larger than the service request bandwidth, and the residual computing resources of the data center connected with the path are larger than the service request computing resources, determining the number n of service partitions, and jumping to the step S204, otherwise, k + +, and jumping to the step S203;
s203: checking the size of K, jumping to step S202 if K is less than or equal to K, otherwise jumping to step S4;
s204: allocating the first n-1 sub-services to the first n-1 larger frequency spectrum blocks on the path, pre-allocating the nth sub-service to the residual frequency spectrum blocks on the path, and calculating the corresponding path frequency spectrum fragment degree valueAllocating the nth sub-service toWithin the smallest spectral block;
s205: allocating computing resources for traffic, i.e. Ck-c, saving the split traffic into the set VS and jumping to step S1.
The delay constraint based multi-path traffic splitting strategy of the present invention will be described in more detail with reference to fig. 6. In the stage of determining the division number of the frequency spectrum, in order to reduce the division number of the services and enable each path to transmit more services as much as possible, the invention adopts a heuristic algorithm to firstly divide the services into 2 sub-services for multi-path transmission, if the transmission of the 2 sub-services fails and the degree of network fragmentation is higher, the services are divided into 3 sub-services for transmission, and so on. In the elastic optical interconnection data center network, because the number of the data centers is limited, the service has the maximum division number which is set as nmaxThen n ismaxThe smaller of the number of data centers and the number of shares the traffic is partitioned at the minimum partition granularity. After the service division number is determined, in the path selection stage, in order to select the multipath combination with the minimum delay difference, firstly, K candidate paths are arranged, combined and grouped according to the division number n, and the total number isGroups of n paths each, denoted asLength is expressed as Lj＝{l1,l2,...,lnIn the difference Δ L between the maximum path length and the minimum path length in each set of pathsjIndicating the multipath delay. Δ LjThe larger the path difference and the more severe the delay, the larger Δ L is selectedjThe multipath transmission with the minimum value can effectively improve the time delay sensitive service performance. As shown in fig. 6, which is a schematic diagram of traffic division and multipath transmission, it is known from fig. 6 that neither spectrum resources nor computation resources on three paths can satisfy traffic division and single-path transmission, and therefore, traffic is processed by multiple data centers through multipath transmission. Calculating the parameter values m (p) of the three pathsk) Are respectively m1(p1)＝135、m2(p2) 20 and m3(p3) 15 due to m1(p1)＞m2(p2)＞m3(p3) Then the sub-traffic sizes on path 1 and path 2 are determined by the smaller of the maximum spectrum blocks on the path and the data center computing resources. As shown in fig. 3(a), the maximum spectrum block size on path 1 is 4, and the computation resource size is 12, so the spectrum resource request of sub-service 1 transmitted on path 1 is 4FSs, and the computation resource request is 3. Similarly, as shown in fig. 3(b), although the maximum spectrum block size of path 2 is 5, the remaining computing resource of the data center is 2, and only the service bandwidth request of 2FSs can be processed, so that the spectrum resource request of sub-service 2 is 3FSs, and the computing resource request is 2. After the size of the current n-1 sub-services is determined, if the size of the nth sub-service is the remaining resource request, the 3 rd sub-serviceThe spectrum resource request is 2FSs and the computing resource request is 1. Since the 3 rd path has enough resources to transmit the sub-service 3, the size of the sub-service is determined. In order to reduce the generation of spectral fragments, each sub-service will be transmitted through a spectral block that minimizes the spectral fragments of the path, and the result of the traffic division multi-path transmission spectrum allocation is shown in fig. 6.
The specific general calculation process of the multi-path service partitioning strategy based on the time delay limitation can be divided into the following steps:
s301, setting the service segmentation number n to 2;
s302, the K candidate paths are arranged, combined and grouped according to the service division number n, and the service division number n is totalGroups, forming sets of pathsCalculating the path difference value Delta L of each group of path setsjSet the paths as Δ LjSorting the values, and making j equal to 1;
s303, selecting delta LjThe path set with the minimum value calculates the parameter value m (p) of each pathk) And each path is arranged according to m (p)k) Sorting from big to small;
s304, determining the size of the first n-1 sub-services transmitted on the first n-1 paths with larger parameter values;
s305, checking the nth sub-service resource request and the nth path residual resource condition, if the nth path residual resource is enough, jumping to the step S308, otherwise jumping to the step S306;
s306, checking the size of j, ifJumping to step S303, otherwise jumping to step S307;
s307, checking the size of n, if n is less than or equal to nmaxN + +, and jumps to step S302, otherwise, the service is blocked and jumps to step S5;
s308, distributing each sub-service to a frequency spectrum block which enables the path fragment degree to be minimum, distributing corresponding computing resources to each data center, storing the segmentation services to a set VS, and jumping to the step S1;
the method for allocating anycast service resources to reduce spectrum fragmentation and delay according to the present invention will be described in detail with reference to fig. 7, and the specific process may be divided into the following steps:
s1, judging the nature of the new event, if the new service R arrives, jumping to step S2, otherwise, if the new event is the service R ' is finished, the new event leaves, merging and transmitting the split services which have the same source node with R ' and the resource request is less than the resource released by R ';
s2, executing the route and frequency spectrum allocation strategy of the fragmentation avoidance, if the allocation is successful, jumping to the step S1, otherwise jumping to the step S3;
s3, executing the single path service segmentation strategy based on fragment avoidance, if the distribution is successful, storing the segmented service into a container VS, and jumping to the step S1, otherwise jumping to the step S4;
s4, executing the multi-path service dividing strategy based on the time delay limitation, if the distribution is successful, storing the divided service into a container VS, and jumping to the step S1, otherwise, jumping to the step S5 if the service is blocked;
and S5, judging whether the transmission of the service is finished, if not, jumping to the step S1, otherwise, finishing the algorithm.
Finally, it is noted that the above-mentioned preferred embodiments illustrate rather than limit the invention, and that, although the invention has been described in detail with reference to the above-mentioned preferred embodiments, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the invention as defined by the appended claims.

Claims (4)

1. A method for allocating anycast service resources to reduce spectrum fragmentation and time delay is characterized in that: the method comprises the following steps:
s1, judging the nature of the new event; if the new event is that a new service R arrives, the next step is carried out, otherwise, if the new event is that the service R ' is out of service, the split services which have the same source node with the R ' and the resource request is smaller than the resource released by the R ' are merged and transmitted;
s2, executing a route spectrum allocation strategy of fragmentation avoidance; if the distribution is successful, the step S1 is carried out, otherwise, the next step is carried out;
s3, executing the single-path service segmentation strategy based on fragment avoidance; if the distribution is successful, storing the segmentation services into a container VS, and performing a step S1, otherwise, performing the next step;
s4, executing a multi-path service division strategy based on the time delay limit; if the distribution is successful, storing the segmentation service into a container VS, and performing step S1, otherwise, blocking the service, and performing the next step;
s5, judging whether the service is transmitted; if the transmission is not completed, the step S1 is performed, otherwise, the algorithm ends.
2. The anycast service resource allocation method for reducing spectrum fragmentation and latency of claim 1, wherein: the route spectrum allocation strategy for debris avoidance comprises the following steps:
s201: determining the frequency spectrum resource required by the service according to the path length and the size of the service bandwidth request;
s202: checking whether a spectrum block with the size exactly equal to the spectrum resource required by the service exists in the path residual spectrum resource; if so, allocating the service to the frequency spectrum block; if not, the service is pre-distributed to all available spectrum blocks, the corresponding path spectrum fragmentation degree is calculated, and the service is distributed to the spectrum block with the lowest path fragmentation degree.
3. The anycast service resource allocation method for reducing spectrum fragmentation and latency of claim 1, wherein: the single-path service segmentation strategy based on fragment avoidance comprises the following steps:
s301: selecting a single path with the most residual resources, sequencing all available spectrum blocks from large to small, dividing services according to the sizes of the first n-1 larger spectrum blocks, and distributing the first n-1 sub-services to corresponding spectrum blocks;
s302: and pre-distributing the remaining nth sub-service to all the remaining available spectrum blocks, calculating the corresponding path spectrum fragmentation degree, and distributing the nth sub-service to the spectrum block with the lowest path fragmentation degree.
4. The anycast service resource allocation method for reducing spectrum fragmentation and latency of claim 1, wherein: the multi-path service division strategy based on the time delay limitation comprises the following steps:
s401: setting all available paths as K, arranging, combining and grouping according to the number n of the divided pathsGrouping, wherein each group comprises n paths, then calculating the path difference of the maximum length and the minimum length in each group of paths, and selecting the multipath with the minimum path difference to carry out spectrum allocation through combination;
s402: in the process of spectrum allocation, n multipaths are firstly sequenced from large to small according to the maximum residual spectrum block, then services are divided according to the maximum spectrum block size in the first n-1 paths, sub-services are allocated to the corresponding maximum spectrum block, for the nth sub-service, the nth sub-service is pre-allocated to all available spectrum blocks on the nth path, the corresponding path spectrum fragmentation degree is calculated, and the nth sub-service is allocated to the spectrum block which enables the path fragmentation degree to be the lowest.
CN201710297444.XA 2017-04-28 2017-04-28 Anycast service resource allocation method for reducing frequency spectrum fragments and time delay Active CN107135056B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710297444.XA CN107135056B (en) 2017-04-28 2017-04-28 Anycast service resource allocation method for reducing frequency spectrum fragments and time delay

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710297444.XA CN107135056B (en) 2017-04-28 2017-04-28 Anycast service resource allocation method for reducing frequency spectrum fragments and time delay

Publications (2)

Publication Number Publication Date
CN107135056A CN107135056A (en) 2017-09-05
CN107135056B true CN107135056B (en) 2019-12-27

Family

ID=59716081

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710297444.XA Active CN107135056B (en) 2017-04-28 2017-04-28 Anycast service resource allocation method for reducing frequency spectrum fragments and time delay

Country Status (1)

Country Link
CN (1) CN107135056B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108737031B (en) * 2018-05-22 2020-09-25 北京邮电大学 Data spectrum reconstruction transmission method, device, controller and storage medium
CN108834003B (en) * 2018-06-28 2021-04-30 安徽继远软件有限公司 Electric power optical carrier communication multidimensional resource distribution optimization method for quantum communication service
CN110708616B (en) * 2018-07-10 2022-04-29 中兴通讯股份有限公司 Spectrum allocation method, device and computer storage medium for optical network
CN110035337B (en) * 2019-04-16 2021-09-07 郑州轻工业学院 Active multi-stream spectrum allocation method based on exponential division frequency block in elastic optical network
CN111010626B (en) * 2019-11-11 2021-09-03 郑州轻工业学院 Method for allocating transparent advance reserved service resources in elastic optical network

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102752757A (en) * 2011-04-19 2012-10-24 深圳清华大学研究院 Method for optimizing frequency spectrum allocation according to minimal waste criterion in frequency spectrum aggregation process
CN105490934A (en) * 2016-01-25 2016-04-13 重庆邮电大学 Fair routing and spectrum assignment method based on fragment processing
CN105915279A (en) * 2016-05-18 2016-08-31 烽火通信科技股份有限公司 Method for triggering protection switching based on OTU alarm in WSON

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102752757A (en) * 2011-04-19 2012-10-24 深圳清华大学研究院 Method for optimizing frequency spectrum allocation according to minimal waste criterion in frequency spectrum aggregation process
CN105490934A (en) * 2016-01-25 2016-04-13 重庆邮电大学 Fair routing and spectrum assignment method based on fragment processing
CN105915279A (en) * 2016-05-18 2016-08-31 烽火通信科技股份有限公司 Method for triggering protection switching based on OTU alarm in WSON

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
Spectrum allocation based on spectrum;Huanlin Liu1 ✉, Bangtao Zhou1, Yong Chen2;《IET Journals》;20160302;全文 *

Also Published As

Publication number Publication date
CN107135056A (en) 2017-09-05

Similar Documents

Publication Publication Date Title
CN107135056B (en) Anycast service resource allocation method for reducing frequency spectrum fragments and time delay
Abkenar et al. Study and analysis of routing and spectrum allocation (RSA) and routing, modulation and spectrum allocation (RMSA) algorithms in elastic optical networks (EONs)
CN111901424B (en) Cloud edge cooperative network resource smooth migration and reconstruction method and system
Aibin et al. Adaptive modulation and regenerator-aware dynamic routing algorithm in elastic optical networks
CN105141517B (en) A kind of flexible optical-fiber network Anycast business energy-saving routing method based on resource perception
CN108834004B (en) Routing calculation, fiber core selection and frequency spectrum allocation method and system based on cross crosstalk perception
CN107426110B (en) Adaptive load balancing energy consumption optimization method and system in spectrum flexible optical network
JP6852824B2 (en) Optical network system, optical node device, and optical network control method
CN113784373B (en) Combined optimization method and system for time delay and frequency spectrum occupation in cloud edge cooperative network
CN101986619A (en) Bandwidth reservation-based VSAT satellite communication system bandwidth distribution method
WO2008115221A2 (en) Hierarchically clustered p2p streaming system
CN113742046A (en) Flow grooming cloud-side computing network computing resource balanced scheduling method and system
CN111162865A (en) Virtual optical network mapping method for sensing fragments in space division multiplexing elastic optical network
CN101127689A (en) Method for minimizing priority seizing cost
CN114268548A (en) Network slice resource arranging and mapping method based on 5G
CN115460088A (en) 5G power multi-service slice resource allocation and isolation method
CN110035337B (en) Active multi-stream spectrum allocation method based on exponential division frequency block in elastic optical network
CN108184175B (en) MC node limitation-based elastic optical network multicast routing and spectrum allocation method
CN108616782B (en) Frequency spectrum allocation method based on disjoint connection group optimization in elastic optical network
CN113015040A (en) Resource allocation method based on fragment and domain matching degree in multi-core elastic optical network
WO2012126435A2 (en) Method and apparatus for allocating bandwidth resources
CN112291010A (en) Multi-domain optical network traffic grooming method based on matching game
Alyatama Dynamic spectrum allocation in orthogonal frequency-division multiplexing optical networks
CN102420656A (en) Trigger method of frequency spectrum resource reconstruction in all-optical network and system thereof
Mahala et al. Weight distributed spectrum allocation in flexible-grid optical networks

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant