阅读说明：本技术 空分复用弹性光网络中基于精确划分的频谱预分配方法 (Frequency spectrum pre-allocation method based on accurate division in space division multiplexing elastic optical network ) 是由 袁俊岭 李健勇 孟颍辉 张启坤 郭梦飞 南思雨 于 2021-08-17 设计创作，主要内容包括：本发明提出了一种空分复用弹性光网络中基于精确划分的频谱预分配方法,根据不同类型业务的比例和需要的频块尺寸,计算每种类型业务需要的精确的总频片数量；根据需要的精确总频片数量,在前C-1个核芯上为不同类型业务预分配频片,进而在第C个核芯上为不同类型业务预分配频片；为每种类型业务预分配的频片呈现：预分配的频片个数接近业务需要的精确总频片个数；为前I-1种类型业务预分配的频片刚好可以划分为整数个频块；为每种类型业务预分配的频片相对均衡地分布在不同的核芯上。本发明通过将链路上所有核芯上的频片预分配给不同类型的业务,使得预分配的频片个数与业务的需求相匹配、被浪费的频片个数尽量少、不同类型业务受到的串扰尽量均衡。(The invention provides a frequency spectrum pre-allocation method based on accurate division in a space division multiplexing elastic optical network, which calculates the accurate total number of frequency slices required by each type of service according to the proportion of different types of services and the required frequency block size; pre-allocating frequency slices for different types of services on the first C-1 cores according to the required accurate total number of frequency slices, and further pre-allocating frequency slices for different types of services on the C-th core; pre-allocated frequency slice presentation for each type of service: the number of pre-allocated frequency slices is close to the accurate total number of frequency slices required by the service; the frequency slice pre-allocated for the first I-1 type service can be just divided into an integer number of frequency blocks; the frequency slices pre-allocated for each type of traffic are distributed relatively evenly over the different cores. The invention pre-allocates the frequency slices on all the cores on the link to the services of different types, so that the pre-allocated frequency slices are matched with the requirements of the services, the number of the wasted frequency slices is as small as possible, and the crosstalk suffered by the services of different types is balanced as possible.)

1. A frequency spectrum pre-allocation method based on accurate division in a space division multiplexing elastic optical network is characterized by comprising the following steps:

the method comprises the following steps: the service types are sorted from more to less according to the number of frequency slices required by each connection request, and the number of the frequency slices required by each connection request of the various types of services after reordering is recorded as { N₁,N₂,…,N_IIs N₁＞N₂＞…＞N_ICalculating the accurate total frequency chip number required by the ith service type

Step two: the number of frequency slices which are recorded on the c core and pre-allocated for the i type of service is F_c,iCalculating the number of frequency slices pre-allocated for different types of services on the first C-1 cores:

and

c is 1, …, C, I is 1, …, I, C is the number of cores of each link, I is the total number of service types in the network, and F is the number of frequency slices on each core;<r>_modnthe operator is a modulo n integer, and represents a multiple of the integer n nearest to the real number r; n is a radical of_iThe number of frequency slices required by the connection request of the ith type of service;

step three: calculating the number of frequency slices pre-allocated for different types of services on the C core:

and

wherein the content of the first and second substances,taking down an integer operator for modulo n, representing a maximum multiple of integer n that is not greater than real r;

step four: partitioning a chip pre-allocated to an i-th type of service into N_iWhen a connection request of the ith type of service arrives, the frequency slice pre-allocated for the ith type of service is used in units of frequency blocks, I is 1, …, I-1.

2. The method according to claim 1, wherein the precise number of total frequency slices required for calculating the i-th type of service is:

wherein, { p₁,p₂,…,p_IThe number of frequency chips (N) of each type of service₁,N₂,…,N_IThe corresponding ratio.

3. The method according to claim 1 or 2, wherein for the former I-1 type service, the difference between the pre-allocated number of frequency slices and the required precise total number of frequency slices does not exceed N_iI is 1, …, I-1, and the number of frequency slices pre-allocated for the type I service is not less than the precise total number of frequency slices required by the type I service.

4. The method according to claim 3, wherein the pre-allocated frequency slice for the first I-1 type service can be divided into an integer number of N-containing N_iA frequency slice, wherein I ═ 1., I-1; the frequency slice pre-allocated for the type I service may not be divided into an integer number of N-containing channels_IFrequency block of one frequency chip, but the number of frequency chips wasted on each core does not exceed N_I-1。

Technical Field

The invention relates to the technical field of optical networks, in particular to a frequency spectrum pre-allocation method based on accurate division in a space division multiplexing elastic optical network.

Background

With the rapid development of technologies such as 5G mobile communication, cloud computing, edge computing, and the like, the bandwidth demand of services in a network is higher and higher, and the difference of the bandwidth demand is also larger and larger. The traditional optical network based on the wavelength division multiplexing technology can only provide a transmission channel with fixed bandwidth, so that the bandwidth difference requirement of the service cannot be met. Elastic Optical Networks (EONs) are based on the Optical orthogonal frequency division multiplexing technology, and spectrum resources can be flexibly allocated to Optical Networks (EONs) according to bandwidth requirements of different services by means of equipment such as bandwidth-adjustable Optical repeaters and bandwidth-adjustable Optical cross connectors, so that the bandwidth difference requirements of the services can be met. In recent years, with the development of Space Division Multiplexing technologies such as Multi-Core Fiber (Multi-Core Fiber), the concept of Space Division Multiplexing flexible Optical Networks (SDM-EONs) has been proposed. In the spatial multiplexing elastic optical network, each link comprises a plurality of cores, and different cores can use the same frequency spectrum to transmit signals, so that the bandwidth of the network is greatly improved.

In the elastic optical network, spectrum resources on a link are generally divided into Frequency Slices (FS) having the same spectrum width and much smaller than the spectrum width occupied by each wavelength in the wavelength division multiplexing optical network, and each service may use a plurality of adjacent Frequency slices. In the space division multiplexing flexible optical network, each link comprises a plurality of cores, the frequency spectrum resources on each core are divided into frequency slices with the same frequency spectrum width, and each service can use a plurality of adjacent frequency slices on the same core but cannot use a plurality of frequency slices across the cores. As in the elastic optical network, three constraints are also satisfied when spectrum resources are allocated for services in the sdm elastic optical network: spectral continuity, spectral contiguity, spectral non-overlap. These limitations lead to spectrum resource fragmentation problems in space division multiplexed resilient optical networks. In addition, in the spatial division multiplexing flexible optical network, signals transmitted on different cores of the same link may interfere with each other, thereby causing inter-core crosstalk problems. Fragmentation problems and inter-core crosstalk problems can affect the effective use of spectrum resources, ultimately leading to an increase in traffic blocking rates.

In order to reduce the fragmentation degree of the spectrum resources, scholars propose methods for pre-allocating the spectrum resources to different types of services, such as a Spatial-Partition (Spatial-Partition) method, a Spectral-Partition (Spectral-Partition) method, and the like. The basic idea of the spectrum preallocation method is that according to the type of service in the network, the spectrum resource on the link is divided into a plurality of parts, the frequency slice in each part is preallocated to one type of service, the frequency slices are further divided into frequency blocks with the size equal to the size of the service, and the service uses the spectrum resource preallocated for the service by taking the frequency blocks as units. These spectrum pre-allocation methods can reduce the fragmentation degree of spectrum resources, and thus reduce the traffic blocking rate, but they have the following two disadvantages:

(1) the number of wasted frequency chips is large. In a spatial division multiplexing elastic optical network, each service cannot use multiple frequency slices across the core. When the number of frequency slices allocated for a type of traffic on a core is not an integer multiple of the traffic size, the extra frequency slices are wasted. If the pre-allocation scheme is not properly designed, more frequency slices are wasted on multiple cores.

(2) Different types of traffic are subject to crosstalk imbalance. In the elastic optical network of space division multiplexing, the crosstalk suffered by different cores on a link is different, the crosstalk suffered by a middle core is larger, and the crosstalk suffered by an edge core is smaller. If the chips on the intermediate core are pre-assigned to the same type of traffic, this type of traffic suffers much more crosstalk than other types of traffic.

Therefore, it is necessary to design a more reasonable spectrum resource pre-allocation method to achieve pre-allocation of spectrum resources while wasting frequency slices as little as possible and considering crosstalk balance.

Disclosure of Invention

Aiming at the technical problems that the existing pre-allocation method wastes a large number of frequency slices and suffers from unbalanced crosstalk among cores, the invention provides a frequency spectrum pre-allocation method based on accurate division in a space division multiplexing elastic optical network, which is characterized in that the accurate total number of the frequency slices required by each service is calculated according to the proportion of different types of services and the required size of frequency blocks, and then the frequency slices on each core are allocated to the different types of services according to the proportion and in consideration of the principle of integer division; the number of pre-allocated frequency slices can be matched with the requirement of the service, so that the number of wasted frequency slices is reduced as much as possible, and the crosstalk suffered by different types of services is balanced as much as possible.

In order to achieve the purpose, the technical scheme of the invention is realized as follows: a frequency spectrum pre-allocation method based on accurate division in a space division multiplexing elastic optical network comprises the following steps:

the method comprises the following steps: the service types are sorted from more to less according to the number of frequency slices required by each connection request, and the number of the frequency slices required by each connection request of the various types of services after reordering is recorded as { N₁,N₂,…,N_IIs N₁＞N₂＞…＞N_ICalculating the accurate total frequency chip number required by the ith service type

Step two: the number of frequency slices which are recorded on the c core and pre-allocated for the i type of service is F_c,iCalculating the number of frequency slices pre-allocated for different types of services on the first C-1 cores:

and

c, I, C is the number of cores of each link, I is the total number of service types in the network, and F is the number of frequency slices on each core;<r>_modnthe operator is a modulo n integer, and represents a multiple of the integer n nearest to the real number r; n is a radical of_iThe number of frequency slices required by the connection request of the ith type of service;

step three: calculating the number of frequency slices pre-allocated for different types of services on the C core:

and

wherein the content of the first and second substances,taking down an integer operator for modulo n, representing a maximum multiple of integer n that is not greater than real r;

step four: partitioning a chip pre-allocated to an i-th type of service into N_iWhen a connection request of the ith type of service arrives, the frequency slice pre-allocated for the ith type of service is used in units of frequency blocks, I is 1, …, I-1.

The precise total number of frequency slices required for calculating the ith type of service is as follows:

wherein, { p₁,p₂,…,p_IThe number of frequency chips (N) of each type of service₁,N₂,…,N_IThe corresponding ratio.

For the first I-1 type service, the pre-allocated frequency slice number and the required accurate total frequency slice numberHas a difference of not more than N_iI-1, and the number of frequency slices pre-allocated for the type I service is not less than the precise total number of frequency slices required by the type I service.

The pre-allocated frequency slice for the first I-1 type service can be just divided into an integer number of N_iA frequency slice, wherein I ═ 1., I-1; the frequency slice pre-allocated for the type I service may not be divided into an integer number of N-containing channels_IFrequency block of one frequency chip, but the number of frequency chips wasted on each core does not exceed N_I-1。

Compared with the prior art, the invention has the beneficial effects that: assuming that each link in the space division multiplexing elastic optical network is provided with C cores, and each core is provided with F frequency slices; assuming that I types of services exist in the network, calculating the accurate total number of frequency slices required by each type of service according to the proportion of different types of services and the required frequency block size; pre-allocating frequency slices for different types of services on the first C-1 cores according to the required accurate total number of frequency slices, and further pre-allocating frequency slices for different types of services on the C-th core; the pre-allocated frequency slice for each type of service presents three characteristics: 1) the number of pre-allocated frequency slices is close to the accurate total number of frequency slices required by the service; 2) the frequency slice pre-allocated for the first I-1 type service can be just divided into an integer number of frequency blocks; 3) the frequency slices pre-allocated for each type of traffic are distributed relatively evenly over the different cores. The invention pre-allocates the frequency slices on all the cores on the link to the services of different types, so that the number of the pre-allocated frequency slices is matched with the requirements of the services, the number of the wasted frequency slices is reduced as much as possible, and the crosstalk suffered by the services of different types is balanced as much as possible.

The invention allocates the frequency slices on each core to all types of services according to the accurate frequency slice number required by each type of services, in proportion and in consideration of the integer divisibility of the frequency block size. The number of the distributed frequency chips can be divided by the size of the frequency block corresponding to most types of services, so that the waste of the frequency chips is less; and because the frequency chips on the middle core are distributed to all types of services, the crosstalk suffered by different types of services is relatively balanced.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

FIG. 1 is a schematic flow chart of the present invention.

Fig. 2 is an example of spectrum resources on a link in a space division multiplexing flexible optical network according to the present invention, where the link is implemented by a 7-Core Fiber (Fiber), i.e., there are 7 cores (cores) in the link, Core 1 to Core 6 are around, and Core 7 is in the middle, so that the crosstalk experienced by Core 7 is much larger than that experienced by other cores; the spectral resources on each core are divided into 320 frequency slices.

Fig. 3 is a diagram illustrating the result of pre-allocating spectrum resources on all cores of a 7-core fiber to three types of traffic according to the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without inventive effort based on the embodiments of the present invention, are within the scope of the present invention.

As shown in fig. 1, in embodiment 1, a method for spectrum pre-allocation based on precise partition in a space division multiplexing elastic optical network includes the following steps:

the method comprises the following steps: the service types are sorted from more to less according to the number of frequency slices required by each connection request, and the number of the frequency slices required by each connection request of the various types of services after reordering is recorded as { N₁,N₂,…,N_IIs N₁＞N₂＞…＞N_ICalculating the accurate total frequency chip number required by the ith type of service:

in the space division multiplexing elastic optical network, C is the number of cores of each link, F is the number of frequency slices on each core, and I is the total number of service types in the network; { p₁,p₂,…,p_IThe number of frequency chips (N) of each type of service₁,N₂,…,N_IThe corresponding ratio. The exact total number of frequency slices required for each type of serviceAre positive real numbers, but not necessarily integers.

Step two: the number of frequency slices which are recorded on the c core and pre-allocated for the i type of service is F_c,iWherein C1, 1., C, I1., I; the number of frequency chips pre-allocated for different types of services on the first C-1 cores is calculated according to the following formula

And

wherein the content of the first and second substances,<r>_{mod n}the integer operator, modulo n, represents the nearest multiple of the integer n from the real number r. N is a radical of_iThe number of frequency slices required for the connection request of the ith type of service.

Step three: calculating the number of frequency slices pre-allocated for different types of services on the C core according to the following formula:

and

wherein the content of the first and second substances,the integer operator is taken for modulo n, representing the maximum multiple of integer n that is not greater than the real number r.

Step four: partitioning a chip pre-allocated to an i-th type of service into N_iI-1, when a connection request of the ith type of service arrives, using a frequency slice pre-allocated for the ith type of service in units of frequency blocks.

Two sets of equations are considered in one block:

(1) can not be wasted

For the first I-1 type of traffic, becauseAndare all exactly N_iInteger multiple of the number of the frequency slices, no frequency slices are wasted when further dividing into frequency blocks.

For type I traffic, waste may occur, at most N_IAnd (4) frequency slices. But due to N_IIs all N_iThe smallest of (I ═ 1, 2., I) results in the least waste.

(2) The number of the frequency chips obtained by division is close to the number of the actually required frequency chips

For the first I-1 type of service, the number of frequency chips divided is the sum of the number of frequency chips divided on each core, namely

Andis not more than N_iAnd thus are very close.

For the type I service, the number of the divided frequency slices is also equal to the sum of the number of the frequency slices divided on each core, namely

Since the total number C · F of the frequency slices is not changed, the number of the frequency slices divided by the first I-1 type of service is less than or equal to the number of the frequency slices actually required, and therefore the number of the frequency slices divided by the last type of service is certainly greater than or equal to the number of the frequency slices actually required.

This is so because the assigned tiles for type I traffic may be somewhat wasted and therefore somewhat compensated for when assigned.

(3) The crosstalk is fairer

Since the crosstalk experienced by cores in different locations is not the same, the crosstalk experienced by the core in the center location is the greatest. The frequency slices required by each type of service are distributed on different cores more uniformly, so that the frequency slices of different types of services are more fair.

Therefore, the number of the frequency slices pre-allocated for each type of service is matched with the requirement of the service, and particularly, for the first I-1 type of service, the difference between the number of the frequency slices pre-allocated for the service and the required accurate total number of the frequency slices is not more than N_iI-1, and the number of frequency slices pre-allocated for the type I service is not less than the precise total number of frequency slices required by the type I service.

Less frequency slices are wasted when pre-allocating frequency slices for different types of services, and particularly, the frequency slices pre-allocated for the first I-1 type of services can be just divided into an integral number of N containing_iThe frequency block of each frequency slice cannot be wasted, and the frequency slice pre-allocated for the type I service cannot be usedDivided into integer numbers comprising N_IFrequency slices, thereby generating certain waste, but the number of wasted frequency slices on each core does not exceed N_I-1. If the number of frequency chips reaches N_IIf there are more than one, a frequency block is formed, so the number of wasted frequency slices is at least 0 and at most N_I-1。

The crosstalk experienced by the frequency slices pre-assigned for the different types of traffic is relatively balanced, and in particular, the frequency slices assigned for each type of traffic are relatively evenly distributed across all cores and the frequency slices are relatively aligned, which allows inter-core crosstalk to occur primarily between the frequency slices assigned to the same type of traffic, and thus be relatively balanced.

Embodiment 2, a method for spectrum pre-allocation based on accurate partitioning in a spatial division multiplexing elastic optical network, which is implemented in combination with the link resource situation shown in fig. 2, specifically includes the following steps:

as shown in fig. 2, which is an example of spectrum resources on links in a space division multiplexed resilient optical network, each link contains 7 cores (cores), the 6 cores around the link are numbered 1 to 6, respectively, and the central Core is numbered 7. The signal transmitted on each core may be affected by the signal transmitted on the adjacent core, causing inter-core crosstalk. Each core distributed around has three adjacent cores and the central core has six adjacent cores, so that the crosstalk experienced by the central core is much greater than that experienced by the surrounding cores. The same spectral resources are used on each core and are divided into 320 frequency tiles. Each service may use multiple adjacent tiles on one core, but cannot use multiple tiles across the core.

The method comprises the following steps: in the space division multiplexing flexible optical network shown in fig. 2, the number of cores of each link is 7, and the number of frequency slices on each core is 320; assuming that the network has three types of services in common, the number of frequency slices required by each connection request of the three services is {15, 7, 3}, and assuming that the proportions of the various types of services are { 30%, 30%, 40% }, the accurate total number of frequency slices required by each type of service can be calculated:

step two: pre-allocating frequency slices for three types of services on the first 6 cores, wherein the number of the frequency slices allocated to each type of service is respectively

Step three: pre-allocating frequency slices for three types of services on the 7 th core, wherein the number of the frequency slices allocated to each type of service is respectively

Step four: dividing frequency slices pre-allocated for the three types of services on each core into frequency blocks with corresponding sizes:

(1) on the first six cores, the frequency slices pre-allocated for the first two services can be just divided into an integer number of frequency blocks, for example, 210 frequency slices pre-allocated for the first type of service can be just divided into 210/15-14 frequency blocks, and 77 frequency slices pre-allocated for the second type of service can be just divided into 77/7-11 frequency blocks. The frequency slice pre-allocated for the third service category may not be divided into exactly integer number of frequency blocks, which causes a certain waste, but in this embodiment, 33 frequency slices pre-allocated for the third service category may be divided into exactly 33/3-11 frequency blocks.

(2) On the last core, the frequency slices pre-allocated for the first two types of services may be just divided into an integer number of frequency blocks, for example, 225 frequency slices pre-allocated for the first type of services may be divided into 225/15-15 frequency blocks, and 56 frequency slices pre-allocated for the second type of services may be divided into 56/7-8 frequency blocks. The frequency slice pre-allocated for the third type of service may not be divided into exactly an integer number of frequency blocks, which may cause some waste, but in this example, the 39 frequency slices pre-allocated for the third type of service may be divided into exactly 39/3-13 frequency blocks.

The result of dividing the pre-allocated frequency slices of different types into frequency blocks is shown in fig. 3, wherein B (15), B (7), and B (3) respectively represent the frequency blocks including 15, 7, and 3 frequency slices. It should be explained that when pre-allocating frequency slices for the first two services on the last core, modulo-n round-off (rather than modulo-n round-off) operation is used, this may result in the number of frequency slices pre-allocated for the first two services being less than their exact total number of frequency slices required, while the number of frequency slices pre-allocated for the last service being greater than their exact total number of frequency slices required. The reason for this is as follows: when the frequency slice is divided into frequency blocks, the frequency slices allocated to the first two types of services can be just divided into an integral number of frequency blocks, so that waste is not generated; the frequency slice allocated to the last type of service may be wasted because it cannot be divided into an integer number of frequency blocks, and thus the last type of service is compensated for within a controllable range.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.