SE514355C2 - Method and device for allocating resources to nodes in a network - Google Patents

Abstract

The present invention relates to a resource reallocation scheme and arrangement in a communications network in which network nodes may control their resource ownership and usage and communicate with each other without severely reducing the amount of resources. The invention comprises the steps of assigning priorities to at least a first node (Node i-1) and a second node (Node i), calculating a quotient of resources as the amount of resources of the first node (Node i-1) divided by the amount of resources of the second node (Node i), calculating a quotient of prio ri ties as the priority of the first node (Node i-1) divided by the priority of the second node (Node i), transferring resources from the first node (Node i-1) to the second node (Node i) in dependence of the quotient of resources being larger than the quotient of priorities, and transferring resources from the second node (Node i) to the first node (Node i-1) in d ependence of the quotient of resources being smaller than the quotient of priorities.

Description

The total resource capacity is divided into cycles of 125 microseconds, which are then divided into gaps. A slot comprises 66 bits, where 64 bits form a data frame and two bits are used as control bits. Most of the gaps, or brief gaps, for data transfer. A small number of slots, called static called dynamic slots, are used slots, used for signaling among the nodes, for example in the form of broadcast channels. Each node can have one or more static slots in each cycle.

Each hatch has exactly one owner at a given time. If full control use in terms of channel connection and disconnection and a node has a slot, it has in its way prevents other nodes from using it for data transmission When the system is started, the data slots are assigned to the nodes according to some predefined distribution.

A node can ask other nodes for slots if it does not have enough data slots to connect a requested channel. Such a transfer of gaps from one node to another is called redistribution of gaps.

DTM uses a distribution algorithm1 for slot assignment, among Upon receipt of a user request, the node first considers its where the dynamic slots are distributed the nodes. local pool with free gaps to see if it has enough gaps to satisfy the request. If satisfy one more Otherwise, the other nodes ask for more gaps. the node cannot request sends the channel connection message the destination node or nodes.

This is done by signaling in static gaps. A node that receives such a request and owns unused dynamic slots provides slots to the requesting node.

If a user requests a channel for M slots and the node has N free slots, where N <M, it requests M - N slots from other nodes. The node refers to its state table, which is updated time and time again and includes information about free slots in other nodes, and sends a request to the nearest node with free slots. If this node does not have enough ie. M - N i sends a request to the next nearest number of free slots, in accordance with the condition table, the node with free slots and, if necessary, to correspond to the requested number of slots, also to the third, fourth, etc. nearest node with free gaps. The node waits until a response for each request has been received and grants or denies the channel request, depending on the result of the then redistribution procedure. A node that has J free slots and receives a slot redistribution request if K slots always give away min (J, K) denotes a certain amount of slots that are never given away for slots or at least min (Jn, K), where n to prevent the network is depleted.

The main problem is that the distribution of resources, ie. gaps, are determined when the system is started and cannot be easily controlled during system use.

Another problem is that redistribution of resources is started only with channel connection, which consequently increases the channel connection delay.

Another problem from a fairness point of view may be that each node starts sending requests for resources to other nodes in relative distance order. The average distance can be twice as far from nodes at the end of the bus as for nodes in the middle of the bus. That is, this means that the nodes in the middle of the bus are requested for resource transfers much more often than the nodes at the end of the bus, which means that the nodes in the middle of the bus have fewer resources on average.

Summary of the invention The object of the invention is to provide a method and arrangement for resource redistribution, which dynamically redistributes resources among the nodes in accordance with a changed "quality of service" distribution and / or a changed network use.

Another object is to dynamically reallocate resources in order to minimize channel connection delay. state of the art.

This is accomplished through a resource redistribution procedure and arrangement, which redistributes resources among the nodes depending on the current resource allocation, as well as some predefined static or dynamic priority allocation.

The method and inventive arrangement of at least one comprises assigning priorities first node and a second node, calculating a resource ratio as the amount of resources of the first node divided by the amount of resources of the second node and calculating a priority ratio as the priority of the first node divided by the priority for the second node. It further includes transferring resources from the first node to the second node depending on whether the resource ratio is greater than the priority ratio, and transferring resources from the second node when the resource ratio is less than the priority ratio. the quotas and transfer to calculate are performed at a given time, the steps resources for example when the other node. has 'disconnected or' connected one. channel, men. never this would increase the channel connection, because the channel connection delay. at 10 15 20 25 Preferably, the amount of resources to be transferred is set so that the differences between the two quotas are less than a predetermined value.

The first and second nodes are preferably the nearest logical neighbors in a logic ring with nodes assigned to the network.

If the request for and transfer of resources is required and approved for a channel connection of a node, the node asks its nearest nodes in an order so that the adjacent nodes closest in the assigned logical ring get the question first. The method and arrangement according to the invention redistributes dynamic resources among the nodes, among other things, in order to improve network usage.

An advantage of adjusting for changes in the distribution of quality of service invention is that it can dynamically (QOS) among the nodes and / or changes in a network performance parameter.

Another advantage of the invention is that the channel connection delay is minimized due to fewer requests before channel connection.

Another advantage is that justice, in terms of achieving equal opportunity for two nodes with identical quality of services to use the network resources, is obtained.

BRIEF DESCRIPTION OF THE DRAWINGS The invention is fully understood from the detailed description given below and the accompanying figures, which are given by way of illustration only and are, therefore, not to be limiting of the invention. Fig. 1 shows a possible logical ring assignment in accordance with the invention.

Fig. 2 shows distributions of free and used data gaps among the nodes in a logic ring.

Detailed description of embodiments To enable fair use of the redistribution method, the nodes in a network are sorted in a logical ring, which is independent of the physical network topology, for example double bus topology. The logic ring is used only to determine the order of the nodes, when requesting or transmitting gaps. The order of the nodes 'can' be selected in many different ways, but to minimize the difference between the physical distances of logic node neighbors, the assignment shown in Fig. 1 is preferred. The logic ring is indicated by curved arrows. All nearest logical node neighbors, for example node 0 and node 1, -1, which are also nearest physical neighbors. separated by another node, for example node N- except for two pairs, ie. node i, node i + 1 and node N-1, node 0, Fig. 1 shows only one logic ring assigned to the two buses bus 1, bus 2 in the network, but no restrictions are made in the selection of different logic rings for the two different buses.

An example of data slot distribution among the nodes is shown in Fig. 2, in the case of the logical ring assignment according to Fig. 1. The length of the rectangles is proportional to the number of slots, i.e. the amount of resources belonging to the respective node. The height of the rectangles corresponds to the hatch category; low height corresponds to free gaps and high height corresponds to used gaps. The rectangles corresponding to gaps belonging to the same node have the same pattern. 10 15 20 25 disconnection of I fig. Divided so that they indicate that they can, ha. one. arbitrary 1 and 2 are the buses and the logic rings size.

To enable different quality of services for different network customers, a priority assignment is introduced. Each node has two A higher priority value nw corresponds to a higher probability of having a priority value. one for each bus bus 1, bus 2. large amount of resources, free or in total, and lower channel connection delays in. average. The priority values, the arrangement for reallocating resources, are used in the method according to the invention and as described below.

First, the procedure is started by a node, for example node i, after disconnection of a channel has been considered. 4 Node i. Can request gaps from or transfer gaps to its previous logical neighbor node, ie. node i-1, after 'each call. The procedure is based on the following algorithm: 1. Node i places the gaps that have been made vacant during the end of the call in its pool of free gaps. gaps 2. A resource ratio is calculated as the amount of resources, ie. in node i divided by the amount of resources in node i-1. 3. A priority ratio is calculated as the priority of node i divided by the priority of node i-1. Gaps are transferred from node i to node i-1, if the resource ratio and from node i-1 to node i, if the resource ratio is less than the priority ratio. is greater than the priority ratio, 5. The number of gaps to be transferred is calculated in such a way that the resource ratio is equal to the priority ratio or at least so that the difference between the quotas becomes less than a predetermined value Ö, i.e. so that | resource ratio - priority ratio | <5.

The number of slots to be transmitted may be smaller if a number n is reached, where n refers to a number of slots that a node must always have. These slots are never removed to prevent a node from eventually having any free slots.

When calculating the resource ratio can. the total number of gaps, ie. use gaps' and free gaps, at a. nod beaktas. The node i can acquire the information about the priority and the number of gaps of node i ~ 1 on request or by reading a constantly updated table, which is stored in node i or elsewhere. Alternatively, only vacant gaps are considered, where the number of gaps in node i-1 can be obtained by reading the permit table for node i. The type of resources that are important must be considered.

For example, if the resource allocation process is started in node i with a slot distribution according to Fig. 2, slots are transferred from node i to node i-1 only if free slots are taken into account, and from node i-1 to node i if the total number of slots is considered. Which alternative is to be realized depends on which network performance parameters soul is considered most important, on how fairness between the nodes is valued, and the existence of the offered variation on QOS.

A procedure similar to the above is started by node in after a channel connection. Node i removes the hatches from the pool with free slots, which are used for the duct connection.

The method then follows steps 2-5 of the above algorithm. after using the channel disconnection and / or Through the proposed algorithms - the connection is reduced or even eliminated the need for redistribution of resources in channel connection, ie. that a node does not have sufficient bandwidth to connect a requested channel, especially when using an appropriate priority assignment, as described below.

Sometimes it can be advantageous to carry out the steps of calculating the quotas and transfer the resources at other times. In general, the steps can be carried out at a given time or periodic conditions. at certain or times. when a network performance parameter satisfies a predetermined condition, for example when the traffic intensity is less than a predetermined value.

Two different types of priority allocation are considered; static and dynamic priority assignment.

By using a static priority network, the priorities can only be changed manually on request, for example at management level. Two examples of static priority assignment follow.

The first example is a client-server network with a server node at one end of a dual bus network. All other nodes communicate with this server node by sending their traffic to the server node on a first bus. The second bus is consequently used for traffic from the server node to the other nodes. If the load generated by the client nodes is equally distributed, the ability to request gaps should be equal among In order to achieve a realization of such to the clients. balanced network, where each node has the same possibility to establish channels without. to request gaps from other nodes, the following priority assignment is required. 10 15 20 25 30 10 0 Equal priority for all clients and zero priority for the server on the first bus, where the server is located at the end of the bus. 0 Zero priority for clients and high priority for the server on the second bus, which is directed in the opposite direction to the first bus.

The second example concerns a network with call destinations evenly distributed among the network's nodes. peer-to-peer Here is the load generated by the network depending on the position of the nodes.

For example, it changes linearly along the bus from 0 to 100% in the opposite direction on two opposite buses.

If justice is required, priorities must be set based on the location of the nodes. The sum of the priority numbers on the two buses can be equal. As described, in both examples, a fair distribution of the channel connection probabilities is desirable. However, service quality parameters are also possible to define priority as, one and consequently set different priorities for different nodes in accordance with the sorted QOS.

Another type of priority assignment is dynamic priority assignment. Here the priorities a are calculated as used and / or repeatedly The priorities can also / in posteriori from the nodes' network performance parameters, resource amount and channel connection-disconnection frequency, and are updated as the traffic changes. the place changes in accordance with an a priori-estimated traffic distribution.

If a node lacks the resources to fulfill a user request, it can request resources from other nodes, as described in the prior art. Three different modified 10 15 20 25 30 113 methods for reallocating resources during channel connection are proposed.

The first method simply does not allow door redistribution before duct connection. This simplifies the complexity of the node. If a dynamic priority assignment is used together with an efficient algorithm. it may not cause appreciable deterioration in performance.

The second method is similar to the original algorithm.

The difference is that the order of the door requests is specified in accordance with the logical ring assignment and not by physical distance. Consequently, a node, for example node i, requests resources from other nodes in an order so that resources are requested from the nearest logical neighboring nodes, for example node i-1, only in case it does not have sufficient own resources for a requested channel connection. procedure. the second, for example, node i, The procedure is similar The difference is that a node, third requests resources from only another node, its nearest logical neighbor node, in case it does not have enough own resources for a requested channel connection. If this second, for example, node il, does not satisfy said request requests that, for example, node i. Node, for example node il, its nearest logical neighbor node, for example node i-2, transfer resources to the first node, this until fulfilled or until the number of nodes involved has reached a procedure can be repeated said request has specified maximum value. The latter condition can be monitored by introducing a counter of requests.

In particular, if node i-1 has Jbl free slots and receives a slot redistribution request if I (slots, it gives away my (Jp1-nb1, K) slots to node i, where nbl never denotes the number of slots as node i-1 Then, if Jn <K node i-1 requests that node i-2 transfer K- (Jpl-npl) slots to node i. Node 10 15 20 25 12 which has Jb, free slots, slots to node i, i-2, gives away min (Ji_1-ni-1, K- (Ji_1- np1)) where npl denotes the number of gaps that node i-1 never gives away.This procedure can be repeated, as described above, and consequently more nodes involved.

The method and arrangement according to the invention redistributes resources among the nodes in order to increase the use of the network and reduce average channel connection delays and at the same time enable the nodes to have different priorities, ie. different ability to request resources from an adjacent node.

Advantages of the invention include, inter alia, the following: Better use. Minimizing the number of door requests improves network usage. 0 Lower call set-up delay. More channels can be connected from the pool with their own free slots.

I By using dynamic priorities, the number of free slots can be dynamically adjusted according to a node's future needs. If the use of a sophisticated dynamic priority assignment, for example based on used bandwidth and channel connection and / or disconnection frequency, etc. the number of door requests before duct connection can even be eliminated. 0 Ability to handle. the priority allocation concept as one to service quality. tools for providing network users with varied 10 13 0 Justice. The invention. can. achieve equal. possibility for the users of the nodes to use the network resources by assigning appropriate priority values and introducing the concept of logical ring / nearest logical neighbor node.

One skilled in the art will readily appreciate the various methods described above. as one can vary on Such are considered the scope of the invention. All such modifications as are obvious to the invention many variations not departing from for one skilled in the art are intended to be included within the scope of the following claims.

Claims (20)

10 l5 20 25 14 Patent claim 1. Resource redistribution method 5. a communication network, where the network nodes can control their resource ownership and in their resource use and communicate with each other without significantly reducing the amount of resources, characterized by the steps - assign priorities to at least one first node (node (node). il) and a second node (node i), - calculate a resource ratio such as the resource set of the first (node il) (node i), the node divided by the resource set of the second node - calculate a priority ratio as the priority of the first (node il) (node i), the node divided by the priority of the second node - transfer resources from the first node (node il) to the second node (node jJ when the 'resource ratio' is greater than the priority ratio, and to the first node ( node il) when the resource ratio is less than a priori- - transfer resources from the other node (node i) tet ratio. A method for reallocating resources according to claim 1, characterized by calculating the resource ratio such as the amount of free resources (node i-1) g resources in the first node divided by the amount of free resources in the second node (node i). A method of reallocating resources according to claim 1, characterized by calculating the resource ratio such as the amount of free and used resources of the first node (node il) divided by the amount of free and used resources of the second node (node i). ". .v _ nM-N- »w 10 15 20 25 30 15 Method for 'redistribution' of: resources ~ according to any one of claims 1-3, characterized by the transfer of resources until the resource ratio is equal to the priority ratio. A method of reallocating resources according to any one of the characteristics of the transfer of resources until claims 1-3, the difference between the resource ratio and the priority ratio is less than a predetermined value (8). A method for reallocating resources according to any one of claims 1-5, characterized by performing the steps of calculating quotas and transferring resources when a predetermined period of time has elapsed. A method for reallocating resources according to any one of claims 1-5, characterized by performing the steps of calculating quotas and transferring resources when a network performance parameter satisfies a predetermined condition. A method for reallocating resources according to any one of claims 1-5, characterized by performing the steps of calculating quotas and transferring resources when the second node (node i) has disconnected a channel. Method for 'redistribution' of: resources according to any one of claims 1-5, characterized by performing the steps of calculating quotas and transferring resources, when the second node (node i) has connected a channel. A method for reallocating resources according to any one of claims 1-9, characterized by assigning a logic ring among the nodes in the communication network and assigning the first node (node i-1) and the second node (node i) as the nearest neighboring logic nodes. A method for reallocating resources according to any one of claims 1 to 10, "characterized by assigning priorities to the nodes in the communication network in accordance with a predetermined quality of service distribution among the nodes. Characterized by changing the priority of at least one node in the Resource Redistribution Method according to claim 11, the communication network according to an a priori or an a posteriori estimate of at least one network performance parameter. Claims 1 to 12, A method for reallocating resources according to any one of claims characterized in that the second node (node i) requesting resources from other nodes in an order such that resources are requested from the nearest logical neighboring nodes only in the case where the other the node (node i) does not have sufficient resources for a requested channel connection. Claims 1 to 12, A method for reallocating resources according to any one of characterized by rejecting the second node (node i) request for resources in the event that it does not have sufficient resources for a requested channel connection. 15. Device for reallocating resources in a communication network, where * the network nodes. can control its resource ownership and its. resource use and. communicate. with. without appreciably reducing the amount of resources, characterized by means for carrying out the steps of - assigning priorities to at least a first node (node i-1) and a second node (node i), - calculating a resource ratio as the amount of resources of the first ( node i-1) (node i), the node divided by the amount of resources of the second node - calculate a priority ratio such as the priority of the first node (node i-1) divided by the priority of the second node (node i), 10 15 20 25 ut; i 17 '- transfer resources from the first node (node i-1) to the second node (node i) the ratio, when the resource ratio is greater than a priori- and' to the first node (node i-1) when the resource ratio is less than the priority ratio. - transfer resources from the other node (node i) Device for redistribution of resources according to claim 15, characterized by a needle for performing the steps of calculating quotas and transferring resources when a channel has been disconnected or connected. Node in a communication network, where the network nodes can control their resource ownership and resource use and communicate with each other without appreciably reducing the amount of resources, characterized by being arranged to - store a priority, - store or acquire information about priority and quantity resources in another node (node i-1), - calculate a resource ratio as the amount of resources in said second node (node i-1) divided by its own resource set (node i), - calculate a priority ratio as the priority of the other node ( node i-1) divided by its own priority (node i), (node i-1) when the resource ratio is less than the priority ratio, and - transfer resources to the other node - request resources from the other node (node i-1 ) when the resource ratio is greater than the priority ratio. 10 15 20 i 18 Node according to claim 17, characterized by being arranged to calculate quotas and transfer resources after disconnecting or connecting a channel. Communication network, where the nodes can control their resource access and use and communicate with each other without appreciably reducing the amount of resources, characterized in that a node is arranged to - store a priority, - store or acquire information about priority and amount of resources at another node (node i-1), - calculate a resource ratio as the resource set of the other node (node i-1) divided by the resource set of its own node (node i), - calculate a priority ratio as the priority of a second node (node i-1) divided by the priority of the own node (node i), (node i-1) when the resource ratio is less than the priority ratio, and - transfer resources to the other node - request resources from the other node (node i-1) when the resource ratio is greater than the priority ratio. 20. characterized in that a node is arranged to calculate quotas and transfer resources. The communication network according to claim 19, looks after having disconnected or connected a channel.