Patent Application
20140192683
The invention relates generally to a software-defined network “SDN”. More particularly, the invention relates to a method, a controller system, a network element, and a computer program for configuring one or more network elements of a software-defined network.
Software-defined networking is an emerging architecture for data transfer networks. In a software-defined network “SDN”, the control plane is separated from the data plane so that the control plane is implemented in one or more controllers that can be separate from the network elements and the data plane is implemented in the network elements. The network elements can be, for example, Internet Protocol “IP” routers, multiprotocol label switching “MPLS” nodes, packet optical switches, and/or Ethernet switches. Each network element may consist of a single apparatus or a combination of a plurality of apparatuses. Typically, the software-defined networking allows for quick experimenting and optimization of switching and/or routing policies and external access to the innards of network elements that formerly were closed and proprietary.
Internet Protocol “IP” based networks were initially built based on the concept of Autonomous Systems “AS”. This concept allows networks to scale and extend by connected junctions that forward packets to a reasonable next hop based on partial need-to-know information. The AS principle works much like the traditional post office service, where a postal worker in a given city does not need to know all the tenants of all the streets in another city in order to choose a reasonable next hop for a letter at hand. This approach to networking is simple, and has proven resilient and scalable. This approach has, however, a few drawbacks. It does not allow the designated destinations, or tenants with home mail-boxes, to move without changing their identity as far as the packet delivery service is concerned. The topological location of destinations, which is the network interface they are attached to, dictates their identity related to the packet delivery service. In addition, using only the basic AS principle, it is hard to specify other qualities, such as logical grouping, access control, quality of service, intermediate network processing, or to specify aspects that relate to a sequence of packets that form a flow.
Using the analogy of the postal service, the software-defined networking works, for any given street location, so that all the letters from all the tenants would first be aggregated by a network element on an edge a software-defined network. This network element is configured to examine the current location for each of the letter-destinations using a global lookup mechanism. Based on that global lookup and on other globally defined and globally measured considerations, such as access control or remote location load conditions, the said network element places one or more of the original letters in an additional envelope addressed to each of the street locations where the destinations currently are. It then uses the normal postal service which works like the traditional Internet Protocol “IP” to get these outer envelopes to the remote locations. This is done based on the existing and scalable hop-by-hop forwarding services. The outer letters are then opened by a remote network element and the original envelopes are delivered to the destinations. It is to be noted that the above-presented analogy between the software-defined networking and the postal service is a strong simplification and it gives only a limited viewpoint about the versatile possibilities provided by the software-defined networking.
The software defined networking is, however, not free from challenges. Some of the challenges are related to configuring the network elements so that they are constantly capable of carrying out the above-illustrated tasks and so that the resources of the network elements are utilized in a sufficiently optimal way. Therefore, there is still a need for technical solutions for configuring network elements of software defined networks.
The following presents a simplified summary in order to provide a basic understanding of some aspects of various invention embodiments. The summary is not an extensive overview of the invention. It is neither intended to identify key or critical elements of the invention nor to delineate the scope of the invention. The following summary merely presents some concepts of the invention in a simplified form as a prelude to a more detailed description of exemplifying embodiments of the invention.
In accordance with the invention, there is provided a new method for configuring a network element of a software-defined network. A method according to the invention comprises:
The resource consumption estimate of the look-up system can be used for avoiding situations where a network element is not sufficiently resourced to run a look-up system configured to it. Furthermore, resource consumption estimates related to different network elements can be used for optimizing the software-defined net-work as a whole, e.g. for optimizing the routing topology of the software-defined network.
The look-up table specific resource consumption estimates can be read, for example, from data frames received from the network element to be configured or they can be received from another device or system which is arranged to store sufficient information about the said network element. Furthermore, in methods according to some exemplifying embodiments of the invention, the look-up table specific resource consumption estimates can be read from a local memory which is arranged to store sufficient information about network elements to be configured.
In accordance with the invention, there is provided also a new controller system for configuring a network element of a software-defined network. The controller system can be a single apparatus or a combination of a plurality of apparatuses capable of communicating with each other. A controller system according to the invention comprises a processing system adapted to:
In accordance with the invention, there is provided also a new software-defined network that comprises one or more network elements and a controller system according to the invention for configuring the one or more network elements. It is to be noted that the controller system or one or more parts of it can also act as one or more network elements that can be e.g. an IP (Internet Protocol) router, a multi-protocol label switching “MPLS” node, a packet optical switch, and/or an Ethernet switch.
In accordance with the invention, there is provided also a new network element for a software-defined network. The network element can be, for example, an Internet Protocol “IP” router, a multiprotocol label switching “MPLS” switch, a packet optical switch, and/or an Ethernet switch. The network element may consist of a single apparatus or a combination of a plurality of apparatuses. Furthermore, the network element may have a modular structure so that it comprises one or more line-interface modules and one or more other modules such as a control module. A network element according to the invention comprises:
In accordance with the invention, there is provided also a new computer program for configuring a network element of a software-defined network. A computer program according to the invention comprises computer executable instructions for controlling a programmable processing system to:
In accordance with the invention, there is provided also a new computer program product. The computer program product comprises a non-volatile computer readable medium, e.g. a compact disc “CD”, encoded with a computer program according to the invention.
A number of exemplifying and non-limiting embodiments of the invention are described in accompanied dependent claims.
Various exemplifying and non-limiting embodiments of the invention both as to constructions and to methods of operation, together with additional objects and advantages thereof, will be best understood from the following description of specific exemplifying embodiments when read in connection with the accompanying drawings.
The verbs “to comprise” and “to include” are used in this document as open limitations that neither exclude nor require the existence of also un-recited features. The features recited in depending claims are mutually freely combinable unless otherwise explicitly stated.
The exemplifying and non-limiting embodiments of the invention and their advantages are explained in greater detail below with reference to the accompanying drawings, in which:
As well, the controller system 105 may consist of a single apparatus or a combination of a plurality of apparatuses. In the exemplifying case illustrated in
Each of the network elements 101-104 comprises a control processor section for maintaining a look-up system that comprises at least one look-up table defining actions to be executed in conjunction with managing data frames, and a data forwarding section for managing the data frames in accordance with the look-up system. The at least one look-up table of the look-up system may comprise, for example but not necessarily, one or more successive flow tables and a group table according to the OpenFlow specification. The OpenFlow is managed by the Open Networking Foundation “ONF”. In the exemplifying case where the OpenFlow is used, each flow table in a network element contains a set of flow entries. Each flow entry may consist of match fields, counters, and a set of actions to apply to matching data frames. Matching typically starts at the first flow table and may continue to additional flow tables. Flow entries can be arranged into a priority order and the first matching entry in each table is the one being used. If a matching flow entry is found, the one or more actions associated with this specific flow entry are executed. If no match is found in a flow table, the data frame may be forwarded to the controller system 105 over an OpenFlow channel between the network element under consideration and the controller system 105, the data frame may be dropped, or the data frame may continue to the next flow table or to the group table. Actions associated with each flow entry may comprise for example data frame forwarding, data frame modification, group table processing, and pipeline processing. Pipeline processing actions allow data frames to be sent to subsequent flow tables for further processing and allow information, in the form of metadata, to be communicated between the flow tables. Table pipeline processing stops when the one or more actions associated with a matching flow entry does not specify a next table. At this point the data frame under consideration is usually modified and forwarded. The group table processing actions allow data frames to be sent to the group table for further processing and allow information, in the form of metadata, to be communicated to the group table. The group table contains group entries, where each group entry may contain a list of actions to be directed data frames defined to belong to a particular group.
The controller system 105 is adapted to send to each of the network elements 101-104 configuration data that comprises data items for configuring the network element to maintain the look-up system in accordance with, for example, a pre-determined collection of actions defined by a protocol, e.g. the OpenFlow, for configuring the network elements. Furthermore, the configuration data may comprise one or more configuration programs each of which comprising one or more computer executable instructions defining an action or a chain of actions to be executed in conjunction with managing data frames in the network element under consideration. The control processor section of each of the network elements 101-104 can be adapted to associate each configuration program to the one or more look-up tables of the look-up system so that the configuration program is callable to be executed in conjunction with managing data frames with the aid of the one or more look-up tables. The data forwarding section of each of the network elements 101-104 can be adapted to execute the action or the chain of actions defined by the configuration program in response to a situation in which the managing a particular data frame includes an order to execute the configuration program. The action or the chain of actions defined by the configuration program may comprise for example: reading data from a data frame, modifying the data frame, selecting one or more of egress ports of the network element and forwarding the data frame and its possible duplicates to the selected one or more egress ports, selecting one of the look-up tables maintained by the network element and executing a look-up from the selected look-up table, performing arithmetic operations, branching operations, performing logical operations, reading metadata associated with the data frame, writing metadata associated with the data frame, modifying metadata associated with the data frame, dropping the data frame, and/or duplicating the data frame.
As a preliminary action prior to configuring the network element 101, 102, 103 and/or 104, the controller system 105 retrieves look-up table specific resource consumption estimates from the network element under consideration, or from another device or system which is arranged to store sufficient information about the said network element, or from a local memory of the controller system 105. In a software-defined network according to an exemplifying embodiment of the invention, the controller system 105 sends one or more first data frames to a network element to be configured and reads the look-up table specific resource consumption estimates from one or more second data frames received from the network element. The one or more first data frames contains descriptors of the look-up tables so as to enable the network element to send, to the controller system 105, appropriate look-up table specific resource consumption estimates. Each of the descriptors of the look-up tables may comprise, for example, the maximum number of entries of the corresponding look-up table, the size of one or more look-up keys of the corresponding look-up table, allowed don't care key matching bit positions in the corresponding look-up table, definitions of tasks related to the entries of the corresponding look-up table, maximum look-up frequency of the corresponding look-up table, and/or the average time between successive updates of the corresponding look-up table.
Each look-up table is a potential building block of the look-up system of the network element to be configured and each look-up table specific resource consumption estimate indicates a load factor of at least one component of the network element when running operations related to the corresponding look-up table. The at least one component of the network element may comprise, for example, the data forwarding section of the network element, the control processor section of the network element, one or more random access memories “RAM” of the network element, one or more content access memories “CAM” of the network element, one or more ternary content access memories “TCAM” of the network element, and/or one or more memory busses of the network element.
The controller system 105 is adapted to determine, on the basis of the look-up table specific resource consumption estimates, a resource consumption estimate for at least one look-up system comprising one or more of the look-up tables and having functionality suitable for the network element. The controller system 105 is adapted to check, on the basis of the resource consumption estimate of the look-up system, whether the network element to be configured is capable of running the look-up system under consideration, and to send, to the network element, configuration data defining the look-up system in response to a situation in which the network element is capable of running the said look-up system. The look-up table specific resource consumption estimates may comprise, for example, numerical values indicating load factors of one or more components of the network element. For example, when running operations related to a first exemplifying look-up table, the load factor of the data forwarding section of the network element may be e.g. 50% of the processing capacity, the load factor of random access memories “RAM” of the network element may be e.g. 35% of the memory capacity, the load factor of content access memories “CAM” of the network element may be e.g. 45% of the memory capacity, and the load factor of memory busses may be e.g. 45% of their transfer capacity. Correspondingly, when running operations related to a second exemplifying look-up table, the load factor of the data forwarding section of the network element may be e.g. 46%, the load factor of random access memories “RAM” of the network element may be e.g. 38%, the load factor of content access memories “CAM” of the network element may be e.g. 48%, and the load factor of memory busses may be e.g. 25%. Thus, the look-up table specific resource consumption estimates for the first and second exemplifying look-up tables are the following vectors: [50%, 35%, 45%, 45%] and [46%, 38%, 48%, 25%], respectively.
The controller system 105 is advantageously is adapted to compute, for each of the components of the network element, a sum of the numerical values related to that component of the network element. In this case, the sums related to the components of the network element can be deemed to constitute the resource consumption estimate of a look-up system under consideration. For example, the resource consumption estimate of a look-up system consisting of the above-mentioned first and second exemplifying look-up tables would be [96%, 73%, 93%, 70%]. If any of these numerical values exceeded 100%, the network element under consideration would not be capable of running the look-up system under consideration. In this case, a different look-up system should be defined for the network element. It is worth noting that the controller system 105 does not need to know that e.g. the first elements of the vectors representing the look-up table specific resource consumption estimates are related to the data forwarding section of a network element under consideration. Furthermore, the look-up table specific resource consumption estimates may have a different number of numerical values in conjunction with different network elements. For example, the look-up table specific resource consumption estimates related to the network element 102 can be vectors each having two numerical values whereas the look-up table specific resource consumption estimates related to the network element 104 can be scalars each having only one numerical value. Thus, the controller system 105 can be made capable of understanding the look-up table specific resource consumption estimates irrespective of the implementations of the network elements 101-104.
The resource consumption estimate of a look-up system can be used for avoiding situations where the network element 101, 102, 103 and/or 104 is not sufficiently resourced, i.e. does not have sufficiently effective processors and/or sufficient memory, to run a look-up system configured to it. Furthermore, the resource consumption estimates of the look-up systems of the network elements 101-104 can be used for optimizing the software-defined network 100 as a whole. For example, the controller system 105 can be adapted to configure a routing topology of the software-defined network 100 so that more data traffic load is allocated to a network element under consideration when the resource consumption estimate of the look-up system manifests lower loading of the network element than when the resource consumption estimate of the look-up system manifests higher loading of the network element. The network element may have a modular structure so that it comprises one or more line-interface modules, such as ingress and egress modules, and one or more other modules such as a control module and/or a powering module. In this case, each look-up table specific resource consumption estimate can be provided by a module indicator which expresses the module containing the component whose load factor is indicated by the look-up table specific resource consumption estimate. The module indicator may express the individual module or only a type of the module such as: an ingress module, an egress module, or a module related to control of the network element as whole. The module indicators can be utilized when optimizing the routing topology of the software-defined network 100 for example in a situation where e.g. one or more line-interface modules are at a risk of being overloaded whereas other parts of the network element have margin in their resource consumptions.
The look-up tables, for which the look-up table specific resource consumption estimates are retrieved from the network element under consideration or from another source, may constitute a set of look-up tables with the aid of which alternative look-up systems having functionalities suitable for the network element are capable of being constructed. In a software-defined network according to an exemplifying embodiment of the invention, the controller system 105 is adapted to search, on the basis of the look-up table specific resource consumption estimates, such one of the alternative look-up systems which has, according to a pre-determined criterion, a sufficiently advantageous resource consumption estimate. For example, the controller system 105 can be adapted to compute resource consumption estimates for two or more of the alternative look-up systems on the basis of the look-up table specific resource consumption estimates related to the look-up tables contained by the two or more alternative look-up systems and to select a particular one of the two or more alternative look-up systems which has, according to a pre-determined criterion, the most advantageous resource consumption estimate. For another example, the controller system 105 can be adapted to use a genetic algorithm implementing simulated evolution for searching a suitable one of the alternative look-up systems. For the third example, the controller system 105 can be adapted to use the Dijkstra, A*, or other suitable optimization algorithm for searching a suitable one of the alternative look-up systems from a search network where the look-up tables are related to nodes of the search network and the look-up table specific resource consumption estimates represents costs related to the nodes or to connections between adjacent nodes. An optimal one of the alternative look-up systems corresponds to a minimum cost route through the search network so that look-up tables related to the minimum cost route form a look-up system having the functionality suitable for the network element. The above-mentioned pre-determined criterion can be, for example, such that a greatest one of numerical values constituting the resource consumption estimate of a look-up system should be as small as possible in order to avoid adverse bottle-necks in the operation of the network element.
In a controller system according to an exemplifying embodiment of the invention, the processing system 210 is adapted to configure a routing topology of the software-defined network so that more data traffic load is allocated to the network element if the resource consumption estimate of the look-up system manifests lower loading of the network element than if the resource consumption estimate of the look-up system manifests higher loading of the network element.
In a controller system according to an exemplifying embodiment of the invention, the mutually different look-up tables constitute a set of look-up tables with the aid of which alternative look-up systems having functionalities suitable for the network element are capable of being constructed and the processing system 210 is adapted to search, on the basis of the look-up table specific resource consumption estimates, such one of the alternative look-up systems which has, according to a pre-determined criterion, a sufficiently advantageous resource consumption estimate.
In a controller system according to an exemplifying embodiment of the invention, the processing system 210 is adapted to compute resource consumption estimates for two or more of the above-mentioned alternative look-up systems on the basis of the look-up table specific resource consumption estimates related to the look-up tables contained by the two or more alternative look-up systems and to select a particular one of the two or more alternative look-up systems which has, according to the pre-determined criterion, the most advantageous resource consumption estimate.
In a controller system according to an exemplifying embodiment of the invention, the processing system 210 is adapted to use a genetic algorithm implementing simulated evolution for searching a suitable one of the above-mentioned alternative look-up systems.
In a controller system according to an exemplifying embodiment of the invention, the processing system 210 is adapted to use the Dijkstra or A* optimization algorithm for searching a suitable one of the above-mentioned alternative look-up systems from a search network where the look-up tables are related to nodes of the search network and the look-up table specific resource consumption estimates represents costs related to the nodes or to connections between adjacent nodes. The optimal one of the alternative look-up systems corresponds to a minimum cost route through the search network so that look-up tables related to the the minimum cost route form a look-up system having the functionality suitable for the network element.
In a controller system according to an exemplifying embodiment of the invention, the processing system 210 is adapted to read the look-up table specific resource consumption estimates from one or more data frames received at the controller system.
In a controller system according to an exemplifying embodiment of the invention, the processing system 210 is adapted to control the controller system to send one or more first data frames to the network element and to read the look-up table specific resource consumption estimates from one or more second data frames received from the network element, the one or more first data frames containing descriptors of the look-up tables. Each of the descriptors of the look-up tables may comprise, for example, the maximum number of entries of the corresponding look-up table, the size of one or more look-up keys of the corresponding look-up table, allowed don't care key matching bit positions in the corresponding look-up table, definitions of tasks related to the entries of the corresponding look-up table, maximum look-up frequency of the corresponding look-up table, and/or the average time between successive updates of the corresponding look-up table.
In a controller system according to an exemplifying embodiment of the invention, the look-up table specific resource consumption estimates comprise numerical values indicating load factors of one or more components of the network element and the processing system 210 is adapted to compute, for each of the components of the network element, a sum of the numerical values related to that component of the network element. In this case, the sums related to the components of the network element can be deemed to constitute the resource consumption estimate of the look-up system.
The processing system 210 of the controller system shown in
The processing system 310 of the network element shown in
If the network element is not capable of running the look-up system under consideration, a new candidate for the look-up system has to be defined for the network element. In the exemplifying method illustrated in
A method according to an exemplifying embodiment of the invention comprises configuring a routing topology of the software-defined network so that more data traffic load is allocated to the network element if the resource consumption estimate of the look-up system manifests lower loading of the network element than if the resource consumption estimate of the look-up system manifests higher loading of the network element.
In a method according to an exemplifying embodiment of the invention, the mutually different look-up tables constitute a set of look-up tables with the aid of which alternative look-up systems having functionalities suitable for the network element are capable of being constructed. In this case, the method comprises searching, on the basis of the look-up table specific resource consumption estimates, such one of the alternative look-up systems which has, according to a pre-determined criterion, a sufficiently advantageous resource consumption estimate.
A method according to an exemplifying embodiment of the invention comprises computing resource consumption estimates for two or more of the alternative look-up systems on the basis of the look-up table specific resource consumption estimates related to the look-up tables contained by the two or more alternative look-up systems and selecting a particular one of the two or more alternative look-up systems which has, according to the pre-determined criterion, the most advantageous resource consumption estimate.
In a method according to an exemplifying embodiment of the invention, a suitable one of the alternative look-up systems is searched with a genetic algorithm implementing simulated evolution.
In a method according to an exemplifying embodiment of the invention, a suitable one of the alternative look-up systems is searched with the Dijkstra or A* optimization algorithm from a search network where the look-up tables are related to nodes of the search network and the look-up table specific resource consumption estimates represents costs related to the nodes or to connections between adjacent nodes. An optimal one of the alternative look-up systems corresponds to a minimum cost route through the search network so that look-up tables related to the minimum cost route form a look-up system having the functionality suitable for the network element.
A method according to an exemplifying embodiment of the invention comprises reading the look-up table specific resource consumption estimates from one or more data frames received at the controller system.
A method according to an exemplifying embodiment of the invention comprises sending one or more first data frames to the network element and reading the look-up table specific resource consumption estimates from one or more second data frames received from the network element, the one or more first data frames containing descriptors of the look-up tables. Each of the descriptors of the look-up tables may comprise, for example, the maximum number of entries of the corresponding look-up table, the size of one or more look-up keys of the corresponding look-up table, allowed don't care key matching bit positions in the corresponding look-up table, definitions of tasks related to the entries of the corresponding look-up table, maximum look-up frequency of the corresponding look-up table, and/or the average time between successive updates of the corresponding look-up table.
In a method according to an exemplifying embodiment of the invention, the look-up table specific resource consumption estimates comprise numerical values indicating load factors of one or more components of the network element and the method comprises computing, for each of the components of the network element, a sum of the numerical values related to that component of the network element. The sums related to the components of the network element can be deemed to constitute the resource consumption estimate of the look-up system.
A computer program according to an exemplifying embodiment of the invention for configuring a network element of a software-defined network comprises computer executable instructions for controlling a programmable processing system to carry out a method according to any of the above-described exemplifying embodiments of the invention.
A computer program according to an exemplifying embodiment of the invention comprises software modules for configuring a network element of a software-defined network. The software modules comprise computer executable instructions for controlling a programmable processing system to:
The software modules can be e.g. subroutines or functions implemented with a suitable programming language and with a compiler suitable for the programming language and the programmable processing system.
A computer program product according to an exemplifying embodiment of the invention comprises a computer readable medium, e.g. a compact disc (“CD”), encoded with a computer program according to an exemplifying embodiment of invention.
A signal according to an exemplifying embodiment of the invention is encoded to carry information defining a computer program according to an exemplifying embodiment of invention.
The specific examples provided in the description given above should not be construed as limiting the scope and/or the applicability of the appended claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 20135013 | Jan 2013 | FI | national |
