JP3782063B2

JP3782063B2 - PATH ALLOCATION METHOD, PROGRAM THEREOF, AND STORAGE MEDIUM CONTAINING THE PROGRAM

Info

Publication number: JP3782063B2
Application number: JP2003051369A
Authority: JP
Inventors: 健一松井; 俊之櫻井; 昌樹金田; 純一村山; 啓之石井
Original assignee: Nippon Telegraph and Telephone Corp
Current assignee: Nippon Telegraph and Telephone Corp
Priority date: 2003-02-27
Filing date: 2003-02-27
Publication date: 2006-06-07
Anticipated expiration: 2023-02-27
Also published as: JP2004260719A

Description

【０００１】
【発明の属する技術分野】
本発明は、ネットワーク内のフローを効率良く収容するパス割り当て方法に係り、特に、フォトニック転送技術によって構築された大規模ネットワークにおいて、効率的にパス配置を算出するのに最適なパス割り当て方法及びそのプログラム並びにそのプログラムを記録した記録媒体に関するものである。
【０００２】
【従来の技術】
従来、ネットワーク内のフローを収容するパス割り当て方法として、近似解法の一つである欲張り法を用いた例が考えられる。一般に、欲張り法は、目的関数への貢献度を示す局所的な評価値に基づいて、実行可能解を直接構成する方法である（非特許文献１，２を参照）。
【０００３】
パケットを中継する複数の転送装置から成るネットワークにおいて、このネットワーク内のフローを収容するパス割り当て方法に欲張り法を適用した場合の動作を以下に示す。すなわち、発生したフローを帯域（容量）によって降順に並べ、上位のフローから順にフローの帯域と転送装置間のパスの容量を比較する。フローの帯域がパスの容量以下である場合には、フローをそのパスへ割り当て、割り当てられたパスから順に転送装置のポートを割り当てることを繰り返す。そして、フローの帯域がパスの容量よりも大きい場合には、割り当てを終了する。
【０００４】
【非特許文献１】
柳浦、茨木著，「組合せ最適化問題に対するメタ戦略について」，電子情報通信学会論文誌，電子情報通信学会，２０００年１月２５日，ＶＯＬ．Ｊ８３−Ｄ−ＩＮＯ．１，ｐ．３−２５
【非特許文献２】
柳浦睦憲、“組合せ最適化の数理”、［online］、２００２年１２月、京都大学、［平成１５年２月１９日検索］、インターネット＜ＵＲＬ：http://www-or.amp.i.kyoto-u.ac.jp/members/yagiura/papers/surikagaku200212.pdf＞
【０００５】
【発明が解決しようとする課題】
しかし、従来の欲張り法を適用したパス割り当て方法は、以下の２つの問題があった。第１の問題は、最適なパスの割り当てができないことである。欲張り法は帯域の大きなフローから順にパスへ割り当てるため、割り当て対象となったフローの帯域がパスの容量よりも大きい場合は割り当てを終了する。この場合、割り当て対象となったフローの帯域よりも小さいフローの帯域がパスの容量と等しいまたは小さいときがあり、フローをパスへ割り当てることができる可能性を無視することになるからである。第２の問題は、パス割り当て方法に対して様々な解法を適用した場合には、その適応性を検証することが難しいことである。これは、パス割り当て方法は、欲張り法を前提としてその動作の設計を行っているため、他の解法をその設計に適用させるのが困難だからである。
【０００６】
本発明はこのような問題を考慮してなされたものであり、その目的は、最適な割り当てができない従来の欲張り法を適用したパス割り当て方法を改善し、欲張り法以外の様々な解法を適用可能なパス割り当て方法を提供することにある。
【０００７】
【課題を解決するための手段】
請求項１の発明は、有限個の出力ポートと入力ポートとを各々備えた有限個の転送装置と該転送装置に接続される経路制御装置とを備え、転送装置に備えた１つの出力ポートと他の転送装置に備えた１つの入力ポートとが接続され、前記出力ポートから入力ポートへフローを流すパスを有限個設定して構成されるネットワークにおいて、各々のパスは転送可能な有限の容量を有し、転送装置から他の転送装置へ流れる有限個のフローが存在し、各々のフローは有限の帯域を有し、フローの帯域がパスの容量以下の場合には前記フローがパスを流れ、フローの帯域がパスの容量を越える場合には前記フローがパスを流れず、複数のフローの組合せの帯域の合計値が１つのパスの容量以下の場合には前記複数のフローの組合せが１つのパスを流れ、複数のフローの組合せの帯域の合計値が１つのパスの容量を越える場合には前記複数のフローの組合せが１つのパスを流れない場合に、経路制御装置は、ネットワーク内のそれぞれの転送装置から他の転送装置へ流れるフローの帯域の合計値が前記転送装置の出力ポートから他の転送装置の入力ポートへ接続されているパスの容量の合計値を越えない条件と、ネットワーク内のそれぞれの転送装置から接続されているパスの合計数が前記転送装置に備えた出力ポートの数を超えない条件と、ネットワーク内のそれぞれの転送装置へ接続されているパスの合計数が前記転送装置に備えた入力ポートの数を超えない条件と、ネットワーク内の第１の転送装置から他の第２の転送装置へ向かって接続されているパスの合計数と前記他の第２の転送装置から第１の転送装置へ向かって接続されているパスの合計数とが等しい条件と、ネットワーク内のそれぞれの転送装置に備えた出力ポートの数と入力ポートの数とが等しい条件とをそれぞれ満たし、ネットワーク内の全てのパスを流れる全てのフローの帯域の合計値が最大となるように、単体法及び欲張り法を含む線形計画法毎に、パスを設定し、該設定したパスに割り当てられるフローの帯域を算出し、該フローの帯域の合計値を算出し、前記線形計画法毎に設定したパス、算出したフローの帯域及びその合計値をそれぞれ比較し、前記複数の線形計画法のうちの一つの線形計画法を選択し、該選択した線形計画法により転送装置とパスとの接続関係を決定することを特徴とするパス割り当て方法である。
【０００８】
請求項２の発明は、請求項１に記載のパス割り当て方法において、経路制御装置は、転送装置の数、転送装置が備える入出力ポート数、フローの数、転送装置間のフロー帯域、入出力ポートとパスとの関係及びパスとフローとの関係を有限集合として明示的に規定して、該有限集合の値を組合せて前記各条件を線形関数で表し、該線形関数に基づいて、ネットワーク内の全てのパスを流れる全てのフローの帯域の合計を算出し、転送装置とパスとの接続関係を求めることを特徴とするものである。
【０００９】
請求項３の発明は、請求項１に記載のパス割り当て方法において、経路制御装置は、ネットワーク内の全てのパスを流れる全てのフローの帯域の合計値が最大となるときの、転送装置間のパスを設定し、該設定したパスに割り当てられるフローの帯域を算出し、前記フローの帯域の合計値を算出することを特徴とするものである。
【００１０】
請求項４の発明は、有限個の出力ポートと入力ポートとを各々備えた有限個の転送装置と該転送装置に接続される経路制御装置とを備え、転送装置に備えた１つの出力ポートと他の転送装置に備えた１つの入力ポートとが接続され、前記出力ポートから入力ポートへフローを流すパスを有限個設定して構成されるネットワークにおける経路制御装置を構成するコンピュータに、各々のパスは転送可能な有限の容量を有し、転送装置から他の転送装置へ流れる有限個のフローが存在し、各々のフローは有限の帯域を有し、フローの帯域がパスの容量以下の場合には前記フローがパスを流れ、フローの帯域がパスの容量を越える場合には前記フローがパスを流れず、複数のフローの組合せの帯域の合計値が１つのパスの容量以下の場合には前記複数のフローの組合せが１つのパスを流れ、複数のフローの組合せの帯域の合計値が１つのパスの容量を越える場合には前記複数のフローの組合せが１つのパスを流れない場合に、ネットワーク内のそれぞれの転送装置から他の転送装置へ流れるフローの帯域の合計値が前記転送装置の出力ポートから他の転送装置の入力ポートへ接続されているパスの容量の合計値を越えない条件と、ネットワーク内のそれぞれの転送装置から接続されているパスの合計数が前記転送装置に備えた出力ポートの数を超えない条件と、ネットワーク内のそれぞれの転送装置へ接続されているパスの合計数が前記転送装置に備えた入力ポートの数を超えない条件と、ネットワーク内の第１の転送装置から他の第２の転送装置へ向かって接続されているパスの合計数と前記他の第２の転送装置から第１の転送装置へ向かって接続されているパスの合計数とが等しい条件と、ネットワーク内のそれぞれの転送装置に備えた出力ポートの数と入力ポートの数とが等しい条件とをそれぞれ満たす機能と、ネットワーク内の全てのパスを流れる全てのフローの帯域の合計値が最大となるように、単体法及び欲張り法を含む線形計画法毎に、パスを設定し、該設定したパスに割り当てられるフローの帯域を算出し、該フローの帯域の合計値を算出する機能と、前記線形計画法毎に設定したパス、算出したフローの帯域及びその合計値をそれぞれ比較し、前記複数の線形計画法のうちの一つの線形計画法を選択する機能と、該選択した線形計画法により転送装置とパスとの接続関係を決定する機能とを実現させることを特徴とするパス割り当てプログラムである。
【００１１】
請求項５の発明は、請求項４に記載のパス割り当てプログラムにおいて、経路制御装置を構成するコンピュータに、転送装置の数、転送装置が備える入出力ポート数、フローの数、転送装置間のフロー帯域、入出力ポートとパスとの関係及びパスとフローとの関係を有限集合として明示的に規定して、該有限集合の値を組合せて前記各条件を線形関数で表す機能と、該線形関数に基づいて、ネットワーク内の全てのパスを流れる全てのフローの帯域の合計を算出し、転送装置とパスとの接続関係を求める機能とを実現させることを特徴とするものである。
【００１２】
請求項６の発明は、請求項４に記載のパス割り当てプログラムにおいて、経路制御装置を構成するコンピュータに、ネットワーク内の全てのパスを流れる全てのフローの帯域の合計値が最大となるときの、転送装置間のパスを設定する機能と、該設定したパスに割り当てられるフローの帯域を算出する機能と、前記フローの帯域の合計値を算出する機能とを実現させることを特徴とするものである。
【００１３】
請求項７の発明は、請求項４から６のいずれか１項に記載のパス割り当てプログラムを記録した記録媒体である。
【００１４】
【発明の実施の形態】
以下、本発明の実施の形態について、図面及び数式を参照して詳細に説明する。数１は、本発明の実施の形態に基づいて、線形計画の枠組みに沿って、パス割り当て方法を定式化した数式である。
【数１】

【００１５】
数１は、その値を最大化することを目的とする目的関数１００と、当該目的関数１００を最大化する際の変数の範囲を制約するための制約式１０１〜１０６とから成る。ここで、変数は、ｎ台の転送装置からなる有限集合Ｎ、ｍｎ（ｎ−１）通りのフローからなる有限集合Ｍ（１台の転送装置あたりｍ通りのフローが発生するとする）、転送装置ｉ（ｉ∈Ｎ）から転送装置ｊ（ｊ∈Ｎ）へ向かうｋ番目（ｋ∈Ｎ）のフロー帯域ｆijk、転送装置ｉの出力ポート数ＩＦ＿ＯＵＴi、転送装置ｉの入力ポート数ＩＦ＿ＩＮi、パス容量ｗ、転送装置ｉから転送装置ｊへのパス数Ｘij、フローｆijkがパスを通るとき１、通らないとき０である０−１変数ｙijkに設計されている。
【００１６】
制約式１０１は、転送装置からパスへ出力するフロー帯域の合計が転送装置から出るパスの合計容量を越えないという制約条件を満たすために設計されている。制約式１０２は、転送装置ｉから出るパスの総数がＩＦ＿ＯＵＴiを超えないという制約条件を満たすために設計されている。制約式１０３は、転送装置ｉに入るパスの総数がＩＦ＿ＩＮiを超えないという制約条件を満たすために設計されている。
【００１７】
制約式１０４は、転送装置ｉから転送装置ｊへのパスと同数のパスが転送装置ｊから転送装置ｉへ向かうという制約条件を満たすために設計されている。制約式１０５は、ネットワーク内のそれぞれの転送装置に備えた出力ポートの数と入力ポートの数が等しいという制約条件を満たすために設計されている。制約式１０６は、０または１のみをとることができるという０−１変数ｙijkの性質を表すために設定されている。また、目的関数１００は、制約式１０１〜１０６を満たす条件の下で、パスを通るフロー帯域の合計を最大にするように、有限集合Ｍからフローを選択することを示すために設計されている。
【００１８】
次に、本発明の実施の形態の動作について、具体的な実施例を用いて説明する。図４は、本発明の実施の形態に係るパス割り当て方法を用いるネットワークの概略構成図である。本ネットワークは、互いに接続される転送装置１〜３、管理網４及び経路制御装置１０を備えている。図１に示すように、転送装置１は出力と入力とを兼ねたポート３０１〜３０３を備え、転送装置２は出力と入力とを兼ねたポート３１１〜３１３を備え、転送装置３は出力と入力とを兼ねたポート３２１〜３２３を備えている。また、経路制御装置１０は、管理網４を介して転送装置１〜３と接続され、フローの帯域及びパスの容量等を管理し、目的関数１００及び制約式１０１〜１０６により、以下に説明する手順に従って、パスの割り当てを行う。
【００１９】
表１は、各転送装置１〜３間に流れるフローの帯域の一例である。
【表１】

【００２０】
表１において、ｆijkは、転送装置ｉから転送装置ｊへ向かうｋ番目のフロー帯域を示している。例えば、ｆ121は転送装置１から転送装置２へ向かう１番目のフロー帯域を示しており、表１によりその値は９である。
【００２１】
図１に示したネットワークに数１を適用すると、目的関数１００は数２に示す数式に、制約式１０１〜１０６は数３〜８に示す数式に、それぞれ展開される。
【数２】

【数３】

【数４】

【数５】

【数６】

【数７】

【数８】

【００２２】
数１〜８及び表１を基に、線形計画の解法の一つである単体法を適用すると、図２に示すように転送装置１〜３間にパスが設定され、各パスに割り当てられるフローの帯域は表２に示すようになり、目的関数値であるフロー帯域の合計は３８となる。
【表２】

【００２３】
一方、数１〜８及び表１を基に、線形計画の解法の一つである欲張り法を適用すると、図３に示すように転送装置１〜３間にパスが設定され、各パスに割り当てられるフローの帯域は表３に示すようになり、目的関数値であるフロー帯域の合計は３０となる。さらに、同様の手順で、数２〜８及び表１を基に、任意の線形計画の解法を適用し、パス配置及びフロー帯域並びに目標関数値を求めることができる。
【表３】

【００２４】
また、同様の手順で、任意のネットワーク形態におけるパス割り当て方法を、線形計画の枠組みに沿って、例えば、数２〜８及び表１に示したように展開し、全ての線形計画の解法を適用し、パス配置及びフローの帯域並びに目標関数値を比較評価することができる。これにより、ネットワーク形態に適した解法を選択することができ、パス割り当ての最適性を改善することができる。
【００２５】
このように、経路制御装置１０は、転送装置１〜３のフローの帯域及びパスの容量等を管理することにより、表１に示した転送装置１〜３間に流れるフロー帯域のテーブルを生成し、数１に示した数式から数２〜８に示した数式に展開する。そして、単体法や欲張り法等の線形計画の解法を用いて、表２や表３に示したような、転送装置１〜３間のパスを設定し、設定したパスに割り当てられるフローの帯域を算出し、目的関数値であるフローの帯域の合計値を算出する。
【００２６】
次に、図５を参照して、経路制御装置１０の内部構造を説明する。経路制御装置１０は、オペレータＩＦ１１、パス割り当て機能１２、フロー帯域／パス容量ファイル１３及びパス転送テーブル設定ＩＦ１４を有する。また、パス割り当て機能１２は、フロー帯域／パス容量管理機能２１、フロー帯域テーブル生成機能２２、線形計画数式展開／実行機能２３及びパス配置等算出機能２４を有する。
【００２７】
オペレータＩＦ１１は、オペレータが、図示しないキーボードや表示装置等を用いて、転送装置の数、転送装置が備える入力ポート数、出力ポート数等の必要な変数、目的関数１００、制約式１０１〜１０６、線形計画の解法等を入力すると、これらの情報をパス割り当て機能１２に受け渡す。パス割り当て機能１２のフロー帯域／パス容量管理機能２１は、転送装置１〜３とそれぞれ通信を行い、フロー帯域やパスの容量等の情報を受信し、これらの情報をフロー帯域／パス容量ファイル１３に格納して管理する。フロー帯域テーブル生成機能２２は、フロー帯域／パス容量ファイル１３に格納された情報に基づいて、表１に示したような転送装置１〜３間に流れるフローの帯域のテーブルを生成する。線形計画数式展開／実行機能２３は、線形計画の解法を用いるために、数２〜８に示した数式のように、前記変数を目的関数１００や制約式の数式に当てはめて展開するとともに、前記数式に従って様々な線形計画の解法を実行する。パス配置等算出機能２４は、前記解法によってパスを設定し、設定したパスに割り当てられるフローの帯域を算出し、目的関数値を算出することにより、表２または表３に示したパス転送テーブルを生成する。パス転送テーブル設定ＩＦ１４は、パス配置等算出機能２４が生成したパス転送テーブルのパス配置等の情報を管理網４を介して転送装置１〜３に送信する。
【００２８】
実際には、上記のそれぞれの機能は、ソフトウェアを組み込んだコンピュータによって実現される。また、このソフトウェアは、コンピュータに実行させることのできるプログラムとして磁気ディスク、光ディスク、半導体メモリ等の記録媒体に格納して頒布することもできる。
【００２９】
【発明の効果】
以上説明したように、本発明によれば、経路制御装置が、任意のネットワーク形態におけるパス割り当て方法を、線形計画の枠組みに沿って、例えば数２〜８及び表１に示したように展開するようにした。このため、様々な線形計画の解法を適用して、パス配置及びフローの帯域並びに目的関数値を求め、これらを比較評価することができる。これにより、ネットワーク形態に適した解法を選択することによって、パス割り当ての最適性を改善することができる。
【図面の簡単な説明】
【図１】転送装置のポートを説明する概略構成図である。
【図２】単体法を適用した場合におけるパスが設定された転送装置の概略構成図である。
【図３】欲張り法を適用した場合におけるパスが設定された転送装置の概略構成図である。
【図４】本発明の実施の形態に係るパス割り当て方法を用いるネットワークの概略構成図である。
【図５】経路制御装置の内部機能を説明する内部構造図である。
【符号の説明】
１〜３転送装置
４管理網
１０経路制御装置
１１オペレータＩＦ
１２パス割り当て機能
１３フロー帯域／パス容量ファイル
１４パス転送テーブル設定ＩＦ
２１フロー帯域／パス容量管理機能
２２フロー帯域テーブル生成機能
２３線形計画数式展開／実行機能
２４パス配置等算出機能
３０１〜３０３，３１１〜３１３，３２１〜３２３ポート
４００〜４０６パス[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a path allocation method for efficiently accommodating the flow in the network, in particular, in a large-scale network built by photonic transfer technology, optimal path allocation method to calculate efficiently pass arrangement and The present invention relates to the program and a recording medium on which the program is recorded.
[0002]
[Prior art]
Conventionally, an example using a greedy method, which is one of approximate solutions, can be considered as a path allocation method for accommodating flows in a network. In general, the greedy method is a method of directly constructing an executable solution based on a local evaluation value indicating the degree of contribution to the objective function (see Non-Patent Documents 1 and 2).
[0003]
An operation when the greedy method is applied to a path allocation method for accommodating flows in the network in a network composed of a plurality of transfer devices that relay packets will be described below. That is, the generated flows are arranged in descending order according to the bandwidth (capacity), and the bandwidth of the flow and the capacity of the path between the transfer devices are compared in order from the upper flow. If the bandwidth of the flow is less than or equal to the capacity of the path, the flow is assigned to that path, and the port of the transfer device is assigned in order from the assigned path. If the flow bandwidth is larger than the path capacity, the allocation is terminated.
[0004]
[Non-Patent Document 1]
Yanagiura and Ibaraki, “Meta-strategy for combinatorial optimization problems”, IEICE Transactions, IEICE, 25 January 2000, VOL. J83-DI NO. 1, p. 3-25
[Non-Patent Document 2]
Yasunori Yanagiura, "Mathematics of combinatorial optimization", [online], December 2002, Kyoto University, [searched February 19, 2003], Internet <URL: http: //www-or.amp.i. kyoto-u.ac.jp/members/yagiura/papers/surikagaku200212.pdf>
[0005]
[Problems to be solved by the invention]
However, the path allocation method to which the conventional greedy method is applied has the following two problems. The first problem is that an optimal path cannot be allocated. Since the greedy method assigns to the path in order from the flow with the largest bandwidth, the assignment is terminated when the bandwidth of the flow to be assigned is larger than the capacity of the path. In this case, there is a case where the bandwidth of the flow smaller than the bandwidth of the flow to be assigned is equal to or smaller than the capacity of the path, and the possibility that the flow can be assigned to the path is ignored. The second problem is that it is difficult to verify the adaptability when various solutions are applied to the path allocation method. This is because the path assignment method is designed for operation based on the greedy method, and it is difficult to apply other solutions to the design.
[0006]
The present invention has been made in consideration of such problems, and its purpose is to improve the path allocation method using the conventional greedy method, which cannot be optimally allocated, and to apply various solutions other than the greedy method. Is to provide a simple path allocation method.
[0007]
[Means for Solving the Problems]
The invention of claim 1 comprises a finite number of transfer devices each having a finite number of output ports and input ports, and a path control device connected to the transfer device, and one output port provided in the transfer device; In a network configured by connecting a single input port provided in another transfer device and setting a finite number of paths through which flows flow from the output port to the input port, each path has a finite capacity that can be transferred. And there is a finite number of flows that flow from a transfer device to another transfer device, each flow has a finite bandwidth, and if the flow bandwidth is less than or equal to the capacity of the path, the flow flows through the path, When the flow bandwidth exceeds the capacity of the path, the flow does not flow through the path, and when the total bandwidth of the combination of the plurality of flows is equal to or less than the capacity of one path, the combination of the plurality of flows is one. Flow path When the total value of the bandwidths of a plurality of flow combinations exceeds the capacity of a single path, the routing control device receives each transfer device from the network when the combination of the plurality of flows does not flow through one path. The condition that the total value of the bandwidth of the flow flowing to the other transfer device does not exceed the total value of the capacity of the path connected from the output port of the transfer device to the input port of the other transfer device, and each transfer in the network The total number of paths connected from the device does not exceed the number of output ports provided in the transfer device, and the total number of paths connected to each transfer device in the network is provided in the transfer device. Conditions that do not exceed the number of input ports, the total number of paths connected from the first transfer device to the other second transfer device in the network, and the other second transfer A condition in which the total number of paths connected from the device to the first transfer device is equal, and a condition in which the number of output ports and the number of input ports provided in each transfer device in the network are equal to each other. A path is set for each linear programming method including the simplex method and the greedy method so that the total value of the bandwidths of all the flows flowing through all the paths in the network is maximized and assigned to the set path. Calculating a flow bandwidth, calculating a total value of the bandwidth of the flow, comparing the path set for each linear programming method, the calculated bandwidth of the flow and the total value thereof, of the plurality of linear programming methods; Is selected, and the connection relation between the transfer device and the path is determined by the selected linear programming .
[0008]
According to a second aspect of the present invention, in the path allocation method according to the first aspect, the route control device includes: the number of transfer devices, the number of input / output ports included in the transfer device, the number of flows, the flow bandwidth between the transfer devices, and the input / output The relationship between the port and the path and the relationship between the path and the flow are explicitly specified as a finite set, and the values of the finite set are combined to express each of the conditions as a linear function. Based on the linear function, The total of the bandwidths of all the flows flowing through all the paths is calculated, and the connection relationship between the transfer device and the path is obtained.
[0009]
According to a third aspect of the present invention, in the path allocation method according to the first aspect, the path control apparatus is configured to transfer the transfer between the transfer apparatuses when the total value of the bandwidths of all the flows flowing through all the paths in the network is the maximum. A path is set, a bandwidth of a flow assigned to the set path is calculated, and a total value of the bandwidth of the flow is calculated.
[0010]
The invention of claim 4 comprises a finite number of transfer devices each having a finite number of output ports and input ports, and a path control device connected to the transfer device, and one output port provided in the transfer device, Each path is connected to a computer that constitutes a path control device in a network that is connected to one input port provided in another transfer device and configured by setting a finite number of paths through which flows flow from the output port to the input port. Has a finite capacity that can be transferred, there is a finite number of flows that flow from the transfer device to other transfer devices, each flow has a finite bandwidth, and the flow bandwidth is less than the capacity of the path If the flow flows through the path and the bandwidth of the flow exceeds the capacity of the path, the flow does not flow through the path, and if the total value of the bands of the combination of multiple flows is less than the capacity of one path, the flow Duplicate If a combination of flows flows through one path and the total bandwidth of the combinations of flows exceeds the capacity of one path, the combination of the flows does not flow through one path. A condition in which the total value of the bandwidths of the flows flowing from the respective transfer devices to the other transfer devices does not exceed the total value of the capacities of the paths connected from the output ports of the transfer devices to the input ports of the other transfer devices; The condition that the total number of paths connected from each transfer device in the network does not exceed the number of output ports provided in the transfer device, and the total number of paths connected to each transfer device in the network are: The total number of paths connected from the first transfer device to the other second transfer device in the network under conditions that do not exceed the number of input ports provided in the transfer device The condition that the total number of paths connected from the other second transfer device toward the first transfer device is equal, and the number of output ports and the number of input ports provided in each transfer device in the network A path is set for each linear programming method including the simplex method and the greedy method so that the total value of the functions that satisfy the same conditions and the total bandwidth of all flows that flow through all the paths in the network is maximized. Calculating the bandwidth of the flow assigned to the set path, calculating the total value of the bandwidth of the flow, the path set for each linear programming method, the calculated bandwidth of the flow and the total value thereof, respectively. And a function of selecting one of the plurality of linear programming methods and a function of determining a connection relation between the transfer device and the path by the selected linear programming method. This is a path allocation program.
[0011]
According to a fifth aspect of the present invention, in the path allocation program according to the fourth aspect, the number of transfer devices, the number of input / output ports provided in the transfer device, the number of flows, and the flow between transfer devices are included in the computer constituting the route control device. A bandwidth, a relationship between an input / output port and a path, and a relationship between a path and a flow are explicitly defined as a finite set, and each linear condition is expressed by combining the values of the finite set, and the linear function Based on the above, it is possible to realize a function of calculating a sum of bandwidths of all flows flowing through all paths in the network and obtaining a connection relationship between the transfer apparatus and the paths.
[0012]
The invention of claim 6 is the path allocation program according to claim 4, wherein when the total value of the bandwidths of all the flows flowing through all the paths in the network is maximized in the computer constituting the route control device, A function of setting a path between transfer devices, a function of calculating a bandwidth of a flow allocated to the set path, and a function of calculating a total value of the bandwidth of the flow are realized. .
[0013]
A seventh aspect of the invention is a recording medium on which the path assignment program according to any one of the fourth to sixth aspects is recorded.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings and mathematical expressions. Formula 1 is a mathematical formula that formulates a path allocation method along the framework of the linear programming based on the embodiment of the present invention.
[Expression 1]

[0015]
Equation 1 includes an objective function 100 for maximizing the value and constraint equations 101 to 106 for constraining the range of variables when the objective function 100 is maximized. Here, the variables are a finite set N consisting of n transfer devices, a finite set M consisting of mn (n-1) flows (assuming that m flows are generated per transfer device), and transfer devices. The k-th (kεN) flow band fijk from i (iεN) to the transfer device j (jεN), the number of output ports IF_OUTi of the transfer device i, the number of input ports IF_INi of the transfer device i, and the path capacity w The number of paths Xij from the transfer apparatus i to the transfer apparatus j is designed to be 0-1 variable yijk which is 1 when the flow fijk passes the path and 0 when the flow fijk does not pass.
[0016]
The constraint equation 101 is designed to satisfy the constraint condition that the total flow bandwidth output from the transfer apparatus to the path does not exceed the total capacity of the paths output from the transfer apparatus. The constraint equation 102 is designed to satisfy a constraint condition that the total number of paths from the transfer device i does not exceed IF_OUTi. The constraint equation 103 is designed to satisfy the constraint condition that the total number of paths entering the transfer device i does not exceed IF_INi.
[0017]
The constraint equation 104 is designed to satisfy the constraint condition that the same number of paths from the transfer device i to the transfer device j are directed from the transfer device j to the transfer device i. The constraint equation 105 is designed to satisfy the constraint condition that the number of output ports and the number of input ports provided in each transfer device in the network are equal. The constraint equation 106 is set to represent the property of the 0-1 variable yijk that can take only 0 or 1. The objective function 100 is also designed to show that a flow is selected from the finite set M so as to maximize the total flow bandwidth passing through the path under the conditions satisfying the constraint equations 101-106. .
[0018]
Next, the operation of the embodiment of the present invention will be described using specific examples. FIG. 4 is a schematic configuration diagram of a network using the path allocation method according to the embodiment of the present invention. This network includes transfer devices 1 to 3, a management network 4, and a route control device 10 that are connected to each other. As shown in FIG. 1, the transfer device 1 includes ports 301 to 303 that serve as outputs and inputs, the transfer device 2 includes ports 311 to 313 that serve as outputs and inputs, and the transfer device 3 includes outputs and inputs. Ports 321 to 323 that also serve as the above. The path control device 10 is connected to the transfer devices 1 to 3 via the management network 4 and manages the flow bandwidth, path capacity, and the like, and will be described below using the objective function 100 and the constraint equations 101 to 106. Assign a path according to the procedure.
[0019]
Table 1 shows an example of the bandwidth of the flow that flows between the transfer apparatuses 1 to 3.
[Table 1]

[0020]
In Table 1, fijk represents the k-th flow band from the transfer apparatus i to the transfer apparatus j. For example, f121 indicates the first flow band from the transfer apparatus 1 to the transfer apparatus 2, and its value is 9 according to Table 1.
[0021]
When Expression 1 is applied to the network shown in FIG. 1, the objective function 100 is expanded into the mathematical expression shown in Expression 2, and the constraint expressions 101 to 106 are expanded into the mathematical expressions shown in Expressions 3 to 8, respectively.
[Expression 2]

[Equation 3]

[Expression 4]

[Equation 5]

[Formula 6]

[Expression 7]

[Equation 8]

[0022]
When a simplex method, which is one of linear programming methods, is applied based on Equations 1 to 8 and Table 1, a path is set between the transfer apparatuses 1 to 3 as shown in FIG. The total bandwidth is 38 as the objective function value.
[Table 2]

[0023]
On the other hand, when the greedy method, which is one of the linear programming methods, is applied based on Equations 1 to 8 and Table 1, paths are set between the transfer apparatuses 1 to 3 as shown in FIG. 3, and assigned to each path. The bandwidth of the flow to be obtained is as shown in Table 3, and the sum of the flow bandwidths as the objective function value is 30. Furthermore, in the same procedure, based on Equations 2 to 8 and Table 1, an arbitrary linear programming method can be applied to obtain a path arrangement, a flow band, and a target function value.
[Table 3]

[0024]
Further, in the same procedure, the path allocation method in an arbitrary network form is expanded as shown in Equations 2 to 8 and Table 1 along the linear programming framework, and all linear programming solutions are applied. In addition, the path arrangement and the flow bandwidth and the target function value can be compared and evaluated. As a result, a solution suitable for the network form can be selected, and the optimality of path assignment can be improved.
[0025]
In this way, the path control device 10 generates a flow bandwidth table between the transfer devices 1 to 3 shown in Table 1 by managing the flow bandwidth and path capacity of the transfer devices 1 to 3. , The formula shown in Formula 1 is expanded into formulas shown in Formulas 2-8. Then, using a linear programming solution such as the simplex method or the greedy method, the paths between the transfer apparatuses 1 to 3 as shown in Table 2 and Table 3 are set, and the bandwidth of the flow allocated to the set path is set. Calculate the total value of the flow bandwidth that is the objective function value.
[0026]
Next, the internal structure of the path control device 10 will be described with reference to FIG. The path control device 10 includes an operator IF 11, a path assignment function 12, a flow bandwidth / path capacity file 13, and a path transfer table setting IF 14. The path allocation function 12 includes a flow bandwidth / path capacity management function 21, a flow bandwidth table generation function 22, a linear program expression development / execution function 23, and a path arrangement calculation function 24.
[0027]
The operator IF 11 uses an unillustrated keyboard, display device, or the like to allow the operator to use necessary variables such as the number of transfer devices, the number of input ports provided in the transfer device, the number of output ports, the objective function 100, constraint equations 101 to 106, When a linear programming solution or the like is input, the information is transferred to the path assignment function 12. The flow band / path capacity management function 21 of the path allocation function 12 communicates with each of the transfer apparatuses 1 to 3, receives information such as the flow band and path capacity, and stores these information in the flow band / path capacity file 13. To store and manage. The flow bandwidth table generation function 22 generates a bandwidth table of flows flowing between the transfer apparatuses 1 to 3 as shown in Table 1 based on the information stored in the flow bandwidth / path capacity file 13. The linear programming formula expansion / execution function 23 applies the variables to the objective function 100 and the formulas of the constraint formulas and expands the variables as shown in formulas 2 to 8 in order to use the linear programming solution. Perform various linear programming solutions according to mathematical formulas. The path arrangement etc. calculation function 24 sets the path by the above solution, calculates the bandwidth of the flow allocated to the set path, and calculates the objective function value, thereby obtaining the path transfer table shown in Table 2 or Table 3. Generate. The path transfer table setting IF 14 transmits information such as the path arrangement of the path transfer table generated by the path arrangement etc. calculation function 24 to the transfer apparatuses 1 to 3 via the management network 4.
[0028]
Actually, each of the above functions is realized by a computer incorporating software. The software can also be stored and distributed in a recording medium such as a magnetic disk, an optical disk, or a semiconductor memory as a program that can be executed by a computer.
[0029]
【The invention's effect】
As described above, according to the present invention, the path control device develops a path allocation method in an arbitrary network form as shown in, for example, Equations 2 to 8 and Table 1 along the framework of the linear programming. I did it. For this reason, various linear programming solutions can be applied to obtain the path arrangement, the flow bandwidth, and the objective function value, which can be compared and evaluated. Thereby, the optimality of path allocation can be improved by selecting a solution suitable for the network form.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram illustrating a port of a transfer device.
FIG. 2 is a schematic configuration diagram of a transfer apparatus in which a path is set when the simplex method is applied.
FIG. 3 is a schematic configuration diagram of a transfer apparatus in which a path is set when the greedy method is applied.
FIG. 4 is a schematic configuration diagram of a network using a path allocation method according to an embodiment of the present invention.
FIG. 5 is an internal structure diagram illustrating an internal function of the route control device.
[Explanation of symbols]
1-3 Transfer device 4 Management network 10 Route control device 11 Operator IF
12 Path allocation function 13 Flow bandwidth / path capacity file 14 Path transfer table setting IF
21 Flow Band / Path Capacity Management Function 22 Flow Band Table Generation Function 23 Linear Programming Formula Expansion / Execution Function 24 Path Arrangement Calculation Function 301-303, 311-313, 321-323 Port 400-406 Path

Claims

A finite number of transfer devices each having a finite number of output ports and input ports and a path control device connected to the transfer device are provided. One output port provided in the transfer device and another transfer device are provided. In a network configured by connecting a single input port and setting a finite number of paths through which flows flow from the output port to the input port,
Each path has a finite capacity that can be transferred, there is a finite number of flows that flow from the transfer device to other transfer devices, each flow has a finite bandwidth, and the flow bandwidth is less than the capacity of the path If the flow flows through the path, and the flow bandwidth exceeds the path capacity, the flow does not flow through the path, and the total bandwidth of the combination of multiple flows is less than the capacity of one path When the combination of the plurality of flows flows through one path, and the total bandwidth of the combination of the plurality of flows exceeds the capacity of one path, the combination of the plurality of flows does not flow through one path In addition,
The path control device is a total of the capacities of paths in which the total value of the flow flows from each transfer device to another transfer device in the network is connected from the output port of the transfer device to the input port of the other transfer device. A condition that does not exceed the value, a condition that the total number of paths connected from each transfer device in the network does not exceed the number of output ports provided in the transfer device, and a connection to each transfer device in the network The total number of paths connected to the second transfer device from the first transfer device in the network under the condition that the total number of existing paths does not exceed the number of input ports provided in the transfer device And the total number of paths connected from the other second transfer device to the first transfer device, and the output ports provided in each transfer device in the network. Satisfies the number of the number of bets and the input port and the same conditions, respectively, so that the total value of the bandwidth of all flows through all the paths in the network is maximized, linear programming, including a simplex method and greedy Each time, a path is set, the bandwidth of the flow allocated to the set path is calculated, the total value of the bandwidth of the flow is calculated,
Each of the paths set for each linear programming method, the calculated flow bandwidth and the total value thereof are compared, and one linear programming method is selected from the plurality of linear programming methods,
A path assignment method, wherein a connection relation between a transfer device and a path is determined by the selected linear programming .

2. The path allocation method according to claim 1, wherein the path control device includes: the number of transfer devices; the number of input / output ports included in the transfer device; the number of flows; the flow bandwidth between transfer devices; the relationship between input / output ports and paths; The relationship between paths and flows is explicitly specified as a finite set, and the values of the finite set are combined to express each condition as a linear function, and all the paths that flow through all the paths in the network based on the linear function A path allocation method comprising: calculating a total of the bandwidths of the flows and obtaining a connection relationship between the transfer apparatus and the path.

2. The path allocation method according to claim 1, wherein the path control device sets a path between transfer devices when the total value of the bandwidths of all flows flowing through all the paths in the network is maximized, and the setting is performed. A path allocation method characterized by calculating a bandwidth of a flow allocated to the path and calculating a total value of the bandwidth of the flow.

A finite number of transfer devices each having a finite number of output ports and input ports and a path control device connected to the transfer device are provided. One output port provided in the transfer device and another transfer device are provided. A computer that constitutes a path control device in a network that is connected to one input port and that is configured by setting a finite number of paths that flow from the output port to the input port.
Each path has a finite capacity that can be transferred, there is a finite number of flows that flow from the transfer device to other transfer devices, each flow has a finite bandwidth, and the flow bandwidth is less than the capacity of the path If the flow flows through the path, and the flow bandwidth exceeds the path capacity, the flow does not flow through the path, and the total bandwidth of the combination of multiple flows is less than the capacity of one path When the combination of the plurality of flows flows through one path, and the total bandwidth of the combination of the plurality of flows exceeds the capacity of one path, the combination of the plurality of flows does not flow through one path In addition,
A condition that the total value of the bandwidth of the flow flowing from each transfer device in the network to another transfer device does not exceed the total value of the capacity of the path connected from the output port of the transfer device to the input port of the other transfer device And the condition that the total number of paths connected from each transfer device in the network does not exceed the number of output ports provided in the transfer device, and the total number of paths connected to each transfer device in the network The number of input ports provided in the transfer device does not exceed the number of input ports, the total number of paths connected from the first transfer device to another second transfer device in the network, and the other The total number of paths connected from the two transfer devices to the first transfer device, and the number of output ports and input ports provided in each transfer device in the network. And functions to meet Doo number are equal conditions and, respectively,
A path is set for each linear programming method including the simplex method and the greedy method so that the sum of the bandwidths of all the flows flowing through all the paths in the network is maximized . A function of calculating a bandwidth and calculating a total value of the bandwidth of the flow;
A function for comparing each of the paths set for each of the linear programming methods, the calculated bandwidth of the flow and the total value thereof, and selecting one of the plurality of linear programming methods;
A path allocation program that realizes a function of determining a connection relationship between a transfer apparatus and a path by the selected linear programming.

5. The path allocation program according to claim 4, wherein the number of transfer devices, the number of input / output ports included in the transfer device, the number of flows, the flow bandwidth between the transfer devices, the input / output ports and the paths are included in the computer constituting the route control device. And the relationship between the path and the flow are explicitly specified as a finite set, and the function of expressing each condition as a linear function by combining the values of the finite set, and based on the linear function, A path allocation program characterized by realizing a function of calculating a total bandwidth of all flows flowing through all paths and obtaining a connection relationship between the transfer apparatus and the paths.

5. The path allocation program according to claim 4, wherein a path between transfer apparatuses is set in a computer constituting the path control apparatus when a total value of bandwidths of all flows flowing through all paths in the network is maximized. And a function for calculating a bandwidth of a flow allocated to the set path, and a function for calculating a total value of the bandwidth of the flow.

The recording medium which recorded the path allocation program of any one of Claim 4 to 6.