JP2004140541A - Communication system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a communication system for dynamically reassigning a value of a transmission band exclusively used by each application on a network depending on a band in use by each application when part of paths on the network becomes not available. <P>SOLUTION: The communication system wherein a transmission band on a path of network existing between a server and clients is assigned to an application provided from the server to the clients, is configured such that the transmission band of the application to which the transmission band has been assigned before occurrence of a fault on a path having no fault when a new transmission band on the path having no fault is newly assigned to the application to which the transmission band has been assigned on a path whose communication is disabled due to the occurrence of the fault. <P>COPYRIGHT: (C)2004,JPO

Description

【００３５】
経路Ｂには、ＡＰ＃Ｍ＋１〜ＡＰ＃ＮまでのＮ−Ｍ個のアプリケーションが事前に割り当てられているものとする。
【００３６】
また、ｌ番目のアプリケーションＡＰ＃ｌには、ｚ_ｌ＿ｉの帯域が事前に割り当てられているものとし、ある時刻ｔでの使用帯域をｘ_ｌ（ｔ）とする。
【００３７】
また、経路Ｂの全体の帯域Ｔ_Ｂのうち、ＡＰ＃Ｍ＋１〜ＡＰ＃Ｎに割当てられていない帯域をＷ_Ｂ＿ｉとする。この帯域Ｗ_Ｂ＿ｉは通常、ＡＰ＃Ｍ＋１〜ＡＰ＃Ｎが使用することはできない。
【００３８】
そして、ＡＰ＃Ｍ＋１〜ＡＰ＃Ｎの事前割当帯域の合計をＺ_Ｂ＿ｉとする。
【００３９】
したがって、以下の式が成り立つ。
【数２】

【００４０】
［▲２▼ネットワーク障害の発生］
次に、経路Ａにネットワーク障害が発生し、通信が不能になったとする。このとき、アドミッション制御装置は、経路Ａの伝送帯域を専有して使用できるように設定されていたアプリケーション（ＡＰ＃１〜ＡＰ＃Ｍ）が、経路Ｂの伝送帯域を専有して使用できるように設定しなおすものとする。
したがって、ＡＰ＃１〜ＡＰ＃Ｍにとっては、経路Ｂが迂回経路となる。
【００４１】
［▲３▼各アプリケーションに対する最終割当帯域の算出］
経路Ａで障害が発生した時点で、経路Ｂで提供されていたアプリケーション（ＡＰ＃Ｍ＋１〜ＡＰ＃Ｎ）の帯域値を確保しつつ、経路Ａで提供されていたアプリケーション（ＡＰ＃１〜ＡＰ＃Ｍ）に対して伝送帯域が逐次割り当てられていく。
【００４２】
経路Ａと経路ＢのアプリケーションＡＰ＃１〜ＡＰ＃Ｎで経路Ｂの帯域Ｔ_Ｂ＋αＷ_Ｂを比例配分する（ここでαは、０　＜　α≦　１．０。　）。各々のアプリケーション（ＡＰ＃１〜ＡＰ＃Ｎ）の最終の割当帯域は、以下のように表される。
【数３】

【数４】

【００４３】
［▲４▼割当可能帯域の検証］
ここで、障害発生前から経路Ｂを使用していたＡＰ＃ｌ（Ｍ＋１≦ｌ≦Ｎ）については使用中の帯域が維持される。これらＡＰ＃ｌ（Ｍ＋１≦ｌ≦Ｎ）に割り当てた伝送帯域は、サービス終了時（たとえば、ＴＶ電話であれば、回線断後）に変動される。したがって、提供中のサービスの品質は保証される。
【００４４】
そして、経路Ａに障害が発生した時刻ｔ_０にＡＰ＃Ｍ＋１〜ＡＰ＃Ｎの帯域使用状況を検知し、空き帯域をＡＰ＃１〜ＡＰ＃Ｍに割り当てる。
経路ＢでＡＰ＃ｌ［Ｍ＋１≦ｌ≦Ｎ］が使用中の伝送帯域（使用帯域）ｘ_ｌ（ｔ）は以下のいずれかに該当する。
（１）　使用帯域ｘ_ｌ（ｔ_０）が再配分後の最終割当帯域ｚ_ｌ＿ｆ　以下の場合、すなわち、
【数５】

の場合、ＡＰ＃ｌの時刻ｔ_０の割当帯域ｚ_ｌ（ｔ_０）をｚ_ｌ＿ｆとする。
（２）　使用帯域ｘ_ｌ（ｔ_０）が再配分後の最終割当帯域ｚ_ｌ＿ｆ　より大きい場合、
【数６】

の場合、ＡＰ＃ｌの時刻ｔ_０の割当帯域ｚ_ｌ（ｔ_０）をｘ_ｌ（ｔ_０）とする。
【００４５】
各ＡＰ＃ｌ［Ｍ＋１≦ｌ≦Ｎ］に対して、（１）〜（２）のいずれに該当するか判別して、ＡＰ＃Ｍ＋１〜ＡＰ＃Ｎで使用していない帯域のうち、新たにＡＰ＃１〜ＡＰ＃Ｍで使用できる帯域を見積もる。
【００４６】
ここで（２）より、
【数７】

が、ＡＰ＃ｌ［Ｍ＋１≦ｌ≦Ｎ］が最終割当帯域を超過して使用している帯域である。
【００４７】
また、ＡＰ＃１〜ＡＰ＃Ｎに対して最終割当帯域Ｚ_Ａ＿ｆの合計は、
【数８】

【００４８】
となるが、ＡＰ＃１〜ＡＰ＃Ｎの時刻ｔ_０における割当帯域ｚ_ｋ（ｔ_０）は、
【数９】

となる。
【００４９】
［▲５▼帯域割当の継続］
次に、上記▲４▼の手順を、ある時間ごとに繰り返す。ここでは説明の簡略化のため、一定時間間隔Δｔ毎に繰り返すものとする。すなわち、以下のような手順となる。
【００５０】
時刻ｔ_０＋ｎΔｔ　（ｎは正の整数）において、経路Ｂで使用中のＡＰ＃ｌ［Ｍ＋１≦ｌ≦Ｎ］の使用帯域ｘ_ｌ（ｔ_０＋ｎΔｔ　）は以下のいずれかに該当する。
【００５１】
（３）　使用帯域ｘ_ｌ（ｔ_０＋ｎΔｔ）が再配分後の最終割当帯域ｚ_ｌ＿ｆ以下の場合、すなわち、
【数１０】

の場合、ＡＰ＃ｌの時刻ｔ_０＋ｎΔｔの割当帯域ｚ_ｌ（ｔ_０＋ｎΔｔ　）をｚ_ｌ＿ｆとする。
（４）　使用帯域ｘ_ｌ（ｔ_０＋ｎΔｔ　）が再配分後の最終割当帯域ｚ_ｌ＿ｆより大きい場合、
【数１１】

の場合、ＡＰ＃ｌの時刻ｔ_０＋ｎΔｔの割当帯域ｚ_ｌ（ｔ_０＋ｎΔｔ）をその時点での割当帯域、すなわち、ｘ_ｌ（ｔ_０＋ｎΔｔ）とする。
【００５２】
各ＡＰ＃ｌ［Ｍ＋１≦ｌ≦Ｎ］に対して、（３）〜（４）のいずれに該当するか判別して、ＡＰ＃Ｍ＋１〜ＡＰ＃Ｎで使用されていない帯域の内、新たにＡＰ＃１〜ＡＰ＃Ｍで使用できる割当帯域を算出する。
【００５３】
ここで（２）より、
【数１２】

が、ＡＰ＃ｌ［Ｍ＋１≦ｌ≦Ｎ］が最終割当帯域を超過して使用している帯域である。
【００５４】
また、ＡＰ＃１〜ＡＰ＃Ｍの時刻ｔ_０＋ｎΔｔにおける割当帯域ｚ_ｋ（ｔ_０＋ｎΔｔ）は、
【００５５】
【数１３】

となる。
【００５６】
［▲６▼帯域割当手順の終了］
各ＡＰ＃１〜ＡＰ＃Ｎの割当帯域ｚ_ｋが最終割当帯域ｚ_ｌ＿ｆと同じになった時点で、帯域割当作業は終了する。
【００５７】
図４は、本発明の実施の形態に係る通信システムにおけるアドミッション制御装置の動作の一例を説明する図である。
【００５８】
アドミッション制御装置は、各アプリケーション（ＡＰ＃１〜ＡＰ＃５）に対して事前割当帯域としての伝送帯域を経路Ａおよび経路Ｂで以下のように割り当てているものとする。
【００５９】
経路Ａ（総伝送帯域はＴ_Ａ＝１００Ｍｂｐｓ、総事前割当帯域Ｚ_Ａ＿ｉ＝８０Ｍｂｐｓとする）
ＡＰ＃１（ＴＶ電話サービスＡ　１呼あたり２Ｍｂｐｓ）　ｚ_{１＿ｉ　＝}４０Ｍｂｐｓ（２０呼分接続可能）
ＡＰ＃２（ＴＶ電話サービスＢ　１呼あたり１Ｍｂｐｓ）　ｚ_{２＿ｉ　＝}２０Ｍｂｐｓ（２０呼分接続可能）
ＡＰ＃３（ＴＶ電話サービスＣ　１呼あたり０．５Ｍｂｐｓ）ｚ_{３＿ｉ　＝}２０Ｍｂｐｓ（４０呼分接続可能）
なお、残り２０Ｍｂｐｓに関しては、各アプリケーション＃１〜３がこの時点では使用不可能な、予備の帯域とする。
【００６０】
経路Ｂ（総伝送帯域Ｔ_Ｂ＝１００Ｍｂｐｓ、総事前割当帯域Ｚ_Ａ＿ｉ＝８０Ｍｂｐｓとする）
ＡＰ＃４（ＶｏＤサービスＡ　１ストリームあたり０．５Ｍｂｐｓ）　ｚ_{４＿ｉ　＝}４０Ｍｂｐｓ（８０ストリーム分接続可能）
ＡＰ＃５（ＶｏＤサービスＢ　１ストリームあたり１Ｍｂｐｓ）　ｚ_{５＿ｉ　＝}４０Ｍｂｐｓ（４０ストリーム分接続可能）
なお、残り２０Ｍｂｐｓに関しては、各アプリケーション＃４、５がこの時点では使用不可能な帯域である予備の帯域とする。
【００６１】
ここで、ネットワーク障害発生直前の時点ｔ_０での各アプリケーションの使用帯域は、
ＡＰ＃１　　ｘ_１（ｔ_０）　＝３０Ｍｂｐｓ（１５呼分接続）
ＡＰ＃２　　ｘ_２（ｔ_０）　＝２０Ｍｂｐｓ（２０呼分接続）
ＡＰ＃３　　ｘ_３（ｔ_０）　＝１０Ｍｂｐｓ（２０呼分接続）
ＡＰ＃４　　ｘ_４（ｔ_０）　＝３０Ｍｂｐｓ（６０ストリーム分接続）
ＡＰ＃５　　ｘ_５（ｔ_０）　＝３０Ｍｂｐｓ（３０ストリーム分接続）
であったとする。
【００６２】
上記のような状況で、経路Ａにネットワーク障害が発生し、ＡＰ＃１〜ＡＰ＃３がこの経路Ａで伝送帯域を使用できなくなったとする。このとき、アドミッション制御装置は、経路Ａで伝送帯域を専有していたＡＰ＃１〜ＡＰ＃３に対して経路Ｂにおいて新たな伝送帯域を以下のようにして割り当てる。
【００６３】
アドミッション制御装置は、経路ＢでＡＰ＃４およびＡＰ＃５で確保していた帯域の８０Ｍｂｐｓと、ＡＰ＃４およびＡＰ＃５で使用していなかった帯域２０Ｍｂｐｓのうち１０Ｍｂｐｓの合計９０Ｍｂｐｓを、障害発生前の事前確保帯域に基づいて比例配分してＡＰ＃１〜ＡＰ＃５に割り当てるものとする。
【００６４】
すなわち、９０Ｍｂｐｓを当初の事前確保帯域の比　４０：２０：２０：４０：４０＝２：１：１：２：２　で割り当てるので、ＡＰ＃１〜ＡＰ＃５に対して最終的に割り当てられる各伝送帯域（最終割当帯域）は、
ＡＰ＃１　　ｚ_１＿ｆ　＝２２．５Ｍｂｐｓ
ＡＰ＃２　　ｚ_２＿ｆ　＝１１．２５Ｍｂｐｓ
ＡＰ＃３　　ｚ_３＿ｆ　＝１１．２５Ｍｂｐｓ
ＡＰ＃４　　ｚ_４＿ｆ　＝２２．５Ｍｂｐｓ
ＡＰ＃５　　ｚ_５＿ｆ　＝２２．５Ｍｂｐｓ
となる。
【００６５】
しかしながら、直ちにこの最終割当帯域を実行すれば、あるアプリケーションに対しては、使用可能となる伝送帯域が減少し、使用中の呼やサービスなどが中断されてしまうケースが想定される。そこで、アドミッション制御装置は次のような動作を行うものとする。
【００６６】
経路Ａで障害発生したときに、経路Ｂでは、ＡＰ＃４が３０Ｍｂｐｓ、ＡＰ＃５が３０Ｍｂｐｓの帯域を使用中である。したがって、この時点で各々のＡＰの使用帯域は最終割当帯域を超過しているが、このＡＰ＃４およびＡＰ＃５の帯域は確保されるものとする。
すなわち、この時点ｔ_０での各々の割当帯域は、ＡＰ＃４が３０ＭｂｐｓおよびＡＰ＃５が３０Ｍｂｐになる。
【００６７】
一方、ＡＰ＃１〜ＡＰ＃３に対しては、合計４５Ｍｂｐｓの帯域が最終的に割り当てられることになるが、この時点ではＡＰ＃４とＡＰ＃５とが最終割当帯域を超過して合計６０Ｍｂｐｓを使用中であるため、３０Ｍｂｐｓ（＝９０−６０）がＡＰ＃１〜ＡＰ＃３に割り当てられる。
すなわち、この時点での割当帯域は、ＡＰ＃１が１５Ｍｂｐｓ、ＡＰ＃２が７．５　Ｍｂｐｓ、ＡＰ＃３が７．５Ｍｂｐｓ、となる。
【００６８】
アドミッション制御装置は、ＡＰ＃４およびＡＰ＃５に関しては当面、最終割当帯域の２２．５Ｍｂｐｓを参照値としてアドミッション制御を行う。また、アドミッション制御装置は、ＡＰ＃１〜ＡＰ＃３に関しては１５Ｍｂｐｓ、７．５Ｍｂｐｓ、７．５Ｍｂｐｓを基準としてアドミッション制御を行う。
【００６９】
次に、一定時間経過後の時刻ｔ_１におけるＡＰ＃４およびＡＰ＃５での使用帯域が、
ＡＰ＃４　　ｘ_４（ｔ_１）　＝２２Ｍｂｐｓ（ＶｏＤサービスＡが４４ストリーム分）
ＡＰ＃５　　ｘ_５（ｔ_１）　＝２４Ｍｂｐｓ（ＶｏＤサービスＢが２４ストリーム分）
であったとする。
【００７０】
この時刻ｔ_０〜時刻ｔ_１では、ＡＰ＃４およびＡＰ＃５については新たなストリームが加わることなく、各サービス終了と同時に使用帯域は減少していくはずである。
【００７１】
この時点では、ＡＰ＃５の使用帯域が最終割当帯域を超過しているため、ＡＰ＃５の割当帯域は２４Ｍｂｐｓとする。
【００７２】
また、ＡＰ＃４に関しては、使用帯域が最終割当帯域以下に減少したため、割当帯域を２２．５Ｍｂｐｓとする。
【００７３】
ここで、ＡＰ＃１〜ＡＰ＃３に対しては、新たに発生した帯域１３．５Ｍｂｐｓが比例配分して割り当てられる。したがって、ＡＰ＃１〜ＡＰ＃３の割当帯域は、
ＡＰ＃１　ｚ_１（ｔ_１）＝２１．７５Ｍｂｐｓ
ＡＰ＃２　ｚ_２（ｔ_１）＝１０．８７５Ｍｂｐｓ
ＡＰ＃３　ｚ_３（ｔ_１）＝１０．８７５Ｍｂｐｓ
となる。
【００７４】
さらに、一定時間経過後の時刻ｔ_２におけるＡＰ＃５での使用帯域がｘ_５（ｔ_２）＝２２Ｍｂｐｓ（ＶｏＤサービスＢが２２ストリーム分）であったとすると、この時点でのＡＰ＃１〜ＡＰ＃３の割当帯域は、最終割当帯域と一致する。すなわち、
ＡＰ＃１　ｚ_１（ｔ_２）＝２２．５Ｍｂｐｓ（＝ｚ_１＿ｆ）
ＡＰ＃２　ｚ_２（ｔ_２）＝１１．２５Ｍｂｐｓ（＝ｚ_２＿ｆ）
ＡＰ＃３　ｚ_３（ｔ_２）＝１１．２５Ｍｂｐｓ（＝ｚ_３＿ｆ）
また、ＡＰ＃４およびＡＰ＃５に対して、
ＡＰ＃４　ｚ_４（ｔ_２）＝２２．５Ｍｂｐｓ（＝ｚ_４＿ｆ）
ＡＰ＃５　ｚ_５（ｔ_２）＝２２．５Ｍｂｐｓ（＝ｚ_５＿ｆ）
となる。
【００７５】
アドミッション制御装置は、この時点で割当帯域変更の手順を完了する。
【００７６】
【発明の効果】
以上説明したように、本発明によれば、ネットワーク上の経路で障害が発生した場合に、当該障害が発生した経路上の伝送帯域を割り当てられていたアプリケーションの使用経路を別の経路（迂回経路）に切り替えることにより、各アプリケーションに新たな伝送帯域を割り当てることができる。
そして、本発明によれば、この新たに割り当てられる伝送帯域の値を各アプリケーションの使用帯域に応じて決定できる。
【図面の簡単な説明】
【図１】本発明の実施の形態に係る通信システムの構成を示す図である。
【図２】図１のＩＰネットワーク上の経路でネットワーク障害が発生した様子を示す図である。
【図３】本発明の実施の形態に係る通信システムにおけるアドミッション制御装置の動作を詳細に説明する図である。
【図４】本発明の実施の形態に係る通信システムにおけるアドミッション制御装置の動作の一例を説明する図である。[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a communication system, and particularly to a method for providing mainly streaming applications such as video distribution and two-way video communication on a network using IP (hereinafter referred to as “IP network”). The present invention relates to a communication system in which each application occupies a fixed transmission band.
[0002]
[Prior art]
The center of the services provided on the IP network is currently viewing e-mails and the Web. For this reason, the center of the communication system of the IP network is currently a best-effort type suitable for providing these services.
[0003]
However, with the speeding up of the access network, services utilizing broadband such as TV telephones and VoD (Video on Demand) distribution have been gradually developed. In the future, applications realizing services utilizing this broadband (for example, , Streaming-based applications, etc.) are expected to increase on IP networks.
[0004]
Therefore, in the future, the center of the communication system of the IP network will gradually shift to a communication system suitable for an application that realizes a service utilizing broadband, that is, a communication system in which each application occupies a transmission band in the network. Is assumed.
[0005]
As such a communication system, for example, a communication system described in Patent Literature 1 is known, and the communication system described in Patent Literature 1 has conventionally been a Diffserv (Differentiated services) MPLS on an IP network. (Multi Protocol Label Switching) and the like.
[0006]
[Patent Document 1]
JP 2002-111714 A
[0007]
[Problems to be solved by the invention]
However, the conventional communication system realized by a technology such as Diffserv / MPLS has the following problems.
[0008]
That is, in a network, some routes (for example, between routers) become unusable due to an accidental failure or the like, so that a specific application cannot occupy a transmission band allocated in advance. There is.
[0009]
Therefore, in this case, a new transmission band must be assigned to each application on another route (a detour route) on the network. However, in the above-described conventional communication system realized by a technology such as Diffserv / MPLS, However, there is a problem that this new allocation of the transmission band cannot be effectively performed.
[0010]
Accordingly, the present invention has been made in view of such circumstances, and when some routes are unavailable on the network, the value of the transmission bandwidth occupied by each application on the network is determined according to the bandwidth used by each application. An object of the present invention is to provide a communication system that dynamically reassigns data.
[0011]
[Means for Solving the Problems]
According to the present invention, the above-mentioned object is solved by the means described in the claims.
[0012]
In other words, the invention according to claim 1 is a communication system in which a transmission band on a path of a network existing between a server and a client is allocated to an application provided from the server to the client. The failure occurs when a transmission band on a path in which the failure has not occurred is newly assigned to an application to which a transmission band on a path in which communication cannot be performed due to the failure is newly assigned. A communication system in which a transmission band of an application to which a transmission band is allocated before the occurrence of the failure is changed on a path where the transmission is not performed.
[0013]
In the invention according to claim 2, a transmission band on each path of the network is pre-allocated as a pre-allocated band before a failure occurs to an application provided from a server to a client, and on each path of the network, A spare band that cannot be used by the application is provided, and a normal path to be used when a failure does not occur in the network and a detour path to be used when a failure occurs in the normal path are preset for the application before the failure occurs. When a failure occurs in the network, the pre-allocated bandwidth of an application that uses a certain route as a normal route when a fault occurs and the pre-allocated bandwidth of an application that uses the certain route as a bypass route when a fault occurs. Based on the certain route The total value (z + αw; 0 <α ≦ 1) of the total allocated bandwidth (z), which is the total of the pre-allocated bandwidth, and a part of the spare bandwidth (w) on the path where the failure has occurred is proportionally distributed. Assigned band z _{i_f} Is calculated, and the calculated allocated band (z _{i_f} ) Are respectively assigned to the application used as the normal route and the application used as the bypass route.
[0014]
According to the second aspect of the present invention, admission control can be performed within a range not exceeding the previously used band. According to the second aspect of the present invention, a new allocated bandwidth can be determined immediately after a failure occurs in the normal route.
[0015]
According to a third aspect of the present invention, for an application provided from a server to a client, a transmission band on each path of a network is pre-allocated as a pre-allocated band before a failure occurs, and on each path of the network, A spare band that cannot be used by the application is provided, and a normal path to be used when a failure does not occur in the network and a detour path to be used when a failure occurs in the normal path are preset for the application before the failure occurs. When a failure occurs in the network, the pre-allocated bandwidth of an application that uses a certain route as a normal route when a fault occurs and the pre-allocated bandwidth of an application that uses the certain route as a bypass route when a fault occurs. Based on the certain route The total value (z + αw; 0 <α ≦ 1) of the total allocated bandwidth (z), which is the total of the pre-allocated bandwidth, and a part of the spare bandwidth (w) on the path where the failure has occurred is proportionally distributed. Assigned bandwidth (z _{i_f} ) Is calculated, and for an application that uses the certain route as a bypass route, the assigned bandwidth (z _{i_f} ) For the certain route (Σz _{i_f} ) And a used band (x) which is a transmission band used by an application that makes the certain path a normal path when a failure occurs. _i ) To the allocated band (z _{i_f} ) Is subtracted from the total value (あ る (x _i -Z _{i_f} )), And a transmission band (z + αw− (Σz), which is subtracted from a total value (z + αw; 0 <α ≦ 1) of a part of the total allocated band (z) and the spare band (w). _{i_f} ) -Σ (x _i -Z _{i_f} )) Is allocated proportionally based on the allocated bandwidth before the occurrence of the failure, and the allocated bandwidth (z _{i_f} ), The application using the transmission band (x), which is the transmission band in use, _i ) Is assigned and the assigned band (z _{i_f} ) For applications using the following transmission bands, the allocated band (z _{i_f} ) Is assigned, and the total value (Σ (x _i -Z _{i_f} )), And the assignment is repeated until the total value becomes zero.
[0016]
BEST MODE FOR CARRYING OUT THE INVENTION
FIG. 1 is a diagram showing a configuration of a communication system according to an embodiment of the present invention.
[0017]
The communication system according to the embodiment of the present invention can be realized by, for example, an IP network. On this IP network, as shown in FIG. 1, servers (AP # 1 to AP # N servers) Client (AP # 1 to AP # N client) and an admission control device.
Note that “AP” is an abbreviation for application, but these entities are applications that implement video distribution, TV phone, and the like. These applications may be the same streaming application (for example, AP # 1 and AP # 2 are both applications that implement video distribution), but may be different applications (for example, , AP # 1 is an application that implements video distribution, and AP # 2 is an application that implements a TV phone).
[0018]
A plurality of paths (path A and path B) exist between the server (AP # 1 to AP # N server) and the client (AP # 1 to AP # N client), and are provided from the server to the client. Each of the applications (AP # 1 to AP # N) occupies a usable transmission band in the plurality of paths (path A and path B).
In the present embodiment, the number of paths is two for the sake of simplicity, but the present invention does not limit the number of paths to two.
[0019]
Here, in this specification, allocating a certain application to a certain transmission band on a network is referred to as allocating a transmission band to the application.
[0020]
The admission control device manages the bandwidth used on the network for each application (AP # 1 to AP # N) and, based on the bandwidth used, allows each application to occupy a new transmission bandwidth.・ Perform non-permission control.
[0021]
It is assumed that each path (path A, path B) has a spare band that cannot be used by any application (AP # 1 to AP # N).
In addition, as a specific method of exclusively using a transmission band between each node (router, layer 3 switch, and the like), for example, an existing technology such as Diffserv is used.
[0022]
FIG. 2 is a diagram showing a state in which a network failure has occurred on the route on the IP network in FIG.
[0023]
As shown in FIG. 2, in the communication system according to the embodiment of the present invention, when a network failure occurs on a certain path (for example, path A) on the IP network, the admission control device transmits the information on this path A The application (AP # 1 to AP # M), which has occupied the transmission band in step (1), occupies a new transmission band on another path (for example, path B) on the network.
[0024]
The admission control device determines the value of the transmission band newly occupied by the application (AP # 1 to AP # M) by the following method.
[0025]
The total transmission bandwidth value of the transmission bandwidth newly allocated in the path B is proportionally distributed according to the transmission bandwidth occupied by the application (AP # 1 to AP # N) in the path A or the path B before the network failure occurs. This is the final allocated bandwidth of each application (AP # 1 to AP # N).
The total transmission bandwidth value of the transmission bandwidth to be allocated is the sum of the transmission bandwidth occupied by the path B before the occurrence of the failure and some percentage of the transmission bandwidth value of the protection bandwidth of the path B.
An application (AP # M + 1 to AP # N) that has been using the route B before the occurrence of the network failure in the route A uses a transmission bandwidth equal to or greater than the newly allocated transmission bandwidth value before the occurrence of the network failure. In this case, the used transmission band is allocated to these applications (AP # M + 1 to AP # N) as they are.
The above-mentioned new allocation of the transmission band is periodically performed until a predetermined time elapses from the new allocation until the allocated bandwidths of all the applications (AP # 1 to AP # N) converge on the final allocated bandwidth. (Or irregularly).
[0026]
FIG. 3 is a diagram illustrating the operation of the admission control device in the communication system according to the embodiment of the present invention in detail. Hereinafter, the operation of the admission control device in the communication system according to the embodiment of the present invention will be described with reference to FIG.
[0027]
In the following, it is assumed that it is difficult to prepare the same amount of bandwidth as the normal route for all applications, and a spare bandwidth is prepared in advance and used as an increment of the bypass route bandwidth when a failure occurs. The case will be described.
[0028]
[(1) Setting of pre-allocated bandwidth to each application]
As described above, the communication system according to the embodiment of the present invention is implemented by an IP network including a plurality of routers and nodes. In this IP network, services (applications (AP # 1 to AP # N)) are provided in a server / client type.
[0029]
A plurality of routes (route A, route B) exist between the server and the client. A service (application (AP # 1 to AP # N)) is assigned to each route (route A, route B).
[0030]
Here, it is assumed that M applications from AP # 1, AP # 2, AP # 3 to AP # M (M is a positive integer) are allocated to the route A.
[0031]
Also, the k-th application AP # k has z _{k_i} Band is allocated in advance, and the band used at a certain time t is x _k (T).
[0032]
Further, the entire band T of the path A _A Of the bandwidths not allocated to AP # 1 to AP # M, _{A_i} And This band W _{A_i} Cannot normally be used by AP # 1 to AP # M.
[0033]
Then, the sum of the pre-allocated bands of AP # 1 to AP # M is Z _{A_i} And
[0034]
Therefore, the following equation holds.
(Equation 1)

[0035]
It is assumed that NM applications from AP # M + 1 to AP # N are assigned to the route B in advance.
[0036]
Also, the l-th application AP # l has z _{l_i} Band is allocated in advance, and the band used at a certain time t is x _l (T).
[0037]
Also, the entire band T of the path B _B Of the bandwidths not allocated to AP # M + 1 to AP # N, _{B_i} And This band W _{B_i} Is normally not available to AP # M + 1 to AP # N.
[0038]
Then, the sum of the pre-allocated bands of AP # M + 1 to AP # N is Z _{B_i} And
[0039]
Therefore, the following equation holds.
(Equation 2)

[0040]
[[2] Occurrence of network failure]
Next, it is assumed that a network failure has occurred in the path A and communication has been disabled. At this time, the admission control device allows the applications (AP # 1 to AP # M) that are set to exclusively use the transmission band of the path A so that they can exclusively use the transmission band of the path B. Shall be reset to
Therefore, for AP # 1 to AP # M, path B is a detour path.
[0041]
[3] Calculation of final allocated bandwidth for each application
When a failure occurs in the route A, the applications (AP # 1 to AP #) provided in the route A are secured while securing the bandwidth values of the applications (AP # M + 1 to AP # N) provided in the route B. A transmission band is sequentially allocated to M).
[0042]
The bandwidth T of the route B by the applications AP # 1 to AP # N _B + ΑW _B (Where α is 0 <α ≦ 1.0.). The final allocated bandwidth of each application (AP # 1 to AP # N) is expressed as follows.
[Equation 3]

(Equation 4)

[0043]
[4] Verification of allocatable bandwidth
Here, the band in use is maintained for AP # 1 (M + 1 ≦ l ≦ N) that used the path B before the failure occurred. The transmission band assigned to these AP # 1 (M + 1 ≦ l ≦ N) fluctuates at the end of service (for example, after a line disconnection for a TV phone). Therefore, the quality of the service being provided is guaranteed.
[0044]
Then, the time t at which the failure occurred in the route A ₀ , The band usage status of AP # M + 1 to AP # N is detected, and an empty band is allocated to AP # 1 to AP # M.
Transmission band (used band) x used by AP # 1 [M + 1 ≦ l ≦ N] on route B _l (T) corresponds to any of the following.
(1) Bandwidth x _l (T ₀ ) Is the final allocated bandwidth z after redistribution. _{l_f} In the following cases:
(Equation 5)

, The time t of AP # 1 ₀ Assigned bandwidth z _l (T ₀ ) To z _{l_f} And
(2) Bandwidth x _l (T ₀ ) Is the final allocated bandwidth z after redistribution. _{l_f} If greater than,
(Equation 6)

, The time t of AP # 1 ₀ Assigned bandwidth z _l (T ₀ ) To x _l (T ₀ ).
[0045]
For each AP # 1 [M + 1 ≦ l ≦ N], it is determined which one of (1) and (2) is applicable, and a new one among the bands not used by AP # M + 1 to AP # N is newly determined. The bandwidth that can be used by AP # 1 to AP # M is estimated.
[0046]
Here, from (2),
(Equation 7)

Is a band used by AP # 1 [M + 1 ≦ l ≦ N] exceeding the final allocated band.
[0047]
Also, the final allocated bandwidth Z for AP # 1 to AP # N _{A_f} Is the sum of
(Equation 8)

[0048]
At time t of AP # 1 to AP # N ₀ Assigned bandwidth z at _k (T ₀ )
(Equation 9)

It becomes.
[0049]
[5] Continuation of bandwidth allocation
Next, the procedure of the above (4) is repeated every certain time. Here, for simplification of the description, it is assumed that the process is repeated at regular time intervals Δt. That is, the procedure is as follows.
[0050]
Time t ₀ + NΔt (n is a positive integer), the used band x of AP # 1 [M + 1 ≦ l ≦ N] being used in the path B _l (T ₀ + NΔt) corresponds to any of the following.
[0051]
(3) Bandwidth x _l (T ₀ + NΔt) is the final allocated band z after redistribution. _{l_f} In the following cases:
(Equation 10)

, The time t of AP # 1 ₀ + NΔt assigned bandwidth z _l (T ₀ + NΔt) to z _{l_f} And
(4) Bandwidth x _l (T ₀ + NΔt) is the final allocated band z after redistribution. _{l_f} If greater than,
[Equation 11]

, The time t of AP # 1 ₀ + NΔt assigned bandwidth z _l (T ₀ + NΔt) is assigned bandwidth at the time, that is, x _l (T ₀ + NΔt).
[0052]
For each AP # 1 [M + 1 ≦ l ≦ N], it is determined which of (3) and (4) applies, and a new band among the bands not used by AP # M + 1 to AP # N is newly determined. The allocated bandwidth that can be used by AP # 1 to AP # M is calculated.
[0053]
Here, from (2),
(Equation 12)

Is a band used by AP # 1 [M + 1 ≦ l ≦ N] exceeding the final allocated band.
[0054]
Also, the time t of AP # 1 to AP # M ₀ + NΔt assigned bandwidth z _k (T ₀ + NΔt) is
[0055]
(Equation 13)

It becomes.
[0056]
[6] End of bandwidth allocation procedure
Assigned bandwidth z of each AP # 1 to AP # N _k Is the final assigned band z _{l_f} At this point, the bandwidth allocation operation ends.
[0057]
FIG. 4 is a diagram illustrating an example of an operation of the admission control device in the communication system according to the embodiment of the present invention.
[0058]
It is assumed that the admission control device allocates a transmission band as a pre-allocated band to each application (AP # 1 to AP # 5) on route A and route B as follows.
[0059]
Route A (total transmission bandwidth is T _A = 100 Mbps, total pre-allocated bandwidth Z _{A_i} = 80Mbps)
AP # 1 (TV phone service A 2 Mbps per call) z _{1_i =} 40Mbps (connectable for 20 calls)
AP # 2 (TV Mbps service B 1Mbps per call) z _{2_i =} 20Mbps (connectable for 20 calls)
AP # 3 (TV phone service C 0.5Mbps per call) z _{3_i =} 20Mbps (40 calls can be connected)
Note that the remaining 20 Mbps is a spare band that cannot be used by each of the applications # 1 to # 3 at this time.
[0060]
Path B (total transmission bandwidth T _B = 100 Mbps, total pre-allocated bandwidth Z _{A_i} = 80Mbps)
AP # 4 (VoD service A 0.5 Mbps per stream) z _{4_i =} 40Mbps (80 streams can be connected)
AP # 5 (1 Mbps per stream for VoD service B) z _{5_i =} 40Mbps (40 streams can be connected)
In addition, regarding the remaining 20 Mbps, each of the applications # 4 and # 5 is a spare band that is unusable at this time.
[0061]
Here, the time t just before the network failure occurs ₀ The bandwidth used by each application at
AP # 1 x ₁ (T ₀ ) = 30Mbps (15 calls)
AP # 2 x ₂ (T ₀ ) = 20 Mbps (connection for 20 calls)
AP # 3 x ₃ (T ₀ ) = 10Mbps (connection for 20 calls)
AP # 4x ₄ (T ₀ ) = 30Mbps (connection for 60 streams)
AP # 5x ₅ (T ₀ ) = 30Mbps (connection for 30 streams)
Assume that
[0062]
In the above situation, it is assumed that a network failure occurs in the route A, and the AP # 1 to AP # 3 cannot use the transmission band in the route A. At this time, the admission control device allocates a new transmission band on the path B to the APs # 1 to # 3 that have occupied the transmission band on the path A as follows.
[0063]
The admission control device causes a failure of 80 Mbps of the band secured by AP # 4 and AP # 5 on route B and a total of 90 Mbps of 10 Mbps among the 20 Mbps not used by AP # 4 and AP # 5. It is assumed that proportional allocation is made to AP # 1 to AP # 5 based on the pre-reserved bandwidth before occurrence.
[0064]
That is, since 90 Mbps is allocated at the ratio 40: 20: 20: 40: 40 = 2: 1: 1: 2: 2 of the initially reserved bandwidth, each of the APs to be finally allocated to AP # 1 to AP # 5 is allocated. The transmission band (final allocation band)
AP # 1z _{1_f} = 22.5Mbps
AP # 2z _{2_f} = 11.25 Mbps
AP # 3z _{3_f} = 11.25 Mbps
AP # 4z _{4_f} = 22.5Mbps
AP # 5z _{5_f} = 22.5Mbps
It becomes.
[0065]
However, if this final allocated bandwidth is immediately executed, a transmission band that can be used for a certain application is reduced, and a call or a service in use may be interrupted. Therefore, the admission control device performs the following operation.
[0066]
When a failure occurs in the path A, in the path B, the AP # 4 uses the bandwidth of 30 Mbps, and the AP # 5 uses the bandwidth of 30 Mbps. Therefore, at this point, the used bandwidth of each AP exceeds the final allocated bandwidth, but the bandwidths of AP # 4 and AP # 5 are to be secured.
That is, at this time t ₀ Are 30 Mbps for AP # 4 and 30 Mbps for AP # 5.
[0067]
On the other hand, a total bandwidth of 45 Mbps is finally allocated to AP # 1 to AP # 3, but at this time, AP # 4 and AP # 5 exceed the final allocated bandwidth and a total of 60 Mbps is allocated. Is being used, 30 Mbps (= 90-60) is allocated to AP # 1 to AP # 3.
That is, the allocated bandwidth at this time is 15 Mbps for AP # 1, 7.5 Mbps for AP # 2, and 7.5 Mbps for AP # 3.
[0068]
For the time being, the admission control device performs admission control on AP # 4 and AP # 5 with reference to 22.5 Mbps of the final allocated bandwidth. Further, the admission control device performs admission control on the basis of 15 Mbps, 7.5 Mbps, and 7.5 Mbps for AP # 1 to AP # 3.
[0069]
Next, the time t after a certain time has passed ₁ Bandwidth used by AP # 4 and AP # 5 in
AP # 4x ₄ (T ₁ ) = 22 Mbps (44 streams of VoD service A)
AP # 5x ₅ (T ₁ ) = 24 Mbps (24 streams of VoD service B)
Assume that
[0070]
This time t ₀ To time t ₁ Then, for AP # 4 and AP # 5, the used bandwidth should decrease at the end of each service without adding a new stream.
[0071]
At this point, the allocated bandwidth of AP # 5 is 24 Mbps because the bandwidth used by AP # 5 exceeds the final allocated bandwidth.
[0072]
For AP # 4, the allocated bandwidth is set to 22.5 Mbps because the used bandwidth has been reduced below the final allocated bandwidth.
[0073]
Here, a newly generated bandwidth of 13.5 Mbps is allocated to AP # 1 to AP # 3 in proportion. Therefore, the allocated bandwidth of AP # 1 to AP # 3 is:
AP # 1z ₁ (T ₁ ) = 21.75 Mbps
AP # 2z ₂ (T ₁ ) = 1.875 Mbps
AP # 3z ₃ (T ₁ ) = 1.875 Mbps
It becomes.
[0074]
Further, the time t after a certain time has passed ₂ The bandwidth used by AP # 5 is x ₅ (T ₂ ) = 22 Mbps (22 streams of VoD service B), the allocated bandwidth of AP # 1 to AP # 3 at this time coincides with the final allocated bandwidth. That is,
AP # 1z ₁ (T ₂ ) = 22.5 Mbps (= z _{1_f} )
AP # 2z ₂ (T ₂ ) = 11.25 Mbps (= z _{2_f} )
AP # 3z ₃ (T ₂ ) = 11.25 Mbps (= z _{3_f} )
Also, for AP # 4 and AP # 5,
AP # 4z ₄ (T ₂ ) = 22.5 Mbps (= z _{4_f} )
AP # 5z ₅ (T ₂ ) = 22.5 Mbps (= z _{5_f} )
It becomes.
[0075]
At this point, the admission control device completes the procedure for changing the allocated bandwidth.
[0076]
【The invention's effect】
As described above, according to the present invention, when a failure occurs on a route on a network, the use route of the application to which the transmission band on the failed route has been allocated is changed to another route (a bypass route). ), A new transmission band can be allocated to each application.
Then, according to the present invention, the value of the newly allocated transmission band can be determined according to the band used by each application.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration of a communication system according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a state in which a network failure has occurred on a route on the IP network in FIG. 1;
FIG. 3 is a diagram illustrating in detail an operation of the admission control device in the communication system according to the embodiment of the present invention.
FIG. 4 is a diagram illustrating an example of an operation of the admission control device in the communication system according to the embodiment of the present invention.

Claims (3)

A communication system in which a transmission band on a network path existing between a server and a client is allocated to an application provided from a server to the client,
When a transmission band on a path where no failure has occurred is newly assigned to an application to which a transmission band on a path where communication has become impossible due to a failure that has occurred has been assigned, the failure occurs. A transmission bandwidth of an application to which a transmission bandwidth has been allocated before the occurrence of the failure on a path where no failure has occurred. For the application provided from the server to the client, the transmission bandwidth on each path of the network is pre-allocated as a pre-allocated bandwidth before a failure occurs,
A spare band that cannot be used by the application is provided on each path of the network,
For the application, a normal path to be used when a failure has not occurred in the network and a detour path to be used when a failure has occurred in the normal path are set in advance before the occurrence of the failure,
When a failure occurs in the network, based on the pre-allocated bandwidth of an application that uses a certain route as a normal route when a fault occurs and the pre-allocated bandwidth of an application that uses the certain route as a bypass route when a fault occurs. Sum (z + αw; 0 <α ≦ 1) of the total allocated bandwidth (z), which is the total of the pre-allocated bandwidths on the certain path, and a part of the spare bandwidth (w) on the failed path. ) _Is proportionally distributed to calculate an allocated bandwidth z _{i_f} , and the calculated allocated bandwidth (z _{i_f} ) is allocated to the application used as the normal route and the application used as the bypass route, respectively.
A communication system, comprising: For the application provided from the server to the client, the transmission bandwidth on each path of the network is pre-allocated as a pre-allocated bandwidth before a failure occurs,
A spare band that cannot be used by the application is provided on each path of the network,
For the application, a normal path to be used when a failure has not occurred in the network and a detour path to be used when a failure has occurred in the normal path are set in advance before the occurrence of the failure,
When a failure occurs in the network, based on the pre-allocated bandwidth of an application that uses a certain route as a normal route when a fault occurs and the pre-allocated bandwidth of an application that uses the certain route as a bypass route when a fault occurs. Sum (z + αw; 0 <α ≦ 1) of the total allocated bandwidth (z), which is the total of the pre-allocated bandwidths on the certain path, and a part of the spare bandwidth (w) on the failed path. ) _Is proportionally distributed to calculate an allocated band ( _{zi_f} ),
For applications detour path said certain path, the the allocated bandwidth the certain sum in the path of _{(z i_f) (Σz i_f)} , applications that the certain path normal path when a failure occurs in use the certain sum in the path of a value obtained by subtracting the allocated bandwidth _{(z i_f)} from a transmission band using band _{(x i)} and _{(Σ (x i -z i_f)} ) and, the total allocated bandwidth (z) the portion of the sum of the backup band (w); transmission band obtained by subtracting from the (z + αw 0 <α ≦ 1) (z + αw - (Σz i_f) -Σ (x i -z i_f)) is, failure of the previous generation Allocated proportionally based on the allocated bandwidth,
On the other hand, for an application that uses the transmission band beyond the allocated band (z _{i_f} ) among the applications in which the path in which the failure has occurred is a normal path, the used band (x _i) that is the used transmission band is used. ) Is assigned,
The assigned bandwidth (z _{i_f} ) is assigned to an application that uses a transmission bandwidth equal to or less than the assigned bandwidth (z _{i_f} ) among the applications in which the faulty path is a normal path,
At predetermined time intervals, the sum _{(Σ (x i -z i_f)} ) was calculated, until the total value becomes 0, the allocation is repeated Kaee,
A communication system, comprising:

Images