200807950 九、發明說明： 【發明所屬之技術領域】 本發明係關於網路通訊，尤其是有關於令網路連線裝置 重啟以排除通訊障礙的系統與方法。 【先前技術】 透過廣域網路（例如公眾網際網路）的雙向網路通訊牽 涉到諸多的相關設備。為了便於說明，本說明從用戶端 Φ (client)的觀點提出一個如第la圖所示的簡化的模型。如 圖所示，位於一用戶端的運算裝置1〇1、102是連接在一個區 域網路10上，運算裝置101、1〇2進一步透過一個網路連線 裝置103和廣域網路20連線，進而和遠端的運算裝置2〇1進 行雙向的網路通訊。運算裝置1〇1、1〇2、201可以是各種形 式的運算裝置，例如桌上型電腦、伺服器、甚至是智慧型手 機、PDA等。網路連線裝置1〇3則可以是路由器（r〇uter)、 _ 閘道器（gateway )'ADSL數據機（modem )、纜線數據機(cable modem)、交換器（switch)等等，提供運算裝置1〇1、1〇2通 在廣域網路20的路控上的任何相關設備。圖中雖然只繪出一 個網路連線裝置103，但不以此為限。另外區域網路1〇可以 是符合802.3的有線以太網路、802.1 lx的無線區域網路、有 線與無線混合的網路、802.16Wi-Max或是符合其他標準的區 域網路。廣域網路20則可以是公眾網際網路（internet)、或 是專屬的私人網路、或是2.5G或3G行動數據網路。整個用 200807950 戶端可以是位於一戶住宅、一間辦公室、或是一個機房内等 等。 第la圖所示的網路常發生的一種通訊障礙情形是運算裝 置101、102、201都在正常運轉、廣域網路2〇也是暢通的， 但是網路連線裝置103卻停止運作。網路連線裝置丨〇3並不 是有不可回復的硬體故障，而是其作業系統或控制軟體因某 些原因耗盡資源（例如記憶體、緩衝空間）或進入一種無限 迴圈（infinite loop)或死結（deadl〇ck)的狀態，因此將網 路連線裝置103重啟（reb00t)後就釋放了佔用的資源、或脫 離了這種迴圈、死結的狀態，這種通訊的障礙也就隨之排除。 停止運作的網路連線裝置因為其作業系統或控制軟體已 經無法作用’所以通常只有透過裝置的電源開關、面板上的 按鍵（例如reset鍵）等進行重啟。但是在網路的時代，要重 啟遠端的網路連線裝置就是一個挑戰了。也因此有許多相關 的裝置與方法被提出來解決這樣的需求。例如第lb圖所示的 所明运‘電力控制（remote power control)的作法是在網路 連線裝置103連接電源40的電力線路上串接一個重啟裝置 3〇1。網路連線裝置103的電源開關是固定設定為開啟，而運 算裝置201在發現網路連線裝置103沒有回應時，經由控制 網路30下達指令使重啟裝置301先切斷而後再重新恢復對網 路連線裝置103的供電來重啟網路連線裝置103。控制網路 6 200807950 30和廣域網路20可以是同一個（例如都是公眾網際網路）、 也可以是不同的（例如廣域網路20可以是公眾網際網路，而 控制網路30可以公眾交換電話網路）。 由於這種耗盡資源或進入或死結（deadlock )的情形常會 週期性的發生，所以還有一種重啟網路連線裝置ί〇3的作法 是不等到網路連線裝置103停止運作才重啟，而是週期性的 令網路連線裝置103於深夜或網路通訊流量不繁忙的時受自 Φ 動重啟。例如第lb圖的運算裝置201 (透過廣域網路20)或 是第lc圖的運算裝置102 (透過區域網路10)都可以固定每 隔一定時間，透過重啟裝置重啟網路連線裝置103。重啟裝 置301已是商業化的產品，而運算裝置201或102的定時機 制也多是其作業系統内建的功能，所以有關的細節就不再贅 述° 不論是由遠端或是近端的運算裝置發動重啟，很明顯的 • 都有一個缺點，就是可能會導致許多不必要的重啟（因為裝 置並未耗盡資源或進入死結的情況）。這種不必要的重啟對於 網路連線裝置與網路通訊都會造成不必要的負擔（例如會影 響裝置的硬體壽命、造成通訊的暫時中斷等）。 【發明内容】 因此，本發明提出一種令網路連線裝置重啟的系統與方 法，以更有效率的避免不必要的重啟。 200807950 本發明所出的系統係用於重啟一用戶端的區域網路與廣 域網路之間的一或多個網路連線裝置，以排除這些網路連線 裝置所造成的區域網路與廣域網路間的通訊障礙。本系統包 含至少一主控裝置、一抑制裝置、一連線機制、以及至少一 重啟裝置。其中，每一被重啟的網路連線裝置的供電都是由 一重啟裝置所控制。主控裝置係位於區域網路上，並且係透 過這些網路連線裝置與廣域網路連線。主控裝置並透過獨立 於區域網路以外的連線機制以控制每一個重啟裝置。抑制裝 置則可以是位於於區域網路、或是位於廣域網路上。 主控裝置是每隔一第一週期觸發重啟裝置以重啟網路連 線裝置。當抑制裝置係位於廣域網路上時，抑制裝置是以一 適當地小於第一週期的一第二週期發出抑制訊息，通知主控 裝置不要觸發重啟裝置，而主控裝置收到抑制訊息後即取消 未來最近一次的觸發與重啟，並重行排定每隔第一週期、透 過該連線機制觸發該重啟裝置以重啟該網路連線裝置。當某 一個網路連線裝置耗盡資源或進入死結而停止運作時，抑制 裝置的抑制訊息將無法到達主控裝置，而主控裝置就會如期 的觸發重啟裝置而重啟網路連線裝置。而在網路連線裝置因 重啟而恢復運作後，抑制裝置的抑制訊息就會繼續送達主控 裝置而避免了不必要的重啟。 當抑制裝置係位於區域網路上時，抑制裝置同樣是以一 200807950 適當地小於第一週期的一第二週期運作，當第二週期到達 時，首先會發出一個詢問訊息給廣域網路上的一個回應裝 置，如果回應裝置有傳回回應的訊息，抑制裝置就會接著發 出抑制訊息給主控裝置。當某一個鱗路連線裝置耗盡資源或 進入死結而停止運作時，詢問或回覆的訊息無法成功送達， 抑制裝置就會停止發出抑制訊息給主控裝置，而主控裝置就 會如期的觸發重啟裝置而重啟網路連線裝置。而在網路連線 Φ 裝置恢復運作後，抑制裝置的詢問訊息與回應裝置的回應訊 息都可以成功送達，抑制裝置的抑制訊息就會繼續送達主控 裝置而避免了不必要的重啟。 茲配合所附圖示、實施例之詳細說明及申請專利範圍， 將上述及本發明之其他目的與優點詳述於後。然而，當可了 解所附圖示純係為解說本發明之精神而設，不當視為本發明 範疇之定義。有關本發明範疇之定義，請參照所附之申請專 _ 利範圍。 【實施方式】 - 本發明所提出的網路連線裝置重啟的系統與方法係用於 重啟一用戶端的區域網路與廣域網路之間的一或多個網路連 線裝置，以排除這些網路連線裝置所造成的區域網路與廣域 網路間的通訊障礙。 第2a圖所示係依據本發明所提出系統的第一實施例的示 200807950 意圖。如圖所示，本系統包含至少一主控裝置1〇5、一抑制 裝置202、——連線機制50、以及至少一重啟裝置301。主控 裝置105與抑制裝置202就是一般的運算裝置，他們本身也 可以同時參與和其他運算裝置101、102、2〇1的雙向網路通 訊，只是同時負擔有·「主控」與「抑制」.的功能。或者，主 控裝置105與抑制裝置202也可以是專門只負責「主控」與 「抑制」的功能。 區域網路10與廣域網路20之間有一或多個網路連線裝 置，在第2a圖中共繪出二個網路連線裝置1〇3、1〇4(例如網 路連線裝置1〇3是ADSL數據機而網路連線裝置1〇4是路由 器）重啟裝置301是串接在網路連線裝置ι〇3、1〇4與電源 40之間的供電線路上。重啟裝置301的特徵是具有（1)透過連 線機制50而被觸發的能力；以及(2)被觸發時將該供電線路切 斷而後接續以重啟網路連線裝置1〇3、1〇4的能力。 主控裝置105係位於區域網路1〇上，並且係透過這些網 路連線裝置103、104與廣域網路20連線。如前所述，主控 裝置105是透過一個連線機制5〇與重啟裝置3〇1連接，進而 得以觸發重啟裝置301的作動。請同時參閱第2b圖所示的另 實施例。如第2b圖所示，網路連線裝置1〇3、1〇4的供電 疋刀別由二個重啟裝置301所控制，而主控裝置1〇5是透過 連線機制50同時控制這二個重啟裝置301 (圖中的串接方式 200807950 僅屬例示）。連線機制50最主要的特徵是獨立於區域網路10 之外，以避免區域網路10本身的故障而影響到主控裝置105 對重啟裝置301的控制（例如網路連線裝置104就是區域網 路10裡的一個交換器）。連線機制50本身可以是一個有線/ 無線區域網路、或是控制網路（例如LonWorks®網路、 IEEE-488、802.15.4/Zigbee的短距離無線網路）等，連線機 制50也可以是單純的有線實體線路、紅外線的無線連結、或 者是利用電力線上的60Hz交流電為載波。重啟裝置301和主 控裝置105因此需要具有和連線機制50相匹配的實體界面以 及以一適當協定互相通訊的能力，舉例來說，重啟裝置301 可以是内建網路功能的I/O模組（像是國内研華⑽公司所生 產的具有DI/DO界面的控制器，或一般工控業界具Ethernet 模組的PLC)，而主控裝置105以獨立於區域網路10的另一 區域網路（連線機制50)和此内建網路功能的I/O模組連接。 抑制裝置202是位於廣域網路20上，它和主控裝置105都具 有週期性執行特定工作的能力（例如經由作業系統的工作排 程）。 請注意到，在某些實施例裡，主控裝置105與重啟裝置 301實體上可為同一設備，也就是說，一個硬體設備可以同 時具有主控裝置與重啟裝置的功能。這時，前述的連線機制 50就是經由該硬體設備内部的線路與匯流排的訊號傳輸機 11 200807950 制。 以下以第2a圖的系統為例說明本發明所提出的方法。本 發明所提出的方法係分為兩部分，這兩部分係分別獨立運行 並彼此互動。其中的一部份可以想像是實施在前述的主控裝 置105内。另外一部份可以想像是實施在前述的抑制裝置202 内。這二部分的一實施例的流程步驟分別顯示於第4a、4b圖。 如第4a圖所示，主控裝置105在步驟400設定在距離當時 φ 乃時間以後重啟網路連線裝置。主控裝置105接下來有兩條 可能的流程路徑。一是當乃時間到達以後，流程進入到步驟 401，主控部分經由連線機制50觸發重啟裝置301而使得網 路連線裝置103、104重啟，之後流程並回到步驟400繼續設 定在距離當時乃時間以後重啟網路連線裝置。換言之，主控 裝置105是以T!時間為週期，週期性的重啟網路連線裝置 103、104。步驟400以後的另一路徑是在乃時間到達之前收 Φ 到抑制裝置202送來的「不要重啟網路連線裝置」的抑制訊 息，流程因此進入到步驟403，主控裝置105取消最近一次 步驟401所設定的重啟排程，之後流程同樣回到步驟400重 • 新設定在距離當時几時間以後重啟網路連線裝置。 如第4b圖所示，抑制裝置202 —開始首先在步驟410對 主控裝置105發出一個「不要重啟網路連線裝置」的抑制訊 息。請注意到·，如果參照第4圖可以看出，抑制裝置202和 12 200807950 主控裝置105會因此抑告丨自 μ而同步。接下來，抑制裝置202 在步驟411設定在距離冬脖τ 田時丁2時間以後對主控裝置105發出 「不要重啟網路連線袭置」的抑制訊息。接下來當η時間到 達以後，流程回到步驟410，抑制裝置.對主控裝置1〇5 發出「不要重啟網路連線奘罟 深褒置」的訊息。換言之，抑制裝置 202是以I時間為週期，调 Ί生的送出抑制訊息給主控裝置 105。其實由於步驟41〇和411 Β 丄疋不斷重複循環，第4b圖的 流程也可以是由步驟411先開始。 由前述的說明可知，σ i τ各 /、戈i2適當的小於Τ!，而且網路連 線裝置103、104都在正常運作 1下的話，主控裝置105會被抑制 裝置2〇2「抑制」而不進行會 里敌。但若網路連線裝置103、104 其中之一因資源耗盡或進 八处、纟。而停止運作的話，由於主控 裝置1〇5無法從廣域網路2 收到抑制訊息，網路連線裝置 103、104就會如期的被 攸王控裝置1〇5重啟。當網路連線裝 103、104因重啟而恢復 息 運作4，主控裝置1〇5會繼續被「 制」而不進行重啟。請 处到，週期凡與丁2應具有適當的# 距以將網路僂鈐以，^ ^ 傳輪延遲、重啟所需㈣㈣因素考慮進去 於週期丁丨或Τ的大小 至 的大J則要現網路連線裝置的特性 發生死結的週发來 Λ及可能 4來决疋。主控裝置105重啟 裝置202抑制 和抑制 的週期丁2可以是各自内定於主控裝置1〇, 制裝置2〇2的參數，或者置1〇5和抑 次者也可以是抑制裝置加 13 200807950 裡動態指定主控裝置以T!為週期。 第3a、3b圖所示的實施例分別和第2a、2b圖的實施例 非常類似，其差別在於：（1)抑制裝置106是位於區域網路10 上；以及(2)廣域網路20多增加了 一個回應裝置203。主控裝 置105的運作方式即如第4a圖所示，抑制裝置106的運作則 是依照第4c圖的流程運作。如圖所示，抑制裝置106同樣是 以一適當地小於Τι週期的T2週期運作，首先，抑制裝置106 在步驟420對主控裝置105發出一個「不要重啟網路連線裝 置」的抑制訊息。請注意到，這個抑制訊息的主要目的是讓 抑制裝置106和主控裝置105同步。接下來，抑制裝置106 在步驟421設定在距離當時Τ2時間以後對主控裝置105發出 「不要重啟網路連線裝置」的抑制訊息。接下來當Τ2時間到 達以後，流程進入步驟422，抑制裝置106先會發出一個詢 問訊息給廣域網路20上的回應裝置203(例如使用ping或利 用專屬的TCP/IP封包），如果回應裝置203有傳回回覆的訊 息，流程就會回到步驟420而抑制裝置106就會接著發出抑 制訊息給主控裝置105。反之，當某一個網路連線裝置103 或104耗盡資源或進入死結而停止運作時，詢問或回覆的訊 息無法成功送達，抑制裝置106在得不到回應訊息的情形 下，流程回到步驟421，重新排定在T2時間後再試一次。由 於抑制裝置106沒有發出抑制訊息給主控裝置105，主控裝 14 200807950 置就會如期的觸發重啟裝置301而重啟網路連線裝置103、 104。換言之，抑制裝置106基本上也是以T2時間為週期， 週期性的送出抑制訊息給主控裝置105,而和抑制裝置202 位於廣域網路20的實施例主要的差別在抑制裝置202在週期 到達、送出抑制訊息前，先利用回應裝置203的有無回應來 判斷網路連線裝置103、104是否還在正常運作。其實由於步 驟420、421、和422是不斷重複循環，第4c圖的流程也可以 是由步驟421或422先開始。 回應裝置203也是一般的運算裝置，而且通常是利用廣 域網路20上某些特定的、非常穩定運作的網站或伺服器為對 象（只要他們支援類似ping這樣的詢答協定），所以未必須 要另外架設或提供。 請注意到前述的實施例都是相當簡單的重啟系統與方 法，而同樣的重啟系統與方法可以擴充到更為複雜的環境 中。例如第2a圖所示的位於廣域網路20的抑制裝置202,其 實可以對位於不同客戶端的多個主控裝置105廣播或——傳 送抑制訊息，以抑制這些主控裝置105進行重啟。還有一種 可能也是常見的情形是同一用戶端可以同時連接兩個廣域網 路以互為備援，那麼這個用戶端就可以有兩個主控裝置與至 少兩個重啟裝置以分別負責對這兩個廣域網路的網路連線裝 置的重啟，同時這兩個廣域網路就會需要各自的抑制裝置來 15 200807950 分別抑制這兩個主控裝置。對於這樣的應用環境，也可以採 用一個位於該用戶端區域網路的抑制裝置以廣播或——傳送 抑制訊息給這兩個抑制裝置（如果它們是位於同一區域網路 的話），而這兩個廣域網路也會有各自的回應裝置以供抑制裝 置來分別經由這兩個廣域網路來詢問。 藉由以上較佳具體實施例之詳述，係希望能更加清楚描 述本創作之特徵與精神，而並非以上述所揭露的較佳具體實 Φ 施例來對本創作之範疇加以限制。相反地，其目的是希望能 涵蓋各種改變及具相等性的安排於本創作所欲申請之專利範 圍的範疇内。 【圖式簡單說明】 第1 a圖所示係一簡化的雙向網路通訊的模型示意圖。 第lb圖所示係一種習知的重啟遠端網路連線裝置的系統示 意圖。 # 第lc圖所示係一種習知的週期性重啟網路連線裝置的系統示 意圖。 第2a圖所示係依據本發明所提出系統的第一實施例的示意 圖。 第2b圖所示係依據本發明所提出系統的第二實施例的示意 圖。 第3a圖所示係依據本發明所提出系統的第三實施例的示意 16 200807950 圖。 第3b圖所示係依據本發明所提出系統的第四實施例的示意 圖。 第4a圖所示係本發明所提出方法實施於主控裝置的流程圖。 第4b圖所示係本發明所提出方法實施於遠端抑制裝置的流 程圖。 第4c圖所示係本發明所提出方法實施於近端抑制裝置的流程 圖。 【主要元件符號說明】 10 區域網路 20 廣域線路 30 控制網路 40 電源 50 連線機制 101,102 運算裝置 103,104 網路連線裝置 105 主控裝置 106 抑制裝置 201 運算裝置 202 抑制裝置 203 回應裝置 301 重啟裝置 400〜402 流程步驟 410,411 流程步驟 17200807950 IX. INSTRUCTIONS: TECHNICAL FIELD OF THE INVENTION The present invention relates to network communications, and more particularly to systems and methods for restarting a network connection device to eliminate communication barriers. [Prior Art] Two-way network communication over a wide area network (e.g., the public Internet) involves a number of related devices. For ease of explanation, the present description proposes a simplified model as shown in Figure la from the perspective of the client Φ (client). As shown in the figure, the computing devices 101, 102 located at a subscriber end are connected to a regional network 10, and the computing devices 101, 102 are further connected through a network connection device 103 and the wide area network 20, and further Two-way network communication with the remote computing device 2〇1. The arithmetic units 1〇1, 1〇2, 201 may be various types of arithmetic devices such as a desktop computer, a server, or even a smart phone, a PDA, or the like. The network connection device 1〇3 can be a router (r〇uter), a _ gateway (ADSL modem), a cable modem, a switch, and the like. Any associated device that provides computing devices 1〇1, 1〇2 to the roadway of the wide area network 20 is provided. Although only one network connection device 103 is depicted in the figure, it is not limited thereto. In addition, the area network 1 can be an 802.3-compliant wired Ethernet channel, an 802.1 lx wireless local area network, a wired and wireless hybrid network, 802.16 Wi-Max, or a regional network that conforms to other standards. The WAN 20 can be a public internet, or a private private network, or a 2.5G or 3G mobile data network. The whole use of 200807950 can be located in a house, an office, or a computer room. A communication barrier that often occurs in the network shown in Fig. la is that the computing devices 101, 102, and 201 are all operating normally, and the wide area network 2 is also unblocked, but the network connection device 103 is stopped. The network connection device 丨〇3 is not an unrecoverable hardware failure, but its operating system or control software runs out of resources (such as memory, buffer space) or enters an infinite loop for some reason (infinite loop) Or the state of deadlock (deadl〇ck), so after the network connection device 103 is restarted (reb00t), the occupied resources are released, or the state of the loop and the dead knot is released, and the obstacle of the communication is also Excluded. The network connection device that has stopped operating has no effect because its operating system or control software has failed. Therefore, it is usually only restarted through the power switch of the device, the buttons on the panel (for example, the reset button). But in the age of the Internet, it is a challenge to re-open the remote network connection. There are therefore many related devices and methods that have been proposed to address such needs. For example, the "remote power control" shown in Fig. 1b is to connect a restarting device 3〇1 to the power line connecting the power source 40 to the network connecting device 103. The power switch of the network connection device 103 is fixedly set to be on, and when the computing device 201 finds that the network connection device 103 has not responded, the control device 30 issues an instruction to cause the restarting device 301 to cut off and then resume the pair. The network connection device 103 is powered to restart the network connection device 103. The control network 6 200807950 30 and the wide area network 20 may be the same (for example, both public networks) or different (for example, the wide area network 20 may be a public internet network, and the control network 30 may exchange telephones with the public. network). Since this situation of exhausting resources or entering or deadlocks often occurs periodically, there is also a way to restart the network connection device 〇3 without restarting after the network connection device 103 stops operating. Rather, the network connection device 103 is periodically restarted when the network connection device 103 is not busy at night or when the network communication traffic is not busy. For example, the computing device 201 of the lb diagram (through the wide area network 20) or the computing device 102 of the lcth diagram (through the area network 10) can be fixed at intervals, and the network connection device 103 is restarted by the restarting device. The restarting device 301 is already a commercial product, and the timing mechanism of the computing device 201 or 102 is mostly a built-in function of the operating system, so the details are not described again. Whether by remote or near-end operations The device is restarted, and it is obvious that there is a disadvantage that it may cause many unnecessary restarts (because the device does not run out of resources or enters a deadlock). This unnecessary restart imposes an unnecessary burden on the network connection device and network communication (for example, it may affect the hardware life of the device, causing temporary interruption of communication, etc.). SUMMARY OF THE INVENTION Accordingly, the present invention is directed to a system and method for restarting a network connection device to more efficiently avoid unnecessary restarts. 200807950 The system of the present invention is used to restart one or more network connection devices between a local area network and a wide area network of a client to exclude regional network and wide area network caused by these network connection devices. Communication barriers between. The system includes at least one master device, a suppression device, a connection mechanism, and at least one restart device. The power supply of each restarted network connection device is controlled by a restart device. The master device is located on the local area network and is connected to the WAN through these network connection devices. The master device controls each of the restart devices through a connection mechanism independent of the regional network. The suppression device can be located on a local area network or on a wide area network. The master device triggers the restart device every other first cycle to restart the network connection device. When the suppression device is located on the wide area network, the suppression device sends a suppression message in a second period that is appropriately smaller than the first period, notifying the master device not to trigger the restart device, and the master device cancels the future after receiving the suppression message. The most recent trigger and restart, and re-arranged every other first cycle, triggering the restart device through the connection mechanism to restart the network connection device. When a network connection device runs out of resources or enters a deadlock and stops operating, the suppression device's suppression message will not reach the master device, and the master device will restart the device and restart the network connection device as scheduled. After the network connection device resumes operation due to the restart, the suppression device suppression message will continue to be delivered to the main control device to avoid unnecessary restart. When the suppression device is located on the local area network, the suppression device is also operated with a second cycle of 200807950 which is suitably smaller than the first cycle. When the second cycle arrives, an inquiry message is first sent to a response device on the wide area network. If the responding device has a message to return a response, the suppression device will then send a suppression message to the master device. When a scale connection device runs out of resources or enters a dead knot and stops operating, the inquiry or reply message cannot be successfully delivered, and the suppression device stops sending a suppression message to the master device, and the master device triggers as scheduled. Restart the device and restart the network connection device. After the network connection Φ device resumes operation, the inquiry message of the suppression device and the response message of the response device can be successfully delivered, and the suppression message of the suppression device continues to be delivered to the main control device to avoid unnecessary restart. The above and other objects and advantages of the present invention will be described in detail with reference to the accompanying drawings and claims. However, it is to be understood that the appended drawings are merely illustrative of the scope of the invention. For the definition of the scope of the invention, please refer to the attached application scope. [Embodiment] - The system and method for restarting a network connection device according to the present invention is for restarting one or more network connection devices between a local area network and a wide area network of a client to exclude these networks. The communication barrier between the regional network and the wide area network caused by the connection device. Figure 2a shows the intent of the first embodiment of the system according to the invention. As shown, the system includes at least one master device 〇5, a damper device 202, a connection mechanism 50, and at least one restart device 301. The main control device 105 and the suppression device 202 are general arithmetic devices, and they themselves can simultaneously participate in two-way network communication with other computing devices 101, 102, and 2, but at the same time bear the "master" and "suppression". . Features. Alternatively, the master device 105 and the suppression device 202 may be exclusively responsible for "master" and "suppression" functions. There are one or more network connection devices between the local area network 10 and the wide area network 20, and two network connection devices 1〇3 and 1〇4 are drawn in the second picture (for example, the network connection device 1〇) 3 is an ADSL modem and the network connection device 1〇4 is a router. The restart device 301 is connected in series to the power supply line between the network connection devices ι〇3, 1〇4 and the power source 40. The restarting device 301 is characterized by having (1) the ability to be triggered by the connection mechanism 50; and (2) cutting the power supply line when triggered, and then continuing to restart the network connection device 1〇3, 1〇4 Ability. The master device 105 is located on the local area network 1 and is connected to the wide area network 20 via these network connection devices 103,104. As described above, the main control unit 105 is connected to the restart unit 3〇1 via a connection mechanism 5, and thus the activation of the restart unit 301 is triggered. Please also refer to the other embodiment shown in Figure 2b. As shown in FIG. 2b, the power supply tool of the network connection device 1〇3, 1〇4 is controlled by two restart devices 301, and the main control device 1〇5 controls the two through the connection mechanism 50 at the same time. The restarting device 301 (the serial connection method 200807950 in the figure is merely an example). The most important feature of the connection mechanism 50 is that it is independent of the area network 10 to avoid the failure of the area network 10 itself and affect the control of the restart device 301 by the master device 105 (for example, the network connection device 104 is the area). A switch in network 10). The connection mechanism 50 itself may be a wired/wireless local area network, or a control network (such as a LonWorks® network, IEEE-488, 802.15.4/Zigbee short-range wireless network), etc., the connection mechanism 50 It may be a simple wired physical line, a wireless connection of infrared rays, or a carrier using 60 Hz alternating current on a power line. The reboot device 301 and the master device 105 therefore need to have a physical interface that matches the connection mechanism 50 and the ability to communicate with each other with an appropriate protocol. For example, the reboot device 301 can be an I/O module with built-in network functionality. The group (such as a controller with a DI/DO interface produced by the domestic Advantech (10) company, or a PLC with an Ethernet module in the general industrial control industry), and the main control device 105 is another regional network independent of the regional network 10. The road (connection mechanism 50) is connected to the I/O module of the built-in network function. The suppression device 202 is located on the wide area network 20, and both the host device 105 have the ability to periodically perform specific tasks (e.g., via the work schedule of the operating system). It is noted that in some embodiments, the master device 105 and the reboot device 301 can be physically the same device, that is, a hardware device can have the functions of a master device and a reboot device at the same time. At this time, the aforementioned connection mechanism 50 is made by the signal transmitter 11 200807950 of the line and the bus bar inside the hardware device. The method proposed by the present invention will be described below by taking the system of Fig. 2a as an example. The method proposed by the present invention is divided into two parts, each of which operates independently and interacts with each other. Some of these are conceivable to be implemented in the aforementioned main control unit 105. Another part is conceivable to be implemented in the aforementioned suppression device 202. The flow steps of an embodiment of these two parts are shown in Figures 4a and 4b, respectively. As shown in Fig. 4a, the master device 105, in step 400, sets the network connection device to restart after the time φ is the time. The master device 105 then has two possible process paths. First, after the time arrives, the flow proceeds to step 401, the main control portion triggers the restarting device 301 via the connection mechanism 50 to cause the network connection devices 103, 104 to restart, and then the flow returns to step 400 to continue setting at the time. It is time to restart the network connection device. In other words, the master device 105 periodically restarts the network connection devices 103, 104 in a cycle of T! time. The other path after step 400 is to suppress the message "Don't restart the network connection device" sent from the suppression device 202 before the arrival of the time, and the flow therefore proceeds to step 403, and the master device 105 cancels the last step. The restart schedule set by 401, the flow then returns to step 400. • The new setting restarts the network connection device after a few hours. As shown in Figure 4b, the suppression device 202 begins by first issuing a "no restart network connection device" suppression message to the master device 105 in step 410. Please note that, as can be seen from Figure 4, the suppression devices 202 and 12 200807950 master device 105 will therefore be synchronized from μ. Next, the suppressing means 202 sets a suppression message to the master device 105 "Don't restart the network connection" after the time 2 is set from the winter neck τ field in step 411. Next, when the η time arrives, the flow returns to step 410 to suppress the device. A message "Do not restart the network connection 奘罟 褒" is issued to the master device 〇5. In other words, the suppression means 202 adjusts the transmission suppression message to the main control unit 105 in the I time period. In fact, since steps 41〇 and 411 丄疋 continue to repeat the loop, the flow of Fig. 4b may also start from step 411. It can be seen from the above description that σ i τ / / 戈 i2 are suitably smaller than Τ!, and the network connection devices 103, 104 are all under normal operation 1, the main control device 105 is suppressed by the suppression device 2 〇 2 Not to be enemies in the meeting. However, if one of the network connection devices 103, 104 is exhausted or goes to eight places, the network is exhausted. If the operation is stopped, since the main control device 1〇5 cannot receive the suppression message from the WAN 2, the network connection devices 103 and 104 will be restarted as scheduled by the 攸王控装置1〇5. When the network connection 103, 104 resumes operation 4 due to the restart, the master device 1〇5 will continue to be "made" without restarting. Please be aware that the period and the Ding 2 should have an appropriate # distance to the network, ^ ^ transmission delay, restart required (four) (four) factor into the cycle of the size of the Ding or Τ to the big J The characteristics of the network connection device are dead and the next week is likely to be resolved. The period of the suppression and suppression of the main control device 105 restarting device 202 may be a parameter that is respectively determined by the main control device 1〇, the device 2〇2, or set to 1〇5 and the suppression may also be the suppression device plus 13 200807950 Dynamically specify the master device to cycle with T!. The embodiments shown in Figures 3a and 3b are very similar to the embodiments of Figures 2a and 2b, respectively, in that (1) the suppression device 106 is located on the regional network 10; and (2) the WAN 20 is increased. A response device 203. The operation of the main control unit 105 is as shown in Fig. 4a, and the operation of the suppression unit 106 operates in accordance with the flow of Fig. 4c. As shown, the suppression device 106 also operates at a T2 cycle that is suitably less than a period of time. First, the suppression device 106 issues a "Do not restart the network connection device" suppression message to the master device 105 in step 420. Please note that the primary purpose of this suppression message is to synchronize the suppression device 106 with the master device 105. Next, the suppressing means 106 sets, in step 421, a suppression message to the master device 105 that "Do not restart the network connection device" after the time Τ2. Next, when Τ2 time arrives, the flow proceeds to step 422 where the suppression device 106 first sends an inquiry message to the response device 203 on the wide area network 20 (e.g., using ping or utilizing a proprietary TCP/IP packet) if the response device 203 has Returning the reply message, the flow returns to step 420 and the suppression device 106 then sends a suppression message to the master device 105. Conversely, when a certain network connection device 103 or 104 runs out of resources or enters a deadlock and stops operating, the inquiry or reply message cannot be successfully delivered, and the suppression device 106 returns to the step without receiving the response message. 421, reschedule and try again after T2 time. Since the suppression device 106 does not issue a suppression message to the master device 105, the master device 14 200807950 will trigger the restart device 301 as scheduled to restart the network connection devices 103, 104. In other words, the suppression means 106 basically also periodically sends a suppression message to the main control unit 105 in a period of T2 time, and the main difference between the embodiment in which the suppression means 202 is located in the wide area network 20 is that the suppression means 202 arrives and sends out in a cycle. Before suppressing the message, the response device 203 is used to determine whether the network connection device 103, 104 is still operating normally. In fact, since steps 420, 421, and 422 are repeated loops, the flow of Fig. 4c may also be started by step 421 or 422. The response device 203 is also a general computing device and is typically targeted to certain specific, very stable websites or servers on the wide area network 20 (as long as they support a query protocol like ping), so there is no need to additionally Erected or provided. Please note that the foregoing embodiments are relatively simple to reboot the system and method, and the same reboot system and method can be extended to more complex environments. For example, the suppression device 202 located in the wide area network 20 shown in Fig. 2a can actually broadcast or transmit a suppression message to a plurality of master devices 105 located at different clients to suppress the restart of the master devices 105. Another possibility is that the same user can connect two WANs at the same time to provide mutual backup. Then the user can have two master devices and at least two restart devices to be responsible for the two. The WAN network connection device is restarted, and the two WANs will need their respective suppression devices to suppress the two master devices respectively. For such an application environment, a suppression device located in the local area network may also be used to broadcast or transmit a suppression message to the two suppression devices (if they are located in the same regional network), and the two The wide area network also has its own response device for the suppression device to interrogate via the two wide area networks, respectively. The features and spirit of the present invention are more clearly described in the above detailed description of the preferred embodiments, and the scope of the present invention is not limited by the above-described preferred embodiments. On the contrary, the purpose is to cover a variety of changes and equivalence arrangements within the scope of the patent application to which this creative is intended. [Simple description of the diagram] Figure 1a shows a simplified schematic diagram of a two-way network communication model. Figure lb is a schematic illustration of a conventional system for restarting a remote network connection device. #第lc图 shows a system schematic diagram of a conventional periodic restart network connection device. Figure 2a is a schematic representation of a first embodiment of a system in accordance with the present invention. Figure 2b is a schematic representation of a second embodiment of the system in accordance with the present invention. Figure 3a shows a schematic diagram of a third embodiment of the system according to the invention, 16 200807950. Figure 3b is a schematic representation of a fourth embodiment of the system in accordance with the present invention. Figure 4a is a flow chart showing the implementation of the method of the present invention in a master device. Figure 4b is a flow diagram of the method of the present invention implemented in a distal suppression device. Figure 4c is a flow diagram of the method of the present invention implemented in a proximal suppression device. [Main component symbol description] 10 Regional network 20 Wide area line 30 Control network 40 Power supply 50 Connection mechanism 101, 102 Operation device 103, 104 Network connection device 105 Main control device 106 Suppression device 201 Operation device 202 Suppression device 203 Response device 301 Restarting devices 400-402 process steps 410, 411 process step 17