TWI342142B - Methods and systems for routing protocal - Google Patents

Description

P52950052TW 25185twf.doc/n 九、發明說明： 【發明所屬之技術領域】 本發明大體上關於在無線多跳網路中為主機分派位 址。本發明進一步關於為無線網狀/中繼網路提供路由協 定。 【先前技術】 無線多跳網路通常是沒有基站支援的無線網路，例如行動 自組織網路（mobile ad-hoc network，MANET)、感測器網路 或無線辑脱網路（Wireless me§^netw〇rk，。 站的情況下，無線網路提供較高的靈活性，但局)限 用針對基礎設施設計的現有驗方法。目而，無❹跳網路 中的位址分派可能要更加複雜才能維持效率和唯一性。 傳統上，主機可靜態地或動態地配置其位址。在靜態配 置中:用戶可能需要事先獲得可用位址並手動配置所述位 址然而’要用戶在靈活的多跳網路中進行手動配置可能不 2理。另—種選擇是動態配置，其允許主機例如藉由使用 (Dynamic Host Configuration Protocol > DHCP)來動態地配置位址。 — DHCP是集中式分派機制，其部署至少—侧定词服器 以I理中的所有位址。圖1緣示dhcp系統中的基本 二及八又換的消息。如圖所示’存在兩個DHcp伺服器 合、12和一個用於DHCP伺服器12的即〇>中繼器13。 田新主機14加入多跳無線網路時，其廣播〇isc〇ver 1342142 P52950052TW 25185twf.doc/n .. 消息以請求—位址。網路中除DHCP中繼器13之外的所有 主機均將幫助轉播此消息，直到其到達DHCp伺服器u、 :12為止。DHCP中繼器13將把所述消息直接_給°〇11〇> • 舰^2。DHCP伺服器u、12中的每—者在接收到 DHCP—Discover消息時均將分派一位址，並將dhcp 〇ffer 消息連同所述位址發送回所述新主機14。因此，新主機Μ 可，收多個gHCP一Offer消息以從中進行選擇。新主機14 • 接著廣播DHCP—Request消息，以將其選擇（例如，伺服器 12)告知所有伺服器U、12。所選擇的伺服器12將檢查分 派的地址是否真的可用，且如果可用，那麼向新主機^發 送DHCP_Ack。此時，新主機14可將分派的位址配置成^ 位址，並開始進行重複位址檢測（Duplicate Adda's Deteetion ’ 祕—中沒有其他主機使用此 位址。主機Μ在租f Uease time)之後藉由 =HCP一Request回收所述地址’其中所述租賃時間I由伺服 器12指派以通知主機η可使用分派的位址之多長時間。如 • 果所述地址仍然可用，那麼伺服器12將向主機14發送 DHCP一Ack消息。否則，將發送DHCp—Nack消息且主機 14需要再次從廣播DHCP—Discover開始執行整個程式。 考慮到DHCP的消息流，已知在動態配置中，通常必 須進行廣播以獲得位址或者檢測位址衝突。然而，在盔線多 跳網路中制廣播是非”貴的，且可能會造成廣播風暴 題。 、 在已知的位址分派機制中，自我配置是-種分散式位址 P52950052TW 25185twf.doc/n 分派機制，其中主機可根據不同因素依靠自身來配置其位 址，戶斤述因素例如為其硬體包址(S· Cheshire、B. Aboba以 及 Ε· Guttman 的 Dynamic Configuration of IPv4 Link-Local Addresses » draft-ietf.zeroconf-ipv4-linklocal-17.txt » IETF Zeroconf Working Group，2004 年 7 月；S· Thomson 以及 T. Narten 的 IPv6 Stateless Address Autoconfiguration，RFC 2462，1998 年 12 月）或者瞋機編號（c. E. Perkins、J. T. Malinen、R· Wakikawa、E. M. Belding-Royer 以及 Y· Sun 的 IP Address Autoconfiguration for Ad Hoc Networks， draft-ietfmanet-autoconf-01.txt ，IETF MANET Working Group ’ 2000年7月）。由於自我配置的位址是由各個主機 產生的，因而必須藉由DAD來檢查所述地址以證實其並未 被其他主機佔用。因此’仍然必須進行廣播以檢測位址衝突。 MANETconi( S. Nesargi 以及 R. Prakash 的隐视如沾 Configuration of Hosts in a Mobile Ad Hoc Network ^ INFOCOM 2002)也是MANET中的一種藉由使所有主機記 錄已使用的位址之分散式地址分派方法，其中新主機從其一 個鄰居處獲得地址’且所述鄰居找出不存在其記錄+的位 址’並要求MANET中的其他所有主機予以證實。如果所述 位址未被其絲何线制，那麼其將被指派給所述新主 機。在MANETconf中，由於需要從 實，因而仍然需要進行廣播。 預言式地址分浓（H. Zhou、L Μ Νί丨、；3» λ/r… Μ. Νι 以及 M.W.Mutka 的 P_et add· an0cation f〇r _ 嫩师心， 1342142 P52950052TW 25185twf.doc/n INFOCOM 2003)試圖找出一函數，其可針對每個主機產生 唯一的位址序列，所述位址序列將被依次指派。因此，每個 主機可依照其位址序列的順序直接指派位址，而無需詢問其 他主機。然而，要在多個序列之間維持唯一性可能比較複 雜，且當位址空間較小時難以實現唯一性。因此，預言式位 址分派局限於具有較大位址空間的MANET。 在美國專利公開案第US2004/0174904A1號“MethodP52950052TW 25185twf.doc/n IX. Description of the Invention: TECHNICAL FIELD OF THE INVENTION The present invention generally relates to assigning a address to a host in a wireless multi-hop network. The invention further relates to providing a routing protocol for a wireless mesh/relay network. [Prior Art] A wireless multi-hop network is usually a wireless network without base station support, such as a mobile ad-hoc network (MANET), a sensor network, or a wireless network (Wireless me§). ^netw〇rk, in the case of the station, the wireless network provides greater flexibility, but the bureaux) limits the existing methods of design for infrastructure. As such, address assignments in a non-bounce network may be more complex to maintain efficiency and uniqueness. Traditionally, a host can configure its address statically or dynamically. In a static configuration: the user may need to obtain the available address in advance and manually configure the address. However, it may not be reasonable for the user to manually configure it in a flexible multi-hop network. Another option is dynamic configuration, which allows the host to dynamically configure the address, for example by using (Dynamic Host Configuration Protocol > DHCP). – DHCP is a centralized dispatch mechanism that deploys at least one of the addresses in the server. Figure 1 shows the basic two and eight exchange messages in the dhcp system. As shown, there are two DHcp servers, 12 and one instant > repeater 13 for the DHCP server 12. When Tianxin host 14 joins the multi-hop wireless network, it broadcasts 〇isc〇ver 1342142 P52950052TW 25185twf.doc/n .. message to request-address. All hosts in the network except the DHCP Repeater 13 will help relay this message until it reaches the DHCp server u, :12. The DHCP repeater 13 will direct the message to the ship ^2〇> Each of the DHCP servers u, 12 will dispatch a single address upon receipt of the DHCP-Discover message and send a dhcp 〇ffer message back to the new host 14 along with the address. Therefore, the new host can receive multiple gHCP-Offer messages to select from. The new host 14 • then broadcasts a DHCP_Request message to inform all of the servers U, 12 of its selection (e.g., server 12). The selected server 12 will check if the assigned address is really available and, if available, send a DHCP_Ack to the new host. At this point, the new host 14 can configure the assigned address as a ^ address and begin repeated address detection (Duplicate Adda's Deteetion ' Secret - no other host uses this address. Host Μ after renting f Uease time) The address is reclaimed by =HCP_Request' wherein the lease time I is assigned by the server 12 to inform the host n how long the assigned address can be used. If the address is still available, the server 12 will send a DHCP-Ack message to the host 14. Otherwise, the DHCp-Nack message will be sent and the host 14 will need to start executing the entire program again from the broadcast DHCP-Discover. Considering the message flow of DHCP, it is known that in dynamic configuration, it is usually necessary to broadcast to obtain an address or detect an address conflict. However, broadcasting in the helmet-line multi-hop network is not expensive, and may cause broadcast storm problems. In the known address allocation mechanism, self-configuration is a kind of distributed address P52950052TW 25185twf.doc/ n Dispatch mechanism, in which the host can configure its address according to different factors, such as its hardware address (S·Cheshire, B. Aboba, and Gu· Guttman's Dynamic Configuration of IPv4 Link-Local Addresses) » draft-ietf.zeroconf-ipv4-linklocal-17.txt » IETF Zeroconf Working Group, July 2004; S. Thomson and T. Narten's IPv6 Stateless Address Autoconfiguration, RFC 2462, December 1998) or downtime number (c. E. Perkins, JT Malinen, R. Wakikawa, EM Belding-Royer, and Y. Sun's IP Address Autoconfiguration for Ad Hoc Networks, draft-ietfmanet-autoconf-01.txt, IETF MANET Working Group' July 2000 Since the self-configured address is generated by each host, the address must be checked by DAD to verify that it is not occupied by other hosts. Therefore, 'still must be broadcast to detect address conflicts. MANETconi (S. Nesargi and R. Prakash's Vision, such as Configuration of Hosts in a Mobile Ad Hoc Network ^ INFOCOM 2002) is also a way of making all hosts in MANET A decentralized address assignment method that records the used address, where the new host obtains the address 'from one of its neighbors' and the neighbor finds that there is no record of its record +' and asks all other hosts in the MANET to verify. If the address is not made by its wire, it will be assigned to the new host. In MANETconf, since it needs to be true, it still needs to be broadcast. The prophetic address is thick (H. Zhou, L Μ Νί丨,;3» λ/r... Μ. Νι and MWMutka's P_et add· an0cation f〇r _ tender teacher, 1342142 P52950052TW 25185twf.doc/n INFOCOM 2003) trying to find a function that can be The host generates a unique sequence of addresses that will be assigned in sequence. Therefore, each host can directly assign an address in the order of its address sequence without having to ask other hosts. However, maintaining uniqueness between multiple sequences can be complex, and it is difficult to achieve uniqueness when the address space is small. Therefore, predictive address assignment is limited to MANETs with larger address spaces. In US Patent Publication No. US2004/0174904A1 "Method

of allocating IP address and detecting duplication of IP address in an ad-hoc network environment’，中，每個主機均保存歷史 表矛DAD》主機的歷史表儲存主機的位址歷史資訊。DA。 表儲存所有的鄰居位址資訊。在鄰居之間週期 ‘‘祖L〇，，l“更新歸表。當域需要位址時，盆料 鄰居發送請求消息。如果某個鄰居在其歸表令找到此地 麼其將向主機發送含有可用消息之回應消息，而其他In the case of the allocating IP address and detecting duplication of IP address in an ad-hoc network environment, each host saves the historical table spear DAD host history table storage host address history information. DA. The table stores all neighbor address information. In the period between neighbors, '''''''''''''''''''''''''''''''''''''''''''ssssssssssssssssssssssssssssssssssssssssssssssss Available message response message, while others

盆表中找到此地址則保持沉默。如果主機從 =處接收到回應消息，那麼其發送請求消息連同此位址 =便再:人進行重複檢•重複檢_過程將繼續進行N欠， ==任何回應消息為止。此種方法不涉及在整個 =t 可在網财造成大量錢額外開銷。 中使用==:=吏用的路由協定是自組織網路 協定、前瞻定分類成反應式 按需距_量(祕。^^==1= 1342142 P52950052TW 25185twf.doc/n 可允許-發送主機向其所有鄰居廣播路由請求，每 將所述路由請求轉播給其鄰居，依此類推，直到找到目的地 主機為止。使时瞻式協定（例如最佳化鏈路狀態路由 (Optimized Link State Routing，OLSR))的主機输㈣如 性地交換其賴，《剛输^鱗== 後，混合式協定（例如區域路由協定（以取如 Protocol ZRP))組合反應式協定和前瞻式協定，其 在特定區域内使用鏈路狀態的·***換來維持路9 ZRP的域歸使収赋蚊㈣問位於不同 ^中的域。因此，财這些自組織齡均需要缝他主 ，進行廣泛的資訊錄赠配置路由路徑。這可能是奸· ^的’因為其可導致網路中的額外闕增加且等待時間= 【發明内容】 因此，本發明的實施例將提供—種改進的無線網狀/中 =的路由協定’(例如)可藉由使 ^參考發送主機和目的地域的位絲配置路 現所述改進的路由協議。 田峪仫而只 述的53^°其他優點，且根據所實施的並概括描 路態路由_ 的方與_域之間路由資料 包括使用、及：：主機均具有唯一的數值位址，所述方法 匕括.使用敎機和目的地主機#每_者的唯—數值位址 P52950052TW 25185twf.doc/n 來確疋源主機和目的地主機的最近共同上代主機；從源主 機向所述最近共同上代主機發送資料；從最近共同上代主 機向目的地主機發送資料；以及在目的地主機處處理資料。 進一步根據本發明，提供一種使用有狀態路由協定在包 各夕個主機的網路中在源主機與目的地主機之間路由資料 的方去，其中每個主機均具有唯一的數值位址，所述方法包 括：記錄相鄰主機的位i止；以及向最接近目的地主機的相鄰 主機發送資料。 ^再進-步根據本發明，提供一種使用無狀態路由協定在 多個具有唯一數值位址的聯網主機之間路由資 述系統包括：發駐機，其祕使域和㈣地主機的斤 者的唯-數錄絲確定最近朗上代域，並向所述 取近共同上代主機發送資料；以及目的地域，其從所述最 近共同上代主機接收資料，並處理所述資料。 少又進一步根據本發明，提供一種使用有狀態路由協定在 夕個具有唯-數值位址的聯網主機之間路由資料的系統，所 述系4*包括·送主機’其用以記錄相鄰主機的數值位址； ^及目的地域，其用以接㈣料；其情述發送主機首先 向最接近目的地主機的相鄰主機發送資料。 本發明的額外特徵和優點部分地將細下描述中蘭 日2地職職描述巾容易相，或可11由實踐本發 於所附申請專利範财特別指出的元件和 -且口來μ現和獲得本發明的特徵和優點。 將瞭解’以上概括描述和以下詳細描述均只是示範性的 1342142 P52950052TW 25185twf.d〇c/n 機分派位址（步驟S23)，其中所分派的位址是非根主機的 位址乘以一質數，所述質數從非根主機的位址的最大質因數 開始且依次接著非根主機分派的前一個質數。也就是說，非 根主機可分派這樣的位址，所述位址是其自身的位址乘以從 其自身位址的录U因數開始的質數。因此，如圖3A所示， 對於具有位址{2}的主機B，2的最大質因數是2，所以其可 向主機F、G......指派的位址次序是{2*2=4, 2*3=6, ...}， 其中主機B是主機f、G......的父代’且相應地，主機F、 G......是主機B的子代。相似地，對於具有位址{6丨的主機 (G)’ 6的最大質因數是3，所以其可向主機k、l、M...... 指派的地址次序是{6*3=18, 6*5=30, 6*7=42, .·.}，其中主 機G是主機K、L、Μ......的父代’且相應地，主機κ、L、 Μ•…·.是主機G的子代。基於此質數位址分派過程，每個位 址僅可由單個主機產生。進而，每個主機可藉由用其位址除 以其位址的最大質因數來確定其父代主機的位址。 基於圖2的質數編號位址分派方法，可建立如圖3八所 示的邏輯位址分派樹，以向每個主機分派唯一的位址。在實 踐中’每個主被分派-個位址之後可在彡跳無線網路中 四處移動和/或離開網路。因此’邏輯位址分派樹中的相鄰 關係可能與真實的網路拓撲的不同。圖4八說明具有用於圖 3Α所示的主機的128位址空間的示範性網路拓撲，且說明 質數編號DHCP的消息流。這些消息可由無線 的任何路由協定攜帶。網路中的每個主機均具備位址記錄 41以用於記錄已由此主機分派的位址，以及迴圈列表幻以' 12 1342142 P52950052TW 25185twf.doc/n 用於記錄先前由此主機分派但當前未由其分派的地址。 圖4B說明圖4A所示的任一主機的示範性結構4〇〇。 參看圖4B，主機400可包含一個或一個以上以下元件：炱 少一個中央處理單元（centmi pr〇cessing unjt，cpu) 402， 其經配置以執行電腦程式指令以執行各種過程和方法；隨機 存取§己憶體（random access memory，RAM) 404 以及唯讀 記憶體（rehd only memory，ROM) 406，其經配置以存取和 儲存資訊以及電腦程式指令；記憶體4〇8，其用以儲存資科 和貧訊；一個或一個以上資料庫41〇，其可儲存表、列表或 其他資料結構；一個或一個以上1/〇裝置412 ;—個或一個 以上介面414 個或一個以上天線416等。這些組件中的 每一者均為本行業中眾所周知的，且將不作進一步論述。 圖5說明用於為新主機獲得位址的方法的流程圖。如圖 4A和圖5所示，當新主機n加入網路時，其向其鄰居主機 廣播用於發現鄰居的消息（例如，DHCp_Disc〇ver請求消息） (步驟S51)，如DHCP中定義的那樣。如果沒有鄰居主機， 且因此未接收到任何回應，那麼已知新主機N是根，且為 其配置位址1。否則，如同在此實例中，在從新移動主機N 接收到請求時，鄰居主機L、B和F中的每一者不是將所述 消息轉發到整個網路，是改為根騎述賊編號位址分派 方法指派位址（步驟S52)。圖6進一步說明用於基於上述 質，編號位址分派方法指派位址的方法的流程圖其中如果 在郴居主機的迴圈列表42中存在有位址，那麼鄰居主機指 派其迴圈列表中的最小位址（步驟S61)。如果迴圈列表42 13 1342142 P52950052TW 25185twf.doc/n 中沒有位址，那麼鄰居主機藉由用質數乘以其自身的位址來 產生位址（A)(步驟S62) ’且如果此產生的位址（A)不 大於位址空間，那麼鄰居主機指派此位址（A)(步驟S63); 否則，將DHCPJDiscover請求消息轉播到其父代（步驟 S64) ’且父代主機指派一位址（步驟S65)。在此實例中， 假設鄰居主機L、B、F的迴圈列表中沒有任何位址，則鄰 居主機F (位址=4)指派地址= 4x2 = 8，鄰居主機B (位 址=2)指派地址= 2x5 = 10 ’且鄰居主機(位址= 30)指 ——-、 派地址= 30x5=150。此處，主機B用其位址乘以5 ,因為 其先前用其地址乘以2和3以分別向主機F和G (圖3A) 指派位址’且5是下一個質數。關於主機L進行的位址指派， 鄰居主機L產生的位址= 150大於位址空間128。因此，鄰 居主機L向其父代主機G (圖3A)發送DHCP_Discover請 求消息以尋求幫助，且父代主機G為鄰居主機L產生位址 = 6x11 = 66。 再次參看圖5，在指派了位址之後，鄰居主機l、B、F 中的每一者均向新主機（N)發送提供位址的消息（例如， DHCP—Offer消息），即’所述消息中嵌入有諸如位址661〇,8 (步驟S53)。因此，主機N分別從主機L (或G)、B和F 接收具有三個可用位址66、10和8的三個DHCP Offer消 息。請注意，如果主機N未接收到任何回應（即，j)HCP Offer 消息）’那麼就知道主機N為根主機’且向其指派地址1。 在接收到所述三個DHCP一Offer消息之後，主機N選出其中 一個位址，且在DHCP_Request消息中廣播其選擇（步驟 1342142 P52950052TW 25185twf.doc/nFinding this address in the basin table remains silent. If the host receives a response message from =, then it sends a request message along with this address = then: the person repeats the check • the repeat check _ the process will continue with N owed, == any response message. This method does not involve a large amount of money overhead in the entire network. The routing protocol used in ==:= is a self-organizing network protocol, and the forward-looking classification into a reactive on-demand distance _ quantity (secret. ^^==1= 1342142 P52950052TW 25185twf.doc/n allowable - send host Broadcast routing requests to all of its neighbors, each relaying the routing request to its neighbors, and so on, until the destination host is found. Enables a proactive protocol (eg, optimized Link State Routing, OLSR)) The host loses (4) if it is sexually exchanged, "just after the scale ==, the hybrid agreement (such as the regional routing agreement (such as Protocol ZRP)) combines the reactive protocol and the forward-looking agreement, The use of the link state of the sexual exchange in a specific area to maintain the road 9 ZRP domain is responsible for receiving mosquitoes (4) to ask the domain located in different ^. Therefore, these self-organized ages need to sew the other, for a wide range of information records The configuration routing path. This may be a 'because' it can cause additional enthalpy in the network to increase and wait time = [Summary] Therefore, embodiments of the present invention will provide an improved wireless mesh/middle = routing protocol' For example, the improved routing protocol can be configured by referring to the bit line configuration of the transmitting host and the destination domain. Tian Hao only describes the other advantages of the 53^°, and according to the implementation and general description of the routing The routing information between the _ party and the _ domain includes the use, and:: The host has a unique numeric address, and the method includes: the use of the downtime and the destination host # _ each of the _ only numerical address P52950052TW 25185twf .doc/n to determine the most recent common generation host of the source host and the destination host; send data from the source host to the most recent common generation host; send data from the most recent common generation host to the destination host; and at the destination host Further processing according to the present invention provides a method for routing data between a source host and a destination host in a network of packets of each host using a stateful routing protocol, wherein each host has a unique numeric bit. Address, the method comprising: recording a bit of an adjacent host; and transmitting the data to a neighboring host that is closest to the destination host. ^Re-step according to the present invention, providing a Using a stateless routing protocol to route a system of information between a plurality of networked hosts having unique numerical addresses includes: a resident machine, the secret domain of the host and the (four) host's only number of recorded disks determine the nearest domain And transmitting data to the near-common generation host; and a destination domain that receives data from the most recent common generation host and processes the data. Less and further in accordance with the present invention, providing a stateful routing protocol on the evening A system for routing data between networked hosts having unique-numeric addresses, the system 4* includes a sending host 'used to record a numerical address of an adjacent host; ^ and a destination domain, which is used to connect (four) materials The sent host first sends the data to the neighboring host that is closest to the destination host. Additional features and advantages of the present invention will be described in part in the description of the elements of the present invention, or may be made by the practice of the elements specifically recited in the appended claims. And obtaining the features and advantages of the present invention. It will be understood that the above general description and the following detailed description are merely exemplary 1342142 P52950052TW 25185 twf.d〇c/n machine assignment addresses (step S23), where the assigned address is the address of the non-root host multiplied by a prime number. The prime number begins with the largest prime factor of the address of the non-root host and in turn precedes the previous prime number assigned by the non-root host. That is, a non-root host can dispatch an address that is its own address multiplied by a prime number starting from the recorded U factor of its own address. Therefore, as shown in FIG. 3A, for the host B having the address {2}, the maximum prime factor of 2 is 2, so the address order that it can assign to the hosts F, G, ... is {2* 2=4, 2*3=6, ...}, where host B is the parent of host f, G... and correspondingly, host F, G... is host B Children. Similarly, for the host (G) '6 with the address {6 的, the maximum prime factor is 3, so the address order that it can assign to the hosts k, 1, M... is {6*3= 18, 6*5=30, 6*7=42, .....}, where host G is the parent of host K, L, Μ... and correspondingly, host κ, L, Μ • ..... is the child of the host G. Based on this prime address assignment process, each address can only be generated by a single host. In turn, each host can determine the address of its parent host by dividing its address by the maximum prime factor of its address. Based on the prime numbered address assignment method of Figure 2, a logical address assignment tree as shown in Figure 3 can be created to assign a unique address to each host. In practice, 'each master is assigned an address to move around and/or leave the network in the hopping wireless network. Therefore, the neighbor relationship in the logical address assignment tree may be different from the real network topology. Figure 4 illustrates an exemplary network topology with a 128-bit address space for the host shown in Figure 3, and illustrates a prime numbered DHCP message flow. These messages can be carried by any routing protocol in the wireless. Each host in the network has an address record 41 for recording the address that has been dispatched by this host, and the loop list is illusory with ' 12 1342142 P52950052TW 25185twf.doc/n for recording previously assigned by this host but The address that is currently not assigned by it. FIG. 4B illustrates an exemplary structure of any of the hosts shown in FIG. 4A. Referring to FIG. 4B, host 400 can include one or more of the following elements: a central processing unit (centre pr〇cessing unjt, cpu) 402 configured to execute computer program instructions to perform various processes and methods; random access § Random access memory (RAM) 404 and read only memory (ROM) 406 configured to access and store information and computer program instructions; memory 4〇8 for storage One or more databases 41, which may store tables, lists or other data structures; one or more 1/〇 devices 412; one or more interfaces 414 or more antennas 416, etc. . Each of these components is well known in the art and will not be discussed further. Figure 5 illustrates a flow diagram of a method for obtaining an address for a new host. As shown in FIG. 4A and FIG. 5, when the new host n joins the network, it broadcasts a message for discovering the neighbor (for example, a DHCp_Disc〇ver request message) to its neighbor host (step S51), as defined in DHCP. . If there is no neighbor host and therefore no response is received, the new host N is known to be the root and is configured with address 1 for it. Otherwise, as in this example, when a request is received from the new mobile host N, each of the neighbor hosts L, B, and F does not forward the message to the entire network, instead it is changed to the root riding thief number bit. The address assignment method assigns an address (step S52). Figure 6 further illustrates a flow diagram of a method for assigning an address based on the quality, numbered address assignment method, wherein if there is an address in the loop list 42 of the host, then the neighbor host assigns its list of loops The minimum address (step S61). If there is no address in the loop list 42 13 1342142 P52950052TW 25185twf.doc/n, the neighbor host generates the address (A) by multiplying the prime number by its own address (step S62) 'and if this is generated The address (A) is not greater than the address space, then the neighbor host assigns this address (A) (step S63); otherwise, the DHCPJDiscover request message is relayed to its parent (step S64) 'and the parent host assigns an address ( Step S65). In this example, assuming that there are no addresses in the loop list of neighbor hosts L, B, and F, neighbor host F (address = 4) assigns address = 4x2 = 8, neighbor host B (address = 2) assignment Address = 2x5 = 10 ' and the neighbor host (address = 30) means ---, the address = 30x5 = 150. Here, Host B multiplies its address by 5 because it previously multiplies its address by 2 and 3 to assign addresses ' to hosts F and G (Fig. 3A), respectively, and 5 is the next prime number. Regarding the address assignment by the host L, the address generated by the neighbor host L = 150 is greater than the address space 128. Therefore, the neighbor host L sends a DHCP_Discover request message to its parent host G (Fig. 3A) for help, and the parent host G generates the address for the neighbor host L = 6x11 = 66. Referring again to FIG. 5, after the address is assigned, each of the neighbor hosts 1, B, F sends a message providing the address (eg, a DHCP_Offer message) to the new host (N), ie, An address such as the address 661 〇, 8 is embedded in the message (step S53). Thus, host N receives three DHCP Offer messages with three available addresses 66, 10, and 8 from hosts L (or G), B, and F, respectively. Note that if host N does not receive any response (i.e., j) HCP Offer message), then host N is known to be the root host and address 1 is assigned to it. After receiving the three DHCP-Offer messages, the host N selects one of the addresses and broadcasts its selection in the DHCP_Request message (step 1342142 P52950052TW 25185twf.doc/n

^1)^通知所選主_於位址之成功分派。較佳地，主機 '、位址8以防止樹增長得過快。其鄰居主機L、B ΐΐ 發自DHCP—Request消息，但自主機l轉播到父代 機=的m最後’所選駐機F更新其位址記錄 ’並向主機N發送確認消息（例如，DHCP—Ack消息） =進行證實（步驟S55)。因此，所制的方法藉由單跳廣 播而不是整個網路的廣播來實現對pHcp〇ffer和 DHCP_Request消息的發送。 為了避免位址茂漏問題，主機應當執行優雅離開 i==epa伽1)，這意味著其應當在準備離開網路或關 機時釋放其位址。舉例來說，假設圖4A中的主機κ即將離 開’因此主機Κ向其父代域G發送縣其位址的消自 例來說，DHCP_Release消息）。接著，主機G在宜迴㈣ 表42中記錄錄18，且迴圈列表42中的位址在以後指派 時具有較南的優先權。如果根主機準備離開，那料應 知其最高下代（deseendent)成為根主機，所述下代^ 中具有最大分派位址。舉例來說，主機Μ是圖3A中#， A的最高下代。 T很主機 在無線通信網路中，無線電鏈路可能會斷開或者 機可能會異常離開，例如無線電鏈路在=、， DHCP一Release消息的情況下斷開。為了解決此類事/， 個主機設定-個計時器，所述計時器在發送主機向 = 機發送消息（其可以是除DHCP—Release消息外的 時開始倒計時。如果計時器期滿且未接收到任何回應二卞 15 P52950052TW 25185twf.doc/i =為=的地主機是異常離開的主機。躲那些異常離開的主 4曰如、又，主機可在DHCP租賃時間後重複使用其位址， 果其父代主機也離開了網路，那麼所述位址可變成不可 库^用的。因此’在—個實施例中，根主機A週期性地 廣^用於進行迴_消息（例如，DHCPJleeyde m ^ =求所有域發送其當前狀態’其巾包含所指派的位址和 迴„藉由收魏有域態，根域A可重建位 =派樹’且向主機發送其新的迴_表，其中包含異常離 库祕主機。可在路由協議（例如，DSDV)的消息中捎帶所 二的DHCP—Recycle消息，這樣將不會使位址迴圈增加的 根據另-實施例’源主機可配置一到達任何目的地主機 者路由路徑’而無需在網路中的主機之間交換路由資訊。或 來；^主機使用其自身的位址連同目的地主機的目的地位址 诉確定其最近制上代的健。此最近制上代主機是在從 f主機和目的地主機兩者開始向上穿過網路樹時遇到的第 個共用父代主機。-旦識別出最近共同上代主機，便可將 路由路徑配置成^健段，即從駐制最近朗上代主 機且接著再到目的地主機。舉例來說，參照圖3B所示的 位址分派樹系統實例，假設源主機是p且目的地主 那麼最近共同上代主機是主機F。 圖7說明源主機藉以確定最近共同上代主機的位址的 法的流程圖。首先，源主機獲得其自身的Ip位址和目的 主機的IP位址（S71)，且根據所述資訊獲得源位址和目 1342142 P52950052TW 25185twf.doc/n 的地位址。源位址和目的地位址可以是由源主機提取的Ip 位址（特別是子網路）的一部分。舉例來說，如果源主機P 的IP位址是10.5M00.16，那麼源主機位址可能是16。此 外，如果目的地主機Q的IP位址是1〇 51 1〇〇 36，那麼目 的地主機位址可能是36。所屬領域的技術人員現在將認識 到，有若干種方式來從一 IP位址來確定源位址和目的地位 址，其中IP位址包含（但不限於）使用相關表或對正位址 運行算術演算法。 旦源主機成功地獲得源位址和目的地位址，其便繼續 計，每一者的質因數分解序列（S72)。這涉及對源和目的地 的母個位址進行完全的因數分解，以使得該序列中的所有數 值在彼此相乘時可產生原始位址。可藉由以下數學演算法來 為質因數分解各列建立模式：/^印⑻^^。,/^2…，〜）^^1)^Notifies the successful assignment of the selected master_address. Preferably, host ', address 8 to prevent the tree from growing too fast. Its neighbor hosts L, B ΐΐ are sent from the DHCP-Request message, but from the host l to the parent device = m last 'selected station F updates its address record' and send an acknowledgment message to the host N (for example, DHCP - Ack message) = Confirmation is performed (step S55). Therefore, the method is implemented to transmit PHcp〇ffer and DHCP_Request messages by single-hop broadcast instead of broadcast over the entire network. In order to avoid address leakage problems, the host should perform gracefully leaving i==epa gamma 1), which means that it should release its address when it is ready to leave the network or shut down. For example, assume that the host κ in Figure 4A is about to leave 'so the host sends its county address address to its parent domain G, DHCP_Release message). Next, host G records 18 in the Echo (4) table 42, and the address in the loop list 42 has a more south priority when assigned later. If the root host is ready to leave, it is expected that its highest descendant (deseendent) becomes the root host, and the next generation ^ has the largest assigned address. For example, the host Μ is the highest generation of #, A in Figure 3A. T is very host In the wireless communication network, the radio link may be disconnected or the machine may leave abnormally, for example, the radio link is disconnected in the case of =, DHCP-Release message. In order to solve such a problem, a host sets a timer that starts a countdown when the sending host sends a message to the machine (which may be other than the DHCP-Release message. If the timer expires and is not received) To any response, the host of the P52950052TW 25185twf.doc/i = = is the abnormally left host. To avoid those abnormally left masters, for example, the host can reuse its address after the DHCP lease time. The parent host also leaves the network, so the address can become unusable. Therefore, in an embodiment, the root host A is periodically used to perform a back_message (for example, DHCPJleeyde m ^ = ask all domains to send their current state 'the towel contains the assigned address and back „ by receiving the domain state, the root domain A can rebuild the bit = send the tree' and send its new back_table to the host, where Contains an exception from the library. The DHCP-Recycle message can be piggybacked in the message of the routing protocol (for example, DSDV), so that the address loop will not be increased. According to another embodiment, the source host can be configured. Arrive at any destination host Routing paths' without exchanging routing information between hosts in the network. Or; the host uses its own address along with the destination address of the destination host to determine the health of its most recent generation. The host is the first shared parent host encountered when going through the network tree from both the f host and the destination host. Once the most recent common generation host is identified, the routing path can be configured as a health segment. That is, from the camping agent to the nearest host and then to the destination host. For example, referring to the address assignment tree system example shown in FIG. 3B, it is assumed that the source host is p and the destination master is the most recent common generation host is the host F. Figure 7 is a flow chart showing the method by which the source host determines the address of the most recent common generation host. First, the source host obtains its own IP address and the IP address of the destination host (S71), and obtains the source bit based on the information. The address of the address and destination 1342142 P52950052TW 25185twf.doc/n. The source and destination addresses can be part of the Ip address (especially the subnet) extracted by the source host. For example, If the IP address of the source host P is 10.5M00.16, then the source host address may be 16. In addition, if the IP address of the destination host Q is 1〇51 1〇〇36, then the destination host address may be Yes 36. Those skilled in the art will now recognize that there are several ways to determine the source and destination addresses from an IP address that includes, but is not limited to, the use of related tables or alignment bits. The arithmetic algorithm is run. Once the source host successfully obtains the source address and the destination address, it continues to calculate the prime factor decomposition sequence of each (S72). This involves performing the parent address of the source and destination. Complete factorization so that all values in the sequence can produce the original address when multiplied by each other. The following mathematical algorithms can be used to create a pattern for the prime factor decomposition columns: /^(8)^^. , /^2...,~)^

p, < PiVi < / \\p, =n 八 T ，且，=。 。繼續先前實例，藉由對源主機ρ的 位址16進行完全的因數分解來對源主機ρ進行質因數分 解，其將為PfSeq(16) = (2, 2, 2, 2)。請注意，2是無法被進 一步因數分解的質數，且2x2x2x2 =〗6，這是源主機p的位 址。類似地’目的地主機Q的質因數分解將為ρβ叫(36) = (2, 2, 3, 3)。 旦计算出源主機位址和目的地主機位址的質因數分 解序列，源主機便接著計算那些質因數分解序列的最長共同 首碼（S73 )。為了汁异最長共同首碼，源主機採用兩個質因 數刀解序列（此處為piSeq(16) = (2, 2, 2, 2)和邮邮36) = (2, 2, 3)) ’並從序列開頭開始比較序列内的每個數字。舉例 17 1342142p, < PiVi < / \\p, =n 八 T , and, =. . Continuing with the previous example, the source host ρ is subjected to a prime factorization by fully factoring the address 16 of the source host ρ, which will be PfSeq(16) = (2, 2, 2, 2). Note that 2 is a prime number that cannot be factored further, and 2x2x2x2 = 〖6, which is the address of the source host p. Similarly, the prime factor decomposition of the destination host Q will be ρβ = (36) = (2, 2, 3, 3). Once the prime factor decomposition sequence of the source host address and the destination host address is calculated, the source host then calculates the longest common first code of those prime factor decomposition sequences (S73). For the longest common first code, the source host uses two prime factor scaling sequences (here piSeq(16) = (2, 2, 2, 2) and postal 36) = (2, 2, 3)) 'And compare each number in the sequence from the beginning of the sequence. Example 17 1342142

P52950052TW 25185twf.doc/n 有狀態和無狀態路由協定。 無狀態路由協定的一個實施例是源路由方法，其中源主 機負責設置整個路由路徑。無狀態路由協定的另—實施例是 逐跳路由妓，射沿著從駐翻目的魅機的路程的每 個主機均依靠其自身找到到達目的地主機的路一 主機。 ‘很龈無狀悲路由協定的源路由方法的虛擬 碼，其中源主機（咖）找到到達目的地主機（如）的整個 =路#。圖9說明用於實施所述源路由方法的流程圖。參 源主機首先藉由前述方法（圖7)找到關於目的地 义句丨攻近’、同上代主機（S91)。將圖3B的樹結構用於當 ^ 點P具有位址16，目的地節點Q具有位址％， 且取近共同上代域是F，其位址為4。 (叫可藉由用=為主=:路由路徑的第-段 的最大質因重… + 一特疋主機自身的位址除以其自身位址 源主機ΐ所述特定主機的父代主機。舉例來說， 路由列表二3ί816/2由 ，這對應於主機Ν。因此，P52950052TW 25185twf.doc/n Stateful and stateless routing protocols. One embodiment of a stateless routing protocol is the source routing method, in which the source host is responsible for setting the entire routing path. Another embodiment of a stateless routing protocol is a hop-by-hop routing, where each host that travels along the way from the destination is found to rely on itself to find the way to the destination host. ‘The virtual code of the source routing method of the sinless routing protocol, where the source host (cafe) finds the entire =way# to the destination host (eg). Figure 9 illustrates a flow diagram for implementing the source routing method. The source host first finds the destination semantics by the aforementioned method (Fig. 7), and the host of the same generation (S91). The tree structure of Fig. 3B is used when the point P has the address 16, the destination node Q has the address %, and the near common domain is F, and its address is 4. (The maximum quality factor of the first segment of the routing path can be called by = = = + The address of a special host itself is divided by its own address source host 父 the parent of the specific host. For example, the routing list 2 is 3 ί 816/2 by, which corresponds to the host Ν. Therefore,

主機，所簡主H於8並不對應於最近共同上代 機 N 主機的位址除以其自身位址mu是=由用其父代 現在為(Mh機4確實對 U4Z142 P52950052TW 25185twf.doc/n 因而完成路由難的第—段。® 8中虛擬碼的卜7行對應 於圖9中的步驟S91和S92。 接下來’源主機建立路由路徑的第二段，其中其向路由 列ctf加+代域，而不是像在第—段巾那樣添加父代主機 在此触巾，源域首級帛目馳线位址除以 最近^上代主機位址來計算質因數分解序列（如之前所描 =在㈣實例t ’這將是36 (目的地主機q)除以4 (最 二同，代域〉的結果（其為9)的質因數分解序列。9 Ϊ 解序列為PfSeq(9) = (3, 3)。源主機遞迴地進行此 trf列’以計算出路由列表的第二段。在此 貫f中’駐_最近制上代域位址4乘以計算得的質 ==中的第一元素(其為3)來計算路由列表中的 取近共同上代主機制下—主機（其為4χ3 = =〇二ΓΓ附加到路由列表中，所述糊 g⑴，千 下來，社機用最近添加的〇主機（位 =產因數分解序列中的第二元素(其 ί得Hist的最後主機。更明確地說，這個乘 ώ ν對應於主機Q，且將這一位址附加到 ==1戶:述路_見在為(8,4,12,36)。因為已經 機9，故完成路由協定的第二段，且甚而，質 ==中不再有其他元素可用。源主機可將此路由ί f ，的資料-起發送（例如，在標頭中)，使得 ϊί擬:ΐϋ:個1主機均知道將資料路由到何處(s94)。圖8 中虛擬碼的8— 17行對應於圖9中的步驟S93和S94。 20 1342142 P52950052TW 25I85twf.d〇c/n 川询不虚擬碼 ~ 』職據無狀態路由協定的_ 跳式方*中，每個具有位址w的當前主 =， 址_的地主機來計算出將資料發 =到_下-主機。雖然此方法與源路由方法她可能會增 ‘母==計Μ荷，但由於無需連 表’因而減少額外開銷。 說日賴於實施逐跳式方法的流程圖。參看圖11， 先藉由前述方法（圖7)計算其自身與目的地主 =的位址小於其自扣位址（S112處為是），那麼备前^ θ知道其位於路由路徑的第—段中n其使^ 將其自身的父代（SU3)^S1 一刼作序列對應於圖10中虛擬碼的丨—4行。 的、 如果’減’最近共同上代主機的㈣大於於 身的位址（S112處為否），那麼去前 1 .、，、自 二段中，遞機於路由路徑的第 機择m田〜^丨，、—個子代主機（S114)。當前主 主機的位址除以其自身的位址並計算 發送到那個子代主機。接著，1 如上所述），來確定將封包 刀鱗列中的最小元素並用其乘以其自身的健。纟 代主機’且當前主機將向所述子代主疋子 _4的操作序列對應於圖1〇令虛擬二：_ 主楼Hff线將㈣發_下—域（父触機或子代 機）’所述下一主機便成為當前主機，且重複所述方法直 21 1342142 P52950052TW 25185twf.doc/n 到到達目的地主機為止。 繼續圖3B所示的先前實例，源主機p計算出其自身和 目的地主機Q的最近共同上代主機是主機F (sm)。由於The host, the simple master H8 does not correspond to the address of the most recent common generation N host divided by its own address mu is = by its parent now (Mh machine 4 indeed to U4Z142 P52950052TW 25185twf.doc/n Thus, the difficult segment of the routing is completed. The row 7 of the virtual code in the 8 corresponds to steps S91 and S92 in Fig. 9. Next, the 'source host establishes the second segment of the routing path, where it adds + to the routing column ctf Instead of adding the parent host to the touchpad as in the first paragraph, the source domain first-level target address is divided by the last-generation host address to calculate the prime factor decomposition sequence (as described earlier). = in (4) instance t 'This will be the prime factor decomposition sequence of 36 (destination host q) divided by 4 (the most identical, domain > result). 9 Ϊ The solution sequence is PfSeq(9) = (3, 3) The source host recursively performs this trf column ' to calculate the second segment of the route list. In this f, the 'resident_recently generated domain address 4 is multiplied by the calculated quality== The first element (which is 3) is used to calculate the proximity of the common generation in the routing list - the host (which is 4 χ 3 = = 〇 ΓΓ attached to the routing column In the table, the paste g(1), thousand down, the social machine uses the recently added 〇 host (bit = the second element in the production factor decomposition sequence (its ί got the last host of Hist. More specifically, this multiplication ώ ν corresponds On the host Q, and append this address to the ==1 household: the description of the road_seeing is (8, 4, 12, 36). Because already the machine 9, the second paragraph of the routing agreement is completed, and even, There are no more elements available in quality==. The source host can send this route ί f , the data is sent (for example, in the header), so that: ΐϋ: 1 host knows how to route data to At (s94), the 8-17 lines of the virtual code in Fig. 8 correspond to steps S93 and S94 in Fig. 9. 20 1342142 P52950052TW 25I85twf.d〇c/n The query is not virtual code ~ 』 _ hopping party *, each host with the address of the current main = address, address _ to calculate the data sent = to _ lower - host. Although this method and source routing method she may increase 'mother = = Calculate the load, but because there is no need to connect the table' thus reduce the overhead. The day depends on the flow chart of the implementation of the hop-by-hop method. Referring to Figure 11, first by the foregoing The method (Fig. 7) calculates that the address of its own and the destination master = is smaller than its self-deducted address (YES at S112), then the pre-preparation ^ θ knows that it is located in the first segment of the routing path, which makes ^ itself The parent (SU3)^S1 sequence corresponds to the 丨-4 line of the virtual code in Fig. 10. If the 'minus' recent common generation host (4) is larger than the body address (No at S112), Then go to the first 1, and, from the second paragraph, the first machine in the routing path is the choice of m field ~ ^ 丨, - a child host (S114). The current host's address is divided by its own address and calculated to be sent to that child. Next, 1 (described above), it is determined that the smallest element in the scale of the packet is multiplied and multiplied by its own health. The operation sequence of the descendant host 'and the current host will be directed to the child master 疋4 corresponds to Figure 1 虚拟 虚拟 virtual 2: _ main building Hff line will (4) send _ down - domain (parent touch or child machine) 'The next host becomes the current host, and the method is repeated until the destination host is reached 21 1342142 P52950052TW 25185twf.doc/n. Continuing with the previous example shown in Figure 3B, source host p calculates that its most recent common generation host of destination host Q is host F (sm). due to

F的位址4小於P的位址16，因而p知道路由路徑目前位The address 4 of F is smaller than the address 16 of P, so p knows the current position of the routing path.

於第一段中（S112處為是）’且因此將資料發送到其父代主 機N (SU3)。最近共同上代主機F的位址4也小於主機N 的位址8 (S112處為是）’且因此主機N類似地將資料轉發 到^父代主機F (SI13)。域|7計算出其位址4大於或等 於最近共同上代主機位址4 (其為最近共同上代主機） (S112處為否），且因此其知道其位於路由路徑的第二段 中，並且必須將資料發送到子代主機（SU4)。為了計算出 正確的子代主機’主機F首先用目的地主機位址除以其自身 的位址以產生36/4 = 9。接下來，主機F計算出9的質因數 =解序列’其為pfSeq(9) = (3,3)。接下來，主機μ所述計 算出的9的質因數分解序列上選出最小的元素，1為3。最 後’主機F用此最小it素乘以其自身的位址，以產生適當的 =代主機位址4x3= 12，其對應於主機〇。因此，主機？將 貧料轉發到主機0。主機〇 _地計算出其位於環路的第 一段中，魏行類似於主機F的那些計# ，物件的子代主機位址36。因此，主機F 料;; 發到位址36 (對應於主機Q)，其為最终目的地。 雖然上述無狀態路由協定可在無需記錄任何資訊的情 配置路由雜，但路由路徑可能較長，此乃因為主機僅 使用位址分派_分支來計算路鱗徑。畴在額外鍵 22 1342142 P52950052TW 25185twf. doc/π 不屬於位址分派樹中，圖3β中用虛線繪示了這些鏈路的實 例如上所述，當添加新主機時，可將其連接到若干主機， 但'堇由-個域對其進行定址，_纽的域成為新主機 的父代主機。因此’網路巾可能有未由位址分賴映射的額 外鏈路可用。有狀態路由協定使用—個或—伽上這些 鏈路來計算較矩的路由。 一 在有狀態路由協定中’每個主機均記錄其鄰居的位址， i 局部拓撲。因此，有狀態路由龄使用逐跳式 方法，其中母個主機僅計算下一跳躍。 碼。二1Vi示根據有狀態路由協定的逐跳式方法的虛擬 ^圖說制於實施此逐跳式方法的流程圖 ==先計算其自身以及其所有鄰居相對於目的地主 A、力數（greatest common divisor，GCD) ( s 131)。 機接著餘麵sm。當前主 如果大於的話（S132 一 A a 1 以大於其自身的gcd。 主機火二 2處為疋），那麼具有最大C3CD的鄰居 ，在其鄰居中間具有最大二==反’ 麼當前主機將眘+:14、¥ ^ & (S132處為否），那 -10杆料* + 托賴其父代主機（S134)。虛擬觸8 ^應於步_2 — “否，，以及⑽。 只如圖3B中所不的先前實例源主機p首先計算其 23 丄丄吁z P52950052TW 25185twf.d〇c/n SI ^ 自身的GCD是4,主機N的⑽是4, ==是12’且主機K的GCD是2。因此，由於 =:=機大於其他任何連接到p的主機 因此，故瑕靠近目的地主機Q的主機是主機〇。In the first paragraph (YES at S112) and thus the data is sent to its parent host N (SU3). The address 4 of the recent common generation host F is also smaller than the address 8 of the host N (YES at S112) and thus the host N similarly forwards the data to the parent host F (SI13). Domain|7 calculates that its address 4 is greater than or equal to the most recent common generation host address 4 (which is the most recent common generation host) (no at S112), and therefore it knows that it is in the second segment of the routing path and must Send the data to the child host (SU4). In order to calculate the correct child host, host F first divides the destination host address by its own address to produce 36/4 = 9. Next, the host F calculates a prime factor = solution sequence of 9 which is pfSeq(9) = (3, 3). Next, the smallest element is selected on the prime factor decomposition sequence of 9 calculated by the host μ, and 1 is 3. The last 'host F is multiplied by its own address with this smallest integer to produce the appropriate = generation host address 4x3 = 12, which corresponds to the host 〇. So the host? Forward the poor material to host 0. The host 〇 _ calculates its position in the first segment of the loop, which is similar to those of host F, and the child's child host address 36. Therefore, the host F material; is sent to the address 36 (corresponding to the host Q), which is the final destination. Although the above stateless routing protocol can configure routing miscellaneous without recording any information, the routing path may be longer because the host only uses the address allocation_branch to calculate the path scale. The domain is in the extra key 22 1342142 P52950052TW 25185twf. doc/π is not in the address assignment tree. The examples of these links are shown by dashed lines in Figure 3β. As mentioned above, when adding a new host, it can be connected to several hosts. , but '堇 is addressed by a domain, and the domain of _New becomes the parent of the new host. Therefore, the 'network towel may have an extra link that is not mapped by the address. Stateful routing protocols use - or - gamma to link these to calculate a more rigid route. In a stateful routing protocol, 'each host records its neighbor's address, i local topology. Therefore, the stateful routing age uses a hop-by-hop approach where the parent hosts only compute the next hop. code. The virtual flow diagram of the hop-by-hop method according to the stateful routing protocol is described in the flow chart of implementing the hop-by-hop method. == Calculate itself and all its neighbors relative to the destination master A, the number of forces (greatest common divisor) , GCD) ( s 131). The machine then has the remaining surface sm. If the current master is greater than (the S132 A A 1 is greater than its own gcd. The host fire is 2 疋), then the neighbor with the largest C3CD has the largest two == anti' in the middle of its neighbor. +:14, ¥ ^ & (No at S132), that -10 shots * + Trusted its parent host (S134). The virtual touch 8 ^ should be in step_2 - "No,, and (10). Only the previous example source host p as shown in Figure 3B first calculates its 23 丄丄 z z P52950052TW 25185twf.d〇c/n SI ^ itself GCD is 4, host N's (10) is 4, == is 12' and host K's GCD is 2. Therefore, since the =:= machine is larger than any other host connected to p, it is close to the host of destination host Q. Is the host 〇.

=算==;，主機_)。主_ r其有鄰居相對於目的地域Q的GCD =);0的咖是W的GCD是4，P_是4,= count ==;, host _). The main _r has a neighbor with respect to the destination domain Q of GCD =); 0's coffee is W's GCD is 4, P_ is 4,

2 !\GCD是36。因此’主機Q具有最大的GCD (sm ’且，0將資料發送到作為目的地之主機Q 躍數目”因路由。使用這種方法，從p到達Q的跳 躍數目疋2 ’而在無狀態路由協定中跳躍數目為4。 正如現在所屬領域的技術人員將容純解，可以用多種 ==的實施例。舉例來說’-個實施例可在無線 另-實施例可在無線感測器網路中實施。在此實施例 中’感測H g卩職集資料並將其路由到感卿網路内的匿節 點（Sink node)。所述匯節點可以是根主機，其負責向其他 所，感測ϋ節點指派地址。依照根據本發明的實施例，感測 器節點可計算到達其他感·節喊到達匯節關路由路 徑。 一實施例可在WiMAX無線通信網路中實施，WiMAX 無線通信網路是一種行動的基於多跳中繼的（m〇bile multi-hop relay based，MMR)網路。WiMAX 網路包括基站 24 1342142 P52950052TW 25185twf.doc/n 和-個或-個以上中繼站，所述中繼站連接一個或一個以上 • 行動訂戶，以擴展網路的範圍和/或提高處理量等。在此實 ：施例中，WiMAX基站或錨節點（anCh0rnode)可以是根主 • 機’其可向巾舰指祕址。以此方式，巾獅和基站可在 無需交換路由資訊的情況下向其他任何站路由資料。圖14 說明此實施例的兩個實例。在一個實例中，相應主機2 !\GCD is 36. Therefore 'Host Q has the largest GCD (sm 'and, 0 sends the data to the number of host Q hops as the destination) due to routing. Using this method, the number of hops from p to Q 疋 2 ' while in stateless routing The number of hops in the agreement is 4. As will now be understood by those skilled in the art, a variety of embodiments can be used. For example, '-an embodiment can be used in the wireless-other embodiment can be in the wireless sensor network. In the embodiment, the H g卩 job set data is sensed and routed to the sink node in the Senqing network. The sink node can be the root host, which is responsible for the other nodes. The sensing node assigns an address. According to an embodiment in accordance with the invention, the sensor node can calculate the arrival of other senses to reach the meeting route. An embodiment can be implemented in a WiMAX wireless communication network, WiMAX wireless The communication network is a mobile multi-hop relay based (MMR) network. The WiMAX network includes a base station 24 1342142 P52950052TW 25185twf.doc/n and one or more relay stations. The relay station connection One or more • Mobile subscribers to extend the range of the network and/or increase the throughput, etc. In this case, the WiMAX base station or anchor node (anCh0rnode) can be the root host machine. In this way, the lion and the base station can route data to any other station without exchanging routing information. Figure 14 illustrates two examples of this embodiment. In one example, the corresponding host

(corresponding host ’ CH )位於 MMR 網路外部，且盥 MMR • 網路内部的移動訂戶（觀）通信。在第二實例中，均位於 MMRJ網路⑽發送和接收軸訂戶主機（娜、祕2)彼 此通信。來自圖K的這兩個實例均顯示根據圖8和圖9的 無狀態路由協定源路由方法。 在又一實施例中’MMR網路除了上述中繼站之外進一 步包括行動中繼站（mobile rday stati〇ns，)。這些mrs 藉由中繼站連接到網路’並向移動訂戶提供網路連接性。由 於MRS是需要與固定中繼站的連接性的行動裝置，因而可 能必顧MRS的移動而將MRS從—辦_移交到另一 個巾繼站。舉例來說，圖15繪示部署德流排上使得匯流 排上的行動訂戶（mobile subscriber，MS )在行進時可推并 網際網路訪問的MRS。當匯流排行進時，可能在匯流排Z 動到不同覆蓋區域時變成必須在不同固定中繼站之間切 換。為了進行所述操作，MRS首先決定需要從其原始^繼 站（relay stati〇n ’ Rs)移交到目標中繼站。在圖15中，目 標中繼站位址是RS 10,且原始中繼站位址是Rs 2。接下 來’ MRS遵照前述方法來計算原始中繼站Rs 2與目標㈣ 25 1342142 P52950052TW 25185twf.doc/n 間的示範性網路拓撲的圖。 ' ®4B是圖4A所示的類型的示範性主機的方塊圖。 : 圖5說明根據本發明實施例在無線多跳網路中按照編 號位址分派法為新主機獲得位址的方法的流程圖。 圖6說明根據本發明實施例在無線多跳網路中按照編 •號位址分派法向新主機指派位址的方法的流程圖。 、 圖7說明計算最近共同上代節點的流程圖。 圖8是實施根據本發明實施例的源路由方法的虛擬碼^ ® 圖9說明實施所述源路由方法的流程圖。 ^ 圖1〇是實施根據本發明實施例的逐跳路由方法的虛擬 碼。 圖11說明實施所述逐跳路由方法的流程圖。 圖12是實施根據本發明實施例的有狀態逐跳路由協定 的虛擬碼。 圖13說明實施有狀態逐跳路由協定的流程圖。 圖14說明根據本發明實施例的WiMAX網路。 • 圖15說明根據本發明實施例的行動轉播站換手。 【主要元件符號說明】 11，12:伺服器 13:中繼器 14:主機 400··主機 402:中央處理單元 27 1342142 P52950052TW 25185twf.doc/n 404:隨機存取記憶體 406:唯讀記憶體 408:記憶體 410:資料庫 412: I/O 裝置 414:介面 416:天線(corresponding host ’ CH ) is located outside the MMR network and communicates with the mobile subscribers (views) within the MMR • network. In the second example, both the MMRJ network (10) transmit and receive axis subscriber hosts (Na, Secret 2) communicate with each other. Both of the examples from Figure K show the stateless routing protocol source routing method according to Figures 8 and 9. In yet another embodiment, the 'MMR network further includes a mobile relay station (mobile rday stati〇ns) in addition to the relay stations described above. These mrs connect to the network by the relay station' and provide network connectivity to mobile subscribers. Since the MRS is a mobile device that requires connectivity with a fixed relay station, it is possible to transfer the MRS from the office to another towel relay in consideration of the movement of the MRS. For example, FIG. 15 illustrates an MRS that is deployed on a traffic queue such that a mobile subscriber (MS) on the busbar can push and access the Internet while traveling. When the busbar travels, it may become necessary to switch between different fixed relay stations when the busbar Z moves to a different coverage area. In order to perform the operation, the MRS first decides that it needs to hand over from its original relay station (relay stati〇n ' Rs) to the target relay station. In Fig. 15, the target relay station address is RS 10, and the original relay station address is Rs 2. Next, the MRS follows the foregoing method to calculate a diagram of an exemplary network topology between the original relay station Rs 2 and the target (four) 25 1342142 P52950052TW 25185twf.doc/n. '®4B is a block diagram of an exemplary host of the type shown in Figure 4A. Figure 5 illustrates a flow diagram of a method for obtaining an address for a new host in accordance with a numbered address assignment method in a wireless multi-hop network in accordance with an embodiment of the present invention. 6 illustrates a flow diagram of a method of assigning an address to a new host in accordance with a coded address assignment in a wireless multi-hop network in accordance with an embodiment of the present invention. Figure 7 illustrates a flow chart for computing the most recent common generation node. 8 is a virtual code implementing a source routing method according to an embodiment of the present invention. FIG. 9 is a flowchart illustrating a method of implementing the source routing. Figure 1A is a virtual code implementing a hop-by-hop routing method in accordance with an embodiment of the present invention. Figure 11 illustrates a flow chart for implementing the hop-by-hop routing method. Figure 12 is a virtual code implementing a stateful hop-by-hop routing protocol in accordance with an embodiment of the present invention. Figure 13 illustrates a flow diagram for implementing a stateful hop-by-hop routing protocol. Figure 14 illustrates a WiMAX network in accordance with an embodiment of the present invention. • Figure 15 illustrates a mobile relay station handoff in accordance with an embodiment of the present invention. [Main component symbol description] 11, 12: Server 13: Repeater 14: Host 400 · Host 402: Central processing unit 27 1342142 P52950052TW 25185twf.doc/n 404: Random access memory 406: Read-only memory 408: Memory 410: Library 412: I/O Device 414: Interface 416: Antenna

Claims (1)

曰修正本 、申請專利範圍·· 1=路由資料的方法’其在包含多 的—源主機盥—曰沾从士加 Μ路中 定，/、 、 機之間使用一無狀態路由Μ :方=些主機中的每-者均具有-唯-數值位址: 值位址，來確域巾每—者_唯—數 代主機，其中該最近共同:;目主==::=同上 為該源主機的該唯一數值位址計算一第 數分解序列； Τπ第一質因 為該目的地域的該唯—數值位 貝因數分解序列；以及 异第一 使用該第一與該第二質因數分解序列以拉ώ 2公約數計算而計算出-最近共同上代主機唯 位址’從而確定該最近共同上代主機； 數值 將資料從該源主機發送到該最近共同 機；Ζ資料從該最近制上代域發送到該目的地主 在該目的地主機處處理該資料。 2.如申請專利範圍第“員所述的 與該第二質因數分解序列進—步包括：〃中使㈣第一 將該第一質因數分解序列與該第 進行比較以確定-最長制“ 因數分解序列 藉由將s亥最長共同首碼中的戶 所有數目一起相乘來計 29 1342142 100-3-3 算該最近共同上代主機唯—數值位址。 專利範圍第1項所述的方法，其中該無狀態路 由協疋的貫施是使用源路由。 4.亡申請專利範㈣3項所述的方法，其進—步包括·· ，由遞迴地添加至少—父代主機至—路由列表，來建 立從Ϊ社機到該最近共同上代主機的-第-路由段； ，由計算㈣目的地主機的該唯—數值位址除以該 同上代主機的該唯—數值位址的結果的一質因數 二解序列以向該路_表添加任何其魅機，來建立從該 敢近共同上代主機到該目的地主機的一第二路由段。 由^如u概圍第1項所述的方法，其中該無狀態路 協疋的實施是使用逐跳式路由。 一者t如中請專利範圍第5項所賴方法，其中該網路中的 田則主，接收資料以發送到該目的地主機，該當前主機： 確定其是位於一第一路由段中或是一第二路由段中； 如果其確定其位於該第一路由段中，那麼將該資料發 送到一父代主機； ' 如果其確定其位於該第二路由段中，那麼將該資料發 适到一子代主機；且 、 診咯ΐ中使用該目的地主機位址的該唯一數值位址除以 ^ =刖主機的該惟一數值位址的結果的一質因數分解序 來選擇該子代主機。 7 一種路由資料的系統，其在多個具有唯一數值位址的 9、祠主機之間使用—無狀態路由協定，所述系統包括： 30 1342142 100-3-3 一發送主機，其適於： 使用邊源主機和該目的地主機中每一者的該唯一數 值位址來確定一最近共同上代主機，且該最近共同上代主 機的確定藉由： 两该源主機的該唯一數值位址計算一第一質因 數分解序列；曰Revised, the scope of patent application··1=The method of routing data' It contains more than one source host 盥 曰 曰 士 士 士 士 士 , , , , , 使用 使用 使用 使用 使用 使用 方 方 方 方 方 方 方 方= Each of these hosts has a --only value address: a value address, to determine the domain towel - each - only - several generations of hosts, where the most recent common:; the main ==:: = the same as The unique numerical address of the source host calculates a first decomposition sequence; Τπ first quality because of the unique-valued bit factorization sequence of the destination domain; and the first first use of the first and the second prime factor decomposition The sequence is calculated by pulling the 2 common divisor to calculate - the most recent common generation host only address 'to determine the most recent common generation host; the value sends data from the source host to the nearest common machine; Ζ data from the nearest generation domain Send to the destination master to process the material at the destination host. 2. As described in the scope of the patent application, the second factor decomposition sequence includes: (4) first comparing the first prime factor decomposition sequence with the first to determine - the longest system The factorization sequence is calculated by multiplying the total number of households in the longest common first code of shai by 29 1342142 100-3-3 to calculate the most recent common generation host-only numerical address. The method of claim 1, wherein the stateless routing is performed using source routing. 4. The method described in the third paragraph of the patent application (4), the further step including: · recursively adding at least the parent host to the routing list to establish the slave to the nearest common generation host - a first-routing segment; , by calculating (iv) the unique-value address of the destination host divided by the prime-factor two-solution sequence of the result of the unique-value address of the same-generation host to add any to the way_table Its charm machine, to establish a second routing segment from the co-professional host to the destination host. The method of clause 1, wherein the stateless routing protocol is implemented using a hop-by-hop routing. One of the methods of claim 5, wherein the field owner of the network receives the data to send to the destination host, the current host: determines that it is located in a first routing segment or Is a second routing segment; if it determines that it is located in the first routing segment, then the data is sent to a parent host; 'If it determines that it is located in the second routing segment, then the data is appropriate Selecting the child to the child host; and selecting the child value by dividing the unique value address of the destination host address by the prime factor decomposition of the result of the unique numeric address of the ^=刖 host Host. A system for routing data, which is used between a plurality of 9. hosts having unique numerical addresses - a stateless routing protocol, the system comprising: 30 1342142 100-3-3 a transmitting host adapted to: Determining a most recent common generation host using the unique numeric address of each of the edge source host and the destination host, and determining the most recent common generation host by: calculating the unique numerical address of the two source hosts First prime factor decomposition sequence; 為該目的地主機的該唯一數值位址計算一第二 質因數分解序列；以及 使用該第一與該第二質因數分解序列以藉由一 位址㈣算出—最近共同上代主機唯一數值 位址，攸而確定該最近共同上代主機； 將資料發制該最近朗上餘機；以及 該㈣目的魅機，其贱最近朗上代主機接收並處理Calculating a second prime factor decomposition sequence for the unique numerical address of the destination host; and using the first and the second prime factor decomposition sequence to calculate by address (4) - the most recent common generation host unique numerical address , and determine the recent co-generation of the host; the data is issued to the nearest Lang Yi; and the (four) purpose of the charm machine, which was recently received and processed on behalf of the host 與該第二;進I:系統’其中使用該第-將該第-質因數分解序列與 進行比較以確定-最長共同首碼、二及—f因數分解序列 藉由將該最長共同首碼中的所 算該最近共同上代域唯—數值位址數目—起相乘，來計 9. 如申請專利範圍第7項所述的 由協定的實施是使用源路由。 ’、、’、"中邊無狀悲路 10. 如申請專利範圍第9 機適於： “的系統’其中該發送主 31 1342142 100-3-3 …藉由遞迴地添加至少一父代主機至—路由列表，來建 立從^源主機到該最近共同上代主機的一第一路由段，·且 藉由計算用該目的地主機的該唯一數值位址除以該 =近共同上代主機的該唯一數值位址的結果的一質因數 ^解序列以向該路由列表添加任何其他主機，來建立從該 最近共同上代主機到該目的地主機的一第二路由段。And the second; into the I: system 'where the first-to-the prime-factor decomposition sequence is compared to determine - the longest common first code, the second and the -f factor decomposition sequence by using the longest common first code The calculation of the nearest common domain domain-number-number of addresses is multiplied by 9. The implementation of the agreement as described in item 7 of the patent application is to use source routing. ',, ', " no trace of sadness in the middle 10. If the scope of application for patent ninth machine is suitable for: "system" where the sender 31 313422 100-3-3 ... by recursively adding at least one parent Substituting a host to a routing list to establish a first routing segment from the source host to the nearest common generation host, and dividing by the unique numeric address of the destination host by the = common common generation host A prime factor resolution sequence of the result of the unique numeric address to add any other host to the routing list to establish a second routing segment from the most recent common generation host to the destination host. 11·如申請專利範圍第7項所述的系統，其中使用逐跳 來實施該無狀態路由協定。 12.如申請專利範圍第11項所述的系統，其中該網路 的—當前主機接收資料以發送到該目的地主機，該當 機適於： 確定其是位於一第一路由段中或是一第二路由段中； 如果其確定其位於該第一路由段中，那麼將該資料發 送到一父代主機：且11. The system of claim 7, wherein the stateless routing protocol is implemented using hop-by-hop. 12. The system of claim 11, wherein the current host of the network receives data for transmission to the destination host, the trigger being adapted to: determine whether it is located in a first routing segment or In the second routing segment; if it is determined to be in the first routing segment, the data is sent to a parent host: 如果其確定其位於該第二路由段中，那麼將該資 知—子代主機； … 、 其中使用該目的地主機的該唯一數值位址除以該當 如主機的該唯一數值位址的結果的一質因數分解序列來 選擇該子代主機。 32If it is determined that it is located in the second routing segment, then the known-child host; ..., wherein the unique numeric address of the destination host is divided by the result of the unique numerical address such as the host A prime factor decomposition sequence is used to select the child host. 32