1320280 P52950066TW 22244twf.doc/006 九、發明說明： 【發明所屬之技術領域】 本發明是有關於一種網路封包傳送方法與裝置，且特 • 別是有關於一種切割封包的網路封包傳送方法與裝置。 Μ 【先前技術】 被動式乙太光纖網路(EPON: Ethernet Passive Optical Network)擁有高達1Gbps的頻寬，為共享傳輸媒體網路架 構，可大幅降低網路鋪設成本。由於其高速傳輸特性，所 以許多高速的應用，例如多媒體傳播、語音資料、網路電 話（IP Telephony，其中 IP 為 Internet Protoc〇1 之縮寫）、 網路電視(iptv)和高晝質電視(HDTV: ffigh_Deflniti()n τν) 等應用，均可在ΕΡΟΝ上實現。在下行方向，Ερ〇Ν使用 廣播(Broadcast)的方式，在上行方向，epon則使用分時 父割(TDM: Time-Division Multiplexing)的方式，使在共享 媒體的架構上，每個使用者都擁有自己可上傳的時段可 秦防止碰撞(Collision)的發生。 EPON系統可分為局端(〇LT: 〇ptical Une τ⑽_)和 用戶知(ONU : Optical Network Unit)。一個局端可同時管理 許多個用戶端。局端管理所有使用者的註冊和解註冊，並 統籌管理上行和下行的頻寬，使每個使用者都能符合苴所 簽I的合約内容。在下行方向，局端直接管理每個使用者 的=貝料傳送’因為是廣播的方式’所以不會發生碰撞；而 在上行方向，局端會根據每個用戶端所要求的頻寬，透過 1320280 P52950066TW 22244twf.doc/006 動態頻寬分配(DBA: Dynamic Bandwidth Allocation)的機1320280 P52950066TW 22244twf.doc/006 IX. Description of the Invention: [Technical Field] The present invention relates to a network packet transmission method and apparatus, and particularly to a network packet transmission method for a cutting packet and Device. Μ [Prior Art] EPON (Ethernet Passive Optical Network) has a bandwidth of up to 1 Gbps and is a shared transmission media network architecture, which can greatly reduce the cost of network deployment. Due to its high-speed transmission characteristics, many high-speed applications, such as multimedia communication, voice data, Internet telephony (IP Telephony, where IP is the abbreviation of Internet Protoc〇1), Internet TV (iptv) and high-definition TV (HDTV) Applications such as ffigh_Deflniti()n τν) can be implemented on ΕΡΟΝ. In the downlink direction, Ερ〇Ν uses the broadcast method. In the uplink direction, epon uses the Time-Division Multiplexing (TDM) method to make each user on the shared media architecture. Have a time period that you can upload to prevent Collision. The EPON system can be divided into a central office (〇LT: 〇ptical Une τ(10)_) and an ONU (Optical Network Unit). A central office can manage many clients at the same time. The central office manages the registration and deregistration of all users, and manages the upstream and downstream bandwidths in a coordinated manner so that each user can meet the contract content of the signed I. In the downstream direction, the central office directly manages each user's = bedding transmission 'because it is a broadcast mode' so there is no collision; in the uplink direction, the central office will pass the bandwidth required by each client. 1320280 P52950066TW 22244twf.doc/006 Dynamic Bandwidth Allocation (DBA: Dynamic Bandwidth Allocation)
制’計算每個用戶端可以擁有的上傳時間，以避免發生碰 撞。 X 局端會傳送門戶封包(GATE)給每個用戶端，其中帶有 - 允許(GRANT)資訊，亦即每個用戶端可上傳時間的資訊。 用戶端收到此封包後，會對其作處理，並在其指定的時間 内，傳送報告(REPORT)封包給局端，裡面載有其所有佇二 φ (QueUeS)内，準備傳送的封包長度的資訊。局端根據所收 到的所有用戶端的報告資訊，使用動態頻寬分配的機制， .來計算每個用戶端可以使用的上傳頻寬的長度，並計算出 每個用戶端可以上傳的時間，此資訊將包含在門戶封包的 允許資訊内。 由於局端給予的可上傳時間並不能真正反應用戶端 實際要傳送的封包的總長度，這會導致在剩餘的上傳時間 内，由於封包所需上傳時間大於剩餘的時間，而導致封包 無法傳送出去。這段剩餘時間會因為沒有封包傳送而完全 • 浪費掉。如圖ΙΑ * 1B所示，由區域網路端進來的^ P1-P10會經過進入排程器(Ingress Schedule〇而分別進入到 封包佇列(Queue)中，等待被傳送出。待可上傳時間到達， 則依輸出排程器(Egress Scheduler)來依序將封包ρι_料傳 送出去。然而當準備要傳送封包p5的時候，由於剩餘時 間不足’所以封包P5無法被傳送出去。 、 為了提高頻寬使用率，美國專利公 2__G5號提供一種界限報告(Threshold Rep()rting)方 1320280 P52950066TW 22244twf.doc/006 法。在此方法中，用戶端會向局端報告兩個數字，也就^ -個預設界限值以下完整封包的總長度LL，以及全部封= 賴長度LH’而局献配給用戶端的可上料間就是= A LH。如此雖然局端分配的可上傳㈣比較能反應用戶 、《實際需要的時間’但是若有需要優先傳送的控制用 產生’這些封包必須優先傳送，就會擾亂之前的計算盘八 配，此時健會有頻m上述的美國 ^ # 20030048805號另外提出一種方法，是在封包仵歹 長度合適的封包以盡量利用剩餘的可上傳時間， .===:運!又複雜，也不容易朗和“ 時間長度相付合的封包，如此仍然會有浪費的頻寬。 為了提高頻寬使用率，美國專 20050249497號也提供兩種方法。其 料，以充分彻局端分配的 :要;包組合， •，二種方法是前述界限報告方=二=發生。 予改為動匕、調整，然而一旦有高優先的控 送，仍然會擾亂之前的計算與分配，仍然會有頻:=傳 【發明内容】 本發明提供一種網路封包值 的頻寬浪費問題，同時避免先前技術的缺=解決先前技術 本發明另提供-種網路封包傳送裝置，以解決先前技 6 1320280 P52950066TW 22244twf.doc/006 何的頻寬浪費問題 ,^ 几几則歿術的缺點。 本發明之範例提出一種網路封 步:。首先檢查現在可上傳時間的;=否足包::列 :網路封包。若剩餘時間足夠，則直接傳』== ^餘時間不足而且—預設條件不成立，㈣ 傳時間。另一方面，若 扪等待下次可上 立，則將此網路封包切割為一先傳封$士$設條件成 時間傳送先傳封包，後傳‘== 存模 區域網路接收網路封包，暫存 。^ I存核組自 間輸出網路封包。切割控制===傳時 則控制信料㈣封包， 成立，則控制信號為等待狀態，====設條件不 條件成立，難制信號為 不足而且預設 於暫存模組與切·制$封包傳送切割器輕接 控制信號為傳送狀二::封包與控制信號。若 包至被動式乙太光纖網路。若控直接傳送網路封 包傳送切割器等到下次可上傳為，狀態’則封 制信號為切割狀態，則封包逆=麟封包。若控 先傳封包與後傳封包，使用:二封包切割為 封包則等待下次可上傳時^傳枝傳封包，後傳 1320280 P52950066TW 22244twf.doc/006 奉發明不需要自行排程，不改變 序，不需要複雜運算，僅使用簡罩;对包傳送次 使用率。使用上述網路封包吏=序二，高頻寬 f分配到的可上傳時間，和動態頻寬分配⑽理現 無關，因此即使出現較優先的封包需要 ）機制完全 網路封包傳送方法與裝置依朗用，不上述的 為讓本發明之上述特徵和優點能更^顯易懂:浪費。 舉較佳實施例’並配合所附圖式，作詳細說明如^了文特 【實施方式】 路封=是實 將待傳送封包切割成兩個封包，稱 ^ ]不足時， 包’然後賴餘時_送先傳封包，後傳封^等 =傳包的切割並非盲目進行™The system calculates the upload time that each client can have to avoid collisions. The X office will send a Portal Packet (GATE) to each client with the - GRANT message, which is the information that each client can upload time. After receiving the packet, the client will process it and send a report (REPORT) packet to the central office within its specified time, which contains all the packets in the φ (QueUeS) that are ready to be transmitted. Information. The central office uses the dynamic bandwidth allocation mechanism to calculate the length of the upload bandwidth that each client can use according to the received report information of all the clients, and calculates the time that each client can upload. The information will be included in the permission information for the portal package. Since the uploadable time given by the central office does not really reflect the total length of the packet actually transmitted by the UE, this will result in the packet being unable to be transmitted due to the uploading time required for the packet being greater than the remaining time during the remaining uploading time. This remaining time will be completely wasted because there is no packet transmission. As shown in Figure ΙΑ * 1B, ^ P1-P10 coming in from the local network end will enter the packet queue (Queue) and enter the packet queue, waiting to be transmitted. Waiting for upload time When it arrives, the packet is transmitted in order according to the Egress Scheduler. However, when the packet p5 is to be transmitted, the packet P5 cannot be transmitted because the remaining time is insufficient. Wide usage, US Patent 2__G5 provides a threshold report (Threshold Rep() rting) side 1320280 P52950066TW 22244twf.doc/006 method. In this method, the client will report two numbers to the central office, ie ^ The total length LL of the complete packet below the preset threshold value, and the total length of the package = LH' and the available capacity to the client are = A LH. Thus, although the central office can be uploaded (4), it can reflect the user. "The actual time required", but if there is a need to control the transmission of the priority to generate 'these packets must be transmitted preferentially, it will disturb the previous calculation disk eight, at this time there will be frequency m above the United States ^ # Another method is proposed in 20030048805, which is to pack the packet with the appropriate length to make the best use of the remaining uploadable time, .===: shipped! It is complicated, and it is not easy to and "time-length combined packets, so There will still be wasted bandwidth. In order to improve the bandwidth usage, the United States specializes in two methods, No. 20050249497. The materials are distributed in a full-fledged manner: • package combination, •, two methods are the aforementioned boundary reports. Fang=2=occurring. It is changed to move, adjust, but once there is high priority control, it will still disturb the previous calculation and allocation, there will still be frequency: = [invention] The present invention provides a network packet The bandwidth of the value is wasted, while avoiding the lack of prior art. The prior art further provides a network packet transmission device to solve the problem of bandwidth waste of the prior art 6 1320280 P52950066TW 22244twf.doc/006. The shortcomings of several techniques. The example of the present invention proposes a network sealing step: first check the current upload time; = no foot package:: column: network packet. If the remaining time is sufficient, directly 』== ^The remaining time is insufficient and - the default condition is not established, (4) the time is passed. On the other hand, if you wait for the next time to stand up, then cut the network packet into a first pass and then set the condition to the time. The first packet is transmitted, and then the '== memory local area network receives the network packet, temporarily stored. ^ I deposits the core group to output the network packet. Cutting control === transmission time control information (4) packet, established , the control signal is in the waiting state, ==== setting condition is unconditional, the difficult signal is insufficient, and the preset control module and the cutting system packet transmission cutter are connected to the control signal for transmission. Packet and control signals. If you pack to a passive Ethernet network. If the control direct transmission network packet transmission cutter waits until the next time it can be uploaded, the state 'the sealing signal is the cutting state, then the packet is reversed = the collar packet. If the control first transmits the packet and the post-transmission packet, use: the second packet is cut into packets, and the next time the device can be uploaded, the pass-through packet is transmitted. After the 1320280 P52950066TW 22244twf.doc/006, the invention does not need to schedule itself, and does not change the sequence. , does not require complex operations, only uses a simple mask; transmits the secondary usage rate to the packet. Using the above network packet 吏 = sequence 2, the uploadable time allocated to the high frequency width f is independent of the dynamic bandwidth allocation (10), so even if a preferential packet is required, the mechanism is completely network packet transmission method and device. What has not been described above is to make the above features and advantages of the present invention more comprehensible: waste. The preferred embodiment is described in detail with reference to the accompanying drawings, and the detailed description is as follows: [Embodiment] The road seal = is to cut the packet to be transmitted into two packets, and when it is insufficient, the package is then The rest of the time _ send the first pass packet, after the pass seal ^ etc. = the cut of the packet is not blindly carried out TM
圖2為本實關的封包__示意圖在本實施 中必須滿足三個條件才會切割目前待傳送的封包21〇。第 一，傳送封包210所需的時間必須大於現在可上傳時間的 剩餘時間220 ’否則無切割的必要。第二，剩餘時間22〇 必須足夠傳送封包21〇所屬之通訊協杨狀之最短封包 230。本實施例採用國際標準組織（IS〇: lmernati〇nai Orgamzation f0r Standardizati0n)的開放系統介面（〇SI: Open System Interface)架構，封包 21〇 符合 〇SI 第二層(L2) 1320280 P52950066TW 22244twf.doc/006 的乙太網路(Ethernet)通訊協定以及〇SI第三層(L3)的網際 網路通訊協定(IP: Internet Protocol)，因此最短封包230的 長度為乙太網路最小封包的64位元組(byte)加上網際網路 通訊協定的標頭(header)長度。第三，待傳送封包21〇的長 度必須大於設定值240。設定值240可由系統自動計算或 由使用者設定’可以是固定值’也可以根據封包21〇所屬 的通訊協定而決定。舉例而言，有些通訊協定傳遞的封包Figure 2 is a schematic diagram of the packet __ in this implementation. In this implementation, three conditions must be met to cut the packet 21 to be transmitted. First, the time required to transmit the packet 210 must be greater than the remaining time of the current uploadable time 220' or otherwise there is no need for cutting. Second, the remaining time 22〇 must be sufficient to transmit the shortest packet 230 of the communication association Yang to which the packet 21〇 belongs. This embodiment adopts the open system interface (〇SI: Open System Interface) architecture of the International Standards Organization (IS〇: lmernati〇nai Orgamzation f0r Standardizati0n), and the packet 21〇 conforms to the 〇SI second layer (L2) 1320280 P52950066TW 22244twf.doc/ 006 Ethernet protocol and 〇SI Layer 3 (L3) Internet Protocol (IP: Internet Protocol), so the minimum packet size 230 is the minimum 64-bit packet of the Ethernet. Group (byte) plus the length of the header of the Internet Protocol. Third, the length of the packet 21 to be transmitted must be greater than the set value 240. The set value 240 can be automatically calculated by the system or set by the user 'may be a fixed value' or can be determined according to the communication protocol to which the packet 21〇 belongs. For example, some packets passed by a communication protocol
都报小，為了效率起見，可透過設定值24〇使這些通訊協 定的封包不被切割。 圖3為本實施例的網路封包傳送方法流程圖。首先， 在步驟310等待局端分配的可上傳時間。可上傳時間到達 後，在步驟320檢查是否有封包需要傳送，如果沒有就回 到步驟310。如果有封包要傳送，流程會進入步驟33〇，檢 查現在可上__剩餘_是否足_送此封包 餘時間足夠，就進入步驟34〇，直接傳送封包，不 二=回ίΓ 320，檢查是否還有封包要傳送。若剩餘 間不足，流程進入步驟350。 接下來，在步驟35〇檢查剩餘時間是 定的最短封包，例如前述的最= 封包’流程回到步驟，等待下一 1320280 P52950066TW 22244twf.doc/006 程就進人步驟37G的切割程序，將待傳送封包切割 2 2與後傳封包。然後進入步驟380，使用剩餘時 間傳达先傳封包，後㈣包則料下次可上傳時間。 步驟36〇的檢查判斷主要是因應使用者需纟，在其他 實施例中，若使用者無此種需求，可省略步驟％〇。All are reported to be small. For efficiency reasons, the packets of these communication protocols can be cut without being cut through the set value. FIG. 3 is a flowchart of a network packet transmission method according to an embodiment of the present invention. First, in step 310, the uploadable time allocated by the central office is awaited. After the uploadable time has elapsed, it is checked in step 320 whether a packet needs to be transmitted, and if not, it returns to step 310. If there is a packet to be transmitted, the process will go to step 33〇, check whether it can be __ remaining_ is _ is enough to send the packet for the remaining time, then go to step 34〇, directly send the packet, not = Γ Γ 320, check whether There are also packets to be delivered. If the remaining amount is insufficient, the flow proceeds to step 350. Next, in step 35, it is checked that the remaining time is the shortest packet, for example, the above-mentioned most = packet 'flow returns to the step, waiting for the next 1320280 P52950066TW 22244twf.doc/006 to enter the cutting procedure of step 37G, which will be The transmission packet is cut 2 2 and the post-transmission packet. Then, proceeding to step 380, the remaining time is used to convey the first packet, and the (four) packet is expected to be uploaded next time. The check and judgement of step 36〇 is mainly in response to the user's need. In other embodiments, if the user does not have such a requirement, the step %〇 may be omitted.
圖4A為本實施例的封包切割示意圖，其中41〇為切 割前的待傳送封包，與侧分別為切割後的先傳封包 與後傳封包。由於乙太網路的封包不能切割，本實施例切 割第二層的IP封包。如圖4A所示，無論切割前或切割後， 每個封包都包括符合乙太網路通訊協定的第二層標頭部份 (411，421，431)、符合網際網路通訊協定的第三層標頭部份 (412, 422, 432)、酬載(payload)部分(413, 423, 433)、以及符 合乙太網路通訊協定的結尾部分(414, 424, 434)。本實施例 的結尾部分(414, 424, 434)為乙太網路通訊協定的封包檢 查序列(FCS: Frame Check Sequence)。 原待傳封包410的酬載部分413被切割為先傳封包 420的酬載部分423以及後傳封包430的酬載部分433,例 如酬載部分423及433可分別為原酬載部分413的前半段 和後半段。標頭421及431是根據原標頭411以及酬載413 的切割方式而產生，標頭422及432是根據原標頭412以 及酬載413的切割方式而產生。在本實施例中，乙太網路 標頭421及431的封包形態/長度(Type/Length)欄位必須修 改’其他攔位和原標頭411相同。網際網路標頭422及432 的總長度（Total Lengrti)、旗標（Flag)、切割位移 1320280 P52950066TW 22244twf.doc/006 (Fragmentation 0ffset)、和標頭檢查碼(如加㈤冰贿) 搁位必須修改，其他摘位和原標頭412相同。至於封包檢 查序列424及434是分別根據切割後的先傳封包42〇與後 傳封包430進行計算而產生。 、 為了充分利用上傳頻寬，最理想的切割方式是使傳送 先傳封包所需的時間正料於現在可上傳時_剩餘時 間，如圖4B所示，其中先傳封包為PM，而後傳封包為 • P5_2。將剩餘時間能傳送的資料長度扣除所有標頭與結尾 的長度之後，就能決定先傳封包可容納多少酬載，至於剩 .餘的酬載就放入後傳封包。以圖4A的封包41〇為例，'假 設封包的四個組成部分411-414的長度分別為Lh2、 Lm、LP、以及Lfcs，而剩餘時間所能傳送的資料長度為 Lrt，則先傳封包420的酬載長度為& = Lrt-LH2-LH3-Lfcs，後傳封包430的酬載長度為& =4A is a schematic diagram of packet cutting according to the embodiment, wherein 41〇 is a packet to be transmitted before cutting, and the side is a first-pass packet and a post-transport packet after cutting. Since the packet of the Ethernet cannot be cut, this embodiment cuts the IP packet of the second layer. As shown in FIG. 4A, each packet includes a second layer header portion (411, 421, 431) conforming to the Ethernet protocol, and a third network conforming to the Internet Protocol, whether before or after cutting. The layer header portion (412, 422, 432), the payload portion (413, 423, 433), and the end portion of the Ethernet protocol (414, 424, 434). The end portion (414, 424, 434) of this embodiment is a packet check sequence (FCS: Frame Check Sequence) of the Ethernet protocol. The payload portion 413 of the original packet to be transmitted 410 is cut into the payload portion 423 of the first transmission packet 420 and the payload portion 433 of the post-transmission packet 430. For example, the payload portions 423 and 433 may be the first half of the original payload portion 413, respectively. Segment and second half. The headers 421 and 431 are generated based on the cutting method of the original header 411 and the payload 413, and the headers 422 and 432 are generated based on the original header 412 and the cutting mode of the payload 413. In this embodiment, the Packet Type/Length field of the Ethernet headers 421 and 431 must be modified. The other blocks are the same as the original header 411. The total length (Total Lengrti), flag (Flag), cutting displacement 1320280 P52950066TW 22244twf.doc/006 (Fragmentation 0ffset), and header check code (such as plus (5) bribes) of the Internet gateway heads 422 and 432 must be Modified, the other picks are the same as the original header 412. The packet inspection sequences 424 and 434 are generated based on the calculation of the punctured pre-transmission packet 42 〇 and the post-transmission packet 430, respectively. In order to make full use of the upload bandwidth, the most ideal way to cut is to make the time required to transmit the first packet to be the same as the remaining time, as shown in Figure 4B, where the first packet is PM and then the packet is transmitted. For • P5_2. After deducting the length of all headers and endings from the length of the data that can be transmitted in the remaining time, it is possible to determine how much payload can be accommodated in the first packet, and the remaining payload is placed in the post-transfer packet. Taking the packet 41〇 of FIG. 4A as an example, 'assuming that the lengths of the four components 411-414 of the packet are Lh2, Lm, LP, and Lfcs, respectively, and the length of data that can be transmitted in the remaining time is Lrt, the packet is transmitted first. The payload length of 420 is & = Lrt-LH2-LH3-Lfcs, and the payload length of the post-transport packet 430 is &
Lp-Lpi 〇 後傳封包430必須符合所屬通訊協定所規定之最小可 送長度’例如乙太網路軌龄蚊封包不可小於64 位元組，若後傳封包430小於64位元組，必須以墊補資料 (padding)的方式加長至64位元組的最小可傳送長度。 圖5為本實施例節省上傳頻寬的示意圖，其中固定長 f是指封包當中除了酬載以外的部分，在本實施例中就^ 第二層標頭、第三層標頭、以及結尾部分的長度總和。如 圖5所示，若不切割待傳送封包，整個封包都要等到下一 次可上傳時間才能傳送，現在可上傳時間的剩餘時間就不 1320280 P52950066TW 22244twf.doc/〇〇6 ^傳送資料而完全浪f。若將 ，封包’則剩餘時間可傳送先傳封^，在^ = 後傳封包。傳送原封包與後傳封包所需的 、礤差01就是本實施例省下的頻寬。 ，6為本實施例節省上傳頻寬的另—張示意圖，其中 度不足乙太網路通訊協定的下限，因此以 式加長至最小可傳送長度。如圖6所示，傳The Lp-Lpi post-transmission packet 430 must conform to the minimum deliverable length specified in the communication protocol. For example, the Ethernet packet of the Ethernet network age cannot be less than 64 bytes. If the post packet 430 is less than 64 bytes, it must be padded. The padding method is extended to a minimum transferable length of 64 bytes. FIG. 5 is a schematic diagram of saving the upload bandwidth according to the embodiment, where the fixed length f refers to a part other than the payload in the packet, and in this embodiment, the second layer header, the third layer header, and the ending portion The sum of the lengths. As shown in Figure 5, if the packet to be transmitted is not cut, the entire packet will wait until the next upload time to be transmitted. The remaining time of the upload time is now 1320280 P52950066TW 22244twf.doc/〇〇6 ^Transfer the data and complete the wave f. If the packet is 'reserved', the remaining time can be transmitted first, and the packet is transmitted after ^ =. The difference 01 required to transmit the original packet and the post-transmission packet is the bandwidth saved in this embodiment. 6 is another schematic diagram of saving the upload bandwidth of the embodiment, wherein the degree is less than the lower limit of the Ethernet protocol, so the mode is extended to the minimum transmittable length. As shown in Figure 6,
ΐ頻Ξ 後傳封包所需的時間差6G1就是本實施例省下 本發月除了以上的網路封包傳送方法 圖7為依照本發明-實施例的網= j达裝置700的架構圖。網路封包傳送裝置包 切割控制器7〇2、以及封包傳送切割器703， 二 子拉組701又包括封包進入排程器704、封包暫存 器705、以及封包輸出排程器7〇6。 封包暫存器7〇5内建多個仔列(qu㈣，用來暫存 〇圖12繪不二個佇列以及十個封包(P1-P10)，這些數量 只是示範目的，並非限制本發明。由於網路封包傳^裝 在任何時刻都要接收來自區域網路(LAN)的封包卞是 在被動紅太域财(EPQN)的可上料間核傳送封 包’又有上傳時間限制’所以需要封包暫存器7〇5來暫 封包。 封包進入排程器704耦接於封包暫存器705，自區域 網路接收網路封包’依照進人排程演算法將網路封包存入 12 1320280 P52950066TW 22244twf.doc/〇〇6 封包暫存器705的仵列其中 輕接於封包暫存器705盘暫存模二;^排程器706則 可上傳時間依照輪出排程= 傳送的網路封包，並將其輸出至封包傳ί=Γ7Γί 述的進入排程演算法以及輸出 70^上 術，不在此贅述。 ㈣触^法都屬於傳統技 切割控制器702輕接於封包輸 制:號以控制封包傳送切割器7〇3。切割控 相當於圖3的步驟33〇、35〇以及3 —二 制信號的狀態控制封包傳送切割器7〇3是否 包。若可上傳時間的剩餘_足 =或刀。】封 封包，也就是封包輪出排程器观所輸送的網路 信號為傳送狀態。若剩餘時間不足而且_ 而且上述的預設條件成立 右繼化間不足 預贿件如圖3的步驟35〇盥二二，割狀態。這個 必須足夠傳送待傳送封包觀協 二封包必須為預設可切割的封包= 條件=和:面的方法實施例相同，不在此贅述。 割控==二^^ 控制信號為傳送狀態，封包傳唬。若 =至，乙太光織網路。二二會等直=網 封包傳运切割器703會等待一下次可上傳時間 13 1320280 P52950066TW 22244twf.d〇c/〇06 包。若控制佗號為切割狀態，封包傳送切割器7〇3會將網 路封包切割為先傳封包與後傳封包，使用剩餘時間傳送先 傳封包至被動式乙太光纖網路，後傳封包則等待下次可上 傳時間。 網路封包傳送裝置7〇〇其實就是執行前面實施例的網 路封包傳送方法，所以相_技術細節已見於前 例’不在此贅述》The time difference 6G1 required for the transmission of the packet is the network packet transmission method of the present embodiment except for the above. FIG. 7 is an architecture diagram of the network=j to the device 700 according to the embodiment of the present invention. The network packet transfer device packs the cutting controller 702, and the packet transfer cutter 703, which in turn includes a packet entry scheduler 704, a packet buffer 705, and a packet output scheduler 〇6. The packet register 7〇5 has a plurality of built-in queues (qu(4) for temporary storage of the two columns and ten packets (P1-P10), which are for exemplary purposes only and are not intended to limit the invention. Since the network packet transmission and reception is to receive the packet from the local area network (LAN) at any time, it is in the passive red Pacific domain (EPQN) of the core to transfer the packet 'has an upload time limit' so it is needed The packet buffer 7 〇 5 is used to temporarily encapsulate the packet. The packet entry scheduler 704 is coupled to the packet buffer 705, and receives the network packet from the regional network. The network packet is stored in the 12 1320280 according to the inbound scheduling algorithm. P52950066TW 22244twf.doc/〇〇6 The queue of the packet buffer 705 is lightly connected to the packet temporary register 705 disk temporary storage mode 2; ^ the scheduler 706 can upload time according to the round-trip scheduling = transmitted network The packet is output to the packet scheduling algorithm and the output 70^ is described here. (4) The touch method belongs to the traditional technology cutting controller 702, which is connected to the packet transmission: The cutter 7 〇 3 is conveyed by the control packet. The cutting control is equivalent to step 33 of FIG. 3, The status control packet of the 35〇 and 3-2 signals is transmitted by the cutter 7〇3. If the remaining time of the uploadable time is _foot=or knives.] The sealing package, that is, the net of the packet rotation scheduler The road signal is in the transmission state. If the remaining time is insufficient and _ and the above-mentioned preset conditions are established, the short-term pre-bridging component is as shown in step 35 of Figure 3, and the state is cut. This must be sufficient to transmit the packet to be transmitted. The second packet must be the same as the pre-cuttable packet = condition = and the method embodiment of the surface is the same, and will not be described here. Cutting control == two ^^ The control signal is the transmission state, and the packet is transmitted. If = to, Etolight Weaving the network. 22 will wait for the straight = net packet transport cutter 703 will wait for the next upload time 13 1320280 P52950066TW 22244twf.d〇c / 〇 06 package. If the control nickname is cutting state, packet transfer cutter 7 〇3 will cut the network packet into the first packet and the post packet, and use the remaining time to transmit the first packet to the passive Ethernet network, and the later packet will wait for the next upload time. Network packet transmission device 7〇 Oh, that’s actually The network packet transmission method of the previous embodiment is executed, so the phase details have been found in the previous example.
〇本發明並不侷限於被動式乙太光纖網路以及ISO 〇SI „何網路系統’只要有分配上傳時間的機制，而且 封I可切割，均可應用本發明。 ^上所述，本發明不需要任何雙方協調 『ndshaking) ’不需要自行排程，不改變原有的封包傳送 2 不而要複雜運算，僅使用簡單程序，即可達到言镅 I使用率。使用上述封包傳送方法與裝置用。 分配到的可上傳時間，和動態頻寬分配⑽A;:: = 因此即使出現較優先的封包需要立即傳送機上 餘置依然適用，不會造成頻寬浪費。 雖然本發明已以較佳實施例揭露如上然非 限定本發明，任何所屬枯、一、_用以 脫離本發明之精神:====，在不 =本發明之保護範圍當視後附之中請專利範圍i斤=者 【圖式簡單說明】 1320280 P52950066TW 22244twf.doc/006 圖1A與圖IB是習知被動式乙太光纖網路的封包傳送 示意圖。 圖2是依照本發明一實施例的封包切割條件示意圖。 圖3是依照本發明一實施例的網路封包傳送方法流程 圖。 狀 圖4A是依照本發明一實施例的封包切割方 素 圖。 '、~ 圖4B是依照本發明一實施例的封包傳送示肯圖。 圖5與圖6緣示本發明—實施例所節省的上^ 。 圖7是依照本發明-實施例的網路封包傳送裳置二構 【主要元件符號說明】〇 The present invention is not limited to the passive Ethernet network and the ISO 〇SI „he network system', as long as there is a mechanism for allocating the upload time, and the package I can be cut, the invention can be applied. No need to coordinate with any two parties "ndshaking" 'Do not need to schedule itself, do not change the original packet transmission 2 No complicated operation, only use a simple program, you can achieve the use rate of Word I. Use the above packet transmission method and device The available uploadable time, and the dynamic bandwidth allocation (10)A;:: = Therefore, even if a higher priority packet is needed, the remaining on the conveyor is still applicable, and the bandwidth is not wasted. Although the present invention has been better The embodiments are not limited to the above, and any of the following is used to deviate from the spirit of the present invention: ====, in the scope of protection of the present invention, the patent scope is i kg = [FIG. 1A and FIG. 1B are schematic diagrams of packet transmission of a conventional passive Ethernet optical network. FIG. 2 is a packet cutting according to an embodiment of the present invention. Figure 3 is a flow chart of a method for transmitting a network packet according to an embodiment of the present invention. Figure 4A is a block diagram of a packet cutting block according to an embodiment of the present invention. ', ~ Figure 4B is a packet according to an embodiment of the present invention. Figure 5 and Figure 6 show the above-mentioned savings of the present invention. Figure 7 is a diagram of a network packet transmission and placement in accordance with the embodiment of the present invention.
P1-P10、P5-1、P5-2 :網路封包 210 :待傳送封包 220 .可上傳時間的剩餘時間 230 :通訊協定規定的最短封包 240 :設定值 310-380 :流程圖步驟 410、420、430 :網路封包 網路封包組成部分 411-414^ 421-424 ^ 431-434： 501、601 ·傳送封包的時間差 700 :網路封包傳送裝置 701 :暫存模組 15 1320280 P52950066TW 22244twf.doc/006 702 :切割控制器 703 :封包傳送切割器 704 :封包進入排程器 705 :封包暫存器 706 :封包輸出排程器P1-P10, P5-1, P5-2: Network Packet 210: Packet to be transmitted 220. Remaining time of uploadable time 230: Shortest packet 240 specified by the protocol: Setpoint 310-380: Flowchart Steps 410, 420 430: network packet network packet component 411-414^421-424^431-434: 501, 601 · time difference of transmission packet 700: network packet transmission device 701: temporary storage module 15 1320280 P52950066TW 22244twf.doc /006 702: Cutting controller 703: packet transfer cutter 704: packet entry scheduler 705: packet register 706: packet output scheduler