TWI307231B - - Google Patents

Download PDF

Info

Publication number
TWI307231B
TWI307231B TW95117546A TW95117546A TWI307231B TW I307231 B TWI307231 B TW I307231B TW 95117546 A TW95117546 A TW 95117546A TW 95117546 A TW95117546 A TW 95117546A TW I307231 B TWI307231 B TW I307231B
Authority
TW
Taiwan
Prior art keywords
program
dynamic
working
work
processing
Prior art date
Application number
TW95117546A
Other languages
Chinese (zh)
Other versions
TW200744347A (en
Inventor
Jian Xin
Original Assignee
Mitac Res Shanghai Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Mitac Res Shanghai Ltd filed Critical Mitac Res Shanghai Ltd
Priority to TW095117546A priority Critical patent/TW200744347A/en
Publication of TW200744347A publication Critical patent/TW200744347A/en
Application granted granted Critical
Publication of TWI307231B publication Critical patent/TWI307231B/zh

Links

Description

1307231 九、發明說明: 【發明所屬之技術領域】 本發明係關於一種伺服器系統及其工作方法,尤其係關於一 、 種可以快速處理TCP請求的伺服器系統及方法。 1 【先前技術】 目鈾,網絡間交互日夜頻繁,譬如網絡遊戲或網絡聊天,這 些多涉及到信息的迅速並且可靠的傳遞,這些均要求網絡祠服器 ® 系統多程序並發工作且要有很好的回應速度。目前,通常採用多 程序以實現並發工作。然而習知技術的這種做法存有諸多缺陷, 因爲目前的操作系統基本都使用CPU分時的方法來達到程序並 發工作之目的,在工作時,存在程序間上下文切換的開銷,一味 使用多程序達到並發的目的反而會降低系統工作效率,伯用系統 資源,並使回應速度變慢。 【發明内容】 鲁為了解決上述習知技術中的問題與缺陷,本發明的目的之一 在於提供一處理TCP請求之系統,係用以快速處理TCP請求,該 系統包括以下模組:-程序啓動模組,_以在祕啓動時預先 啓動-管理程序和複數個工作程序,並把該複數個工作程序的屬 性標記為靜態工作程序;一程序分配及生成模組,係用以在用戶 端連接到系統後’由所述管理程序根據每個工作程序的工作情況 安排確定用於處理該連接之工作程序或生成新的動態工作程序; 一 S〇Cket異步輸入/輸出控繼、组,係用以透過socket異步輸入 1307231 /輸出魏令触功程相岐㈣料理⑽有Tcp連接及 每一連接的X作處雜ι·以及-_工作程序釋放模組,係用 以根據系統當前之網路連接數量及工作狀況,關所生成之動態 工作程序。 本發明的又-目的在於提供—處理Tcp請求之方法,係用以 快速處理TCP請求’該方法包括以下步驟:—程序啓動步驟,係 用以在系統啓動時預先啓動-管理程序和複數個工作程序,並把 該複數個工作程序的屬性標記為靜態工作程序;—程序分配及生 成步驟,_以在用戶端連接_統後,由所述管理程序根據每 個工作程相功情況安排確定祕處_連接之卫作程序或生 成新的動態工作程序;— SGeket異步輸人/輸出控制步驟,係用 以透過S〇eket異錄人/輸㈣能令所虹作餅_監控由盆 管理的所有TCP連接及每—連接的工作處理狀況;以及—祕二 作私序釋放步驟,伽啸據系統當前之_連紐量及工作 況,關閉所生成之動態工作程序。 、士發縣雌態工作程序和靜態王作料相結合的 2地_系統中程序_數’從喊省了系 占用率並提高了系統的工作效率。此外,本發明中料貝原 s λ /χ. , 了 socket 々月’』輸出功能’使每個工作程序可以負責管理 ;接二單料,式,—的異步 高L=面上發處理’從而減小並發的工作程序數量,: 1307231 綜上所述,與習知技術相比,本發明的優點是:由於本發明 之處理TCP請求之伺服器系統及方法採用動態工作程序和靜態工 作耘序相結合的方式實現程序並發工作,因而提高了系統工作效 率,並降低了系統的資源佔用率,從而使系統工作在較佳的工作 狀態。此外,透過socket異步輸入/輸出功能,系統以單程序工 作模式同時監視多個SGeket的異步處理,從而減小了並發的工作 程序數量,提高了系統效能。 【實施方式】 將、、-σ δ附圖對本發明的優選實施方式作詳細說明 一請參考「第1圖」及「第2圖」,「第1圖」爲本發明之飼服 …系m统架構賴,「第2圖」爲本發明之處理請求之 方法的部分步騎朗。如「第⑼」所示,本發明之値器系 統啓動後,透過-程序啓動模組,啓動—個管理程序ig和p個靜 恶工作程序20。管理程序10監聽向系統發出的TCP連接請求, 分配及生成模組中的平均分配模組根據系統當前連 將TCP連接任務按照平均分配的原則分配給相應 均達到“預:處理’當每個靜態工作程序20的TCP連接個數 祕U統預狄之門檻值M時,將透過程序分配及生成做中的 一門播值報警模組按照程序分配t 、 動態工作程序30,則,告 ▲報3原則,生成一個 分配的频㈣• 的取連接接人時,按照平均 :的原則’新的連接會盡量交給這個動態工作程序3G來處理, 一悲工條序3G具有最讀數之_,當射辦純警原則生 1307231 鱗3G的個數達到系統設置的上限值q時,系統 新的_#=^賴財的—祕關難令錢不再生成 、、作程序,而^系、统出現新的TCP連接請求時,管理程 二〇仍按平均分配的原則提交給合適的工作程序(包括靜態工^ 王和動態工作程序)來處理,但每個工作程序處理的tc ::還是有限制的’限制數為N,當所有— 數都達到N時,系統將拒絕新的連接接人,此 接到另外上去辦。 ^連 另外,本發明之處理TCP請求之系統還包括:- Socket異步 輸入/輸轉韻組’侧以透過SQeket異步輸V輸出功能令 工作私序啊監控由其管理的所有連接及每-賴的工作處理狀 j ’系統錄設置触’伽以在械纽環境下(cpu頻率 咼低、物理内存大小、硬盤速度快慢、是否是多CPU、何種操作 系統等),㈣性㈣試確定味理想的祕倾m統參數 值包括.靜‘紅作轉之個數P、每個卫作程序處理的連接個數 之門松值]VI、生成的動態工作程序之個數的上限值Q以及每個工 作程序所能纽的連接健之極限值N ;以及—動態工作程轉 放板組’係用以根據系統當前之網路連接數量及工作狀況,關閉 所生成之動||二作料,鞠態工伽轉賴崎在下述三種 情況之一的情況下關閉所生成之動態工作程序: 第種情況,系統停止時若存在動態工作程序,則關閉該動 態工作程序;BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a server system and a method for operating the same, and more particularly to a server system and method for rapidly processing TCP requests. 1 [Prior Art] Uranium, network interactions day and night, such as online games or web chat, these involve rapid and reliable transmission of information, which requires the network server system multi-program concurrent work and there must be very Good response speed. Currently, multiple programs are often used to achieve concurrent work. However, this practice of the prior art has many drawbacks, because the current operating system basically uses the CPU time-sharing method to achieve the purpose of the program concurrent work, at work, there is overhead of inter-program context switching, blindly using multiple programs. Concurrency will reduce the efficiency of the system, use system resources, and slow down the response. SUMMARY OF THE INVENTION In order to solve the problems and defects in the above-mentioned prior art, one of the objects of the present invention is to provide a system for processing a TCP request for quickly processing a TCP request, the system comprising the following modules: - Program Startup The module, _ pre-launched at the start of the secret-management program and a plurality of working programs, and marks the attributes of the plurality of working programs as static working programs; a program allocation and generating module is used to connect at the user end After the system, the management program determines the working program for processing the connection or generates a new dynamic working program according to the working condition of each working program; a S〇Cket asynchronous input/output control, group, system Through the socket asynchronous input 1307231 / output Wei Ling touch the power of the phase (four) cuisine (10) have Tcp connection and each connection X work miscellaneous ι· and _ work program release module, according to the current number of network connections according to the system And the working conditions, the dynamic work procedures generated by the off. Yet another object of the present invention is to provide a method for processing a Tcp request for quickly processing a TCP request. The method includes the following steps: - a program startup step for pre-launching at the system startup - a hypervisor and a plurality of jobs a program, and mark the attributes of the plurality of working programs as a static working program; - the program allocation and generating steps, _ after the user terminal is connected, the management program arranges the secret according to each working process. Connected to the program or generate a new dynamic work program; - SGeket asynchronous input / output control steps, through the S〇eket different recording / loss (four) can make the rainbow cake _ monitor all TCP managed by the basin The connection and the working condition of each connection; and the secret second release procedure, the gaer is based on the current system and the working condition, and the generated dynamic working program is closed. The combination of the female work program in Shifa County and the static king work material _ the number of programs in the system _ number from the call to save the system occupancy rate and improve the efficiency of the system. In addition, in the present invention, the material of the original s λ / χ., the socket 々 month '』 output function 'so that each work program can be responsible for management; two materials, the type, - asynchronous high L = face processing" Reducing the number of concurrent work programs, 1307231 In summary, the advantages of the present invention over the prior art are that the server system and method for processing TCP requests of the present invention employ dynamic working procedures and static working sequences. The combined method realizes the concurrent operation of the program, thereby improving the system work efficiency and reducing the resource occupancy rate of the system, thereby making the system work in a better working state. In addition, through the asynchronous input/output function of the socket, the system simultaneously monitors the asynchronous processing of multiple SGekets in a single-program mode, thereby reducing the number of concurrent programs and improving system performance. [Embodiment] The preferred embodiment of the present invention will be described in detail with reference to the drawings and drawings. Please refer to "Fig. 1" and "Fig. 2", and "Fig. 1" is a feeding service of the present invention. The "Architecture 2" is part of the method of processing the request of the present invention. As shown in "(9)", after the device system of the present invention is started, a management program ig and p static work programs 20 are started through the program startup module. The hypervisor 10 listens to the TCP connection request sent to the system, and the average allocation module in the allocation and generation module allocates the TCP connection task according to the current system according to the principle of the average allocation to the corresponding "pre-processing" when each static When the TCP connection number of the working program 20 is the secret value of the pre-Ddi's threshold value M, the program will be assigned and generated by the program distribution alarm module according to the program, and the dynamic work program 30, then, the report 3 Principle, generate an assigned frequency (four) • When taking the connection, according to the principle of averaging: 'The new connection will be handed over to this dynamic working program 3G to handle, a sorrowful sequence 3G has the most reading _, when Shooting pure police principle students 1307231 scale 3G number reaches the upper limit q of the system setting, the system new _#=^赖财--secret is difficult to make money no longer generate, program, and ^ system, When a new TCP connection request occurs, the management process is still submitted to the appropriate working program (including the static worker and the dynamic worker) according to the principle of even distribution, but the tc:: limited 'The limit number is N. When all the numbers reach N, the system will reject the new connection, which is connected to the other. ^ In addition, the system for processing TCP requests of the present invention further includes: - Socket asynchronous input /Transfer to rhyme group side to pass the SQeket asynchronous V output function to make the work private order monitor all the connections managed by it and the work processing of each of them - the system is set to touch the gamma in the mechanical environment ( Cpu frequency is low, physical memory size, hard disk speed, whether it is multi-CPU, what kind of operating system, etc.), (four) sex (four) try to determine the taste of the ideal secret d-parameter parameter value. Static 'red-turned number P , the number of connections processed by each processing program] VI, the upper limit value Q of the generated dynamic working program, and the limit value N of the connection health of each working program; and - dynamic work The process of transferring the board group is used to close the generated motion||two materials according to the current number of network connections and working conditions of the system, and the state is turned off by Laiqi in one of the following three situations: Dynamic working procedure: first In case, if there is a dynamic working program when the system is stopped, the dynamic working program is closed;

1307231 值匕括·靜態工作程序之個數P、每個工作程序處理的 =伽、生成的動態工作程序之個數的上限值Q以及每個2 乍私序所祕理的連接健之極限值N;以及—動態工作程 =驟’係用以峨統當前之網路連接數量及工作狀況,關閉 成之動紅作程序,該_工作程序釋放步驟將在下述三種 情況之一的叙下_所生成之作程序·· =種情況’系統停止時若存在_工作㈣,觸閉該動 悲工作程序, 第-種情況’動態工作程序在運行過程中,其處理之TC 接個數為G時,賴職動態工作程序;以及 第二種情況,當系統根據當前工作情況決定關某些動態工 作私序時,啦統把將要顧之祕丄擁序處理之連接交給其 它工作程序處理後,關閉該動態工作程序。其中在第二種情況下、’ 關閉動作4序之步驟進—步包括以下步驟: .一睛求步驟,係用以將關閉動態工作程序之請求發送給系 、、先’以及斷步驟’係由系統決定是否可以關閉該絲工作程 序(以防止程序生成抖動),若系統允許關閉,則關閉該動態工作 程序。 +雖然本發a㈣前迷之優選實施方式揭露如上,然其並非用以 F疋本發明。本領域的技術人貞應該意識到在*脫離本發明所附 所揭示的本發明之範圍和精神的情況下,所爲之 1307231 ^動與/閏飾’均屬本發明之專利保護細之内。關於本發明所界 疋之保護範圍請參考所附之申請專利範圍。 【圖式簡單說明】 「第1圖」爲本發明之飼服器系統之系統架構簡圖;以及 「第2圖」爲本發明之處理TCP請求之方法的部分步驟流程 圖。 【主要元件符號說明】 !〇 管理程序 20 靜態工作程序 30 動態工作程序 M每個ι作轉處理的連接她之門檻值 N每個工倾序所能處理的雜健之極限值 P 靜態工作程序之個數 Q 生成的動態工作程序之個數的上限值 步驟201系、統啓動時’啟動-管理程序和Μ固靜態工作程 序 步驟202平均分配TCP連接至各工作程序 乂驟203判斷系統當前各程序之連接個數是否達到μ 步驟204生成動態工作程序 步驟205判斷動態工作程序之個數是否達到q 、 步驟206 $再生成動&amp;工作程序,且平均分配新連接至各 工作程序 12 1307231 步驟207判斷各工作程序之連接個數是否均達到N 步驟208拒絕新的連接接入1307231 The value includes the number of static working programs P, the = gamma processed by each work program, the upper limit value of the number of generated dynamic working programs, and the limit of the connection health of each 2 乍 private order. The value N; and - the dynamic working process = step ' is used to shut down the current number of network connections and the working status, and the program is released in the following three cases. _ generated program ·· = kind of situation 'If there is _ work (4) when the system stops, the erroneous work program is closed, the first case 'dynamic work program during the running process, the number of TCs processed is G, the dynamic work process of the Lai; and the second case, when the system decides to close certain dynamic work private orders according to the current work situation, the system will hand over the connection that will take care of the secret processing to other work programs. After that, close the dynamic worker. In the second case, the step of 'closing the action 4 step' includes the following steps: The first step is to send the request to close the dynamic working program to the system, the first 'and the first step' It is up to the system to decide whether the wire work program can be closed (to prevent the program from generating jitter), and if the system allows shutdown, the dynamic work program is closed. Although the preferred embodiment of the prior art (a) is disclosed above, it is not intended to be used in the present invention. It is to be understood by those skilled in the art that within the scope and spirit of the invention disclosed in the appended claims, the <RTI ID=0.0># </ RTI> </ RTI> </ RTI> . Please refer to the attached patent application for the scope of protection of the invention. BRIEF DESCRIPTION OF THE DRAWINGS "FIG. 1" is a schematic diagram of a system architecture of a feeding device system of the present invention; and "FIG. 2" is a partial flow chart of a method for processing a TCP request of the present invention. [Main component symbol description] !〇Management program 20 Static work program 30 Dynamic work program M Each ι is transferred to the connection of her threshold NN The limit value of the fitness that can be handled by each work order P Static work program The upper limit value of the number of dynamic working programs generated by the number Q is step 201, and the startup-management program and the tampering static working program step 202 averagely allocate TCP connections to the respective working programs 203 to determine the current system. Whether the number of connections of each program reaches μ or step 204 generates a dynamic working program. Step 205 determines whether the number of dynamic working programs reaches q, step 206 regenerates the dynamic &amp; working program, and evenly allocates a new connection to each working program 12 1307231 Step 207: determining whether the number of connections of each working program reaches N. Step 208 rejects new connection access.

1313

Claims (1)

,1307231 如r’月/q修正替換頁丨 I*、申請專利範圍: ~ … 求,謗系 、一種處理Tcp請求m 叫速處理TCP誇 統包括以下模組: v t序啓動她,係用以在系聽動時預先㈣一管 獅細㈣糊紐標^ 一程序分配及生成模組’係肋在用戶端連接到系 由所述管理程序根據所述每個工作程序缸作情況安二 用於處理該連接之工作程序或生成新的動態工作程序;又 ocket異步輪j入/輸出控制模組,係用以透過⑽咏t 兴步輪入/輸出功能令所述工作程序同時監控由其管理的戶 有連接及每一連接的工作處理狀況;以及 所 -動恶工作程序釋放模組’制以根據紐當前之網路連 接數量及工作狀況,關閉所生成之動態工作程序。 2、如申請專利範圍第1項所述之處理TCP請求之系統,其中所 述%序分配及生成模組進一步包括以下模組: 一平均分配模組,系用以檢查所述每個工作程序當前處理 之連接數’並按照平均分配之原則,將新連接分配給處理連接 個數較少之工作程序進行處理; —門樓值報警模組,係用以當所述管理程序檢測到當前每 们工作私序處理之連接個數均超過系統預設之門植值後,該管 理程序自動生成一個新的動態工作程序;以及 14 1307231 &amp; 一系統限鑛組’係用以當所述門摄值報賴組生成的動 二、作長序之個數達到系統預設之上限值時,系統不再生成新 /動t工作程序,在系統產生新的連接時,所述管理程序仍按 ‘、、、斤述平均分配原則把新連接提交給合適的工作程序處理,當 斤有的工作程序處理的連接個數均達到所能處理的極限數 時,系統將拒絕新的連接接入。, 1307231 such as r 'month / q correction replacement page 丨 I *, the scope of patent application: ~ ... seeking, 谤, a processing Tcp request m called speed processing TCP boast includes the following modules: vt order to start her, used In the case of listening, (4) a lion (4) paste button ^ a program distribution and generation module ' ribs are connected at the user end by the management program according to each of the work program cylinders Processing the connection work program or generating a new dynamic work program; and the ocket asynchronous wheel j input/output control module is configured to enable the work program to be simultaneously monitored by the (10) 咏t step-in/output function. The managed households have connections and the working processing status of each connection; and the "moving work program release module" is configured to close the generated dynamic working program according to the current number of network connections and working conditions of the New Zealand. 2. The system for processing a TCP request according to claim 1, wherein the %-sequence allocation and generation module further comprises the following modules: an average distribution module for checking each of the working procedures The number of connections currently processed' and according to the principle of average allocation, the new connection is assigned to the work program with a small number of processing connections; the gate value alarm module is used to detect the current each when the management program detects After the number of connections in the private processing of the work exceeds the preset value of the system, the management program automatically generates a new dynamic working program; and 14 1307231 &amp; a system limited mine group is used to take the door When the number of the second and long-order generated by the value report group reaches the upper limit of the system preset, the system no longer generates a new/moving t working program. When the system generates a new connection, the management program still presses The principle of ',, and jin's average allocation is to submit the new connection to the appropriate working procedure. When the number of connections handled by the working procedure reaches the limit that can be processed, the system will reject Access connection. 如申睛專利細第1項所述之處理TCP請求之系統,其中所 、、動L工作k序釋放模組將在下述三種情況之_的情況下關 閉所生成之動態工作程序: 第一種情況,系統停止時若存在動態工作程序,則關閉該 動態工作程序; 第二種情況,動態工作程序在運行過程中,其處理之連接 個數為0時,則關閉該動態工作程序;以及 第二種情況’當系統根據當前卫作情況決定咖某些動態 工作程序時,則祕把將要_〗之_二雜序處理之連接= 給其它工作程序處理後,關閉該動態工作程序。 4、如申請專概圍第3項所述之處理取請求之錢,其中在 所述第二種情況下,所述動態工作程序釋放模組進—步包括以 下模組: .一請求模組,係用以將關閉所述動態工作程序之請求發送 、'、α糸統,及 15 1307231 正魏. 辦/月广修 —決斷模組,係由系統決定是否可以關閉該動態工作程 序’若系統允許關閉’則關閉該動態工作程序。 如申請專利範圍第1項所述之處理TCP請求之系統,其中包 括: 系統參數δ又I模組,係用以在相應系統環境下經過性能 測§式確定系統參數值。 6、 如申請專利範圍第5項所述之處理TCP請求之系,统,其中所 4系統參數值包括:所述靜態工作程序之個數、每個工作程序 處理的連接健之門播值、生成的動態工作程序之個數的上限 值以及每個X作料所能處理的連接贿之極限值。 7、 -種處理TCP請求之方法,係用以快速處理TCp請求,該方 法包括以下步驟: 各序啓動步驟,係肋在纽啓麟預先频—管理程 序和複數個工作程序,並把該複數個工作程序的屬性標記 態工作程序; 野 -程序分配及生成步驟,系㈣在用戶端連接到系 由所述管理程序根_述每個工作㈣的讀軌 用於處雜連接之工作程序或生成新的動態工作程序,·天 - socket異步輪入/輸出控制步驟,係用以透過 異步輸入/輪出功能令所述工作程序同時監控由其0:饮 有連接及每-連接的倾理狀况;以及 、所 16 ^307231 一動態工作程序釋放步驟,係用以根^ 當前之網路連 其中所 接數量及工作狀況,_所生成之動態工作-S、如申請專利範圍第7項所述之處理取。 述輕序分配及生成步驟進一步包括以下步=之方法, 一平均分配步驟,系用以檢查所述每個 辁序自動生成一個新的動態工作程序;以及 At系’苑限制步驟’係用以當所述門才監值報警步驟生成的動 悲工作程序之她翻祕預設之上限辦,系統*再生成新 勺動恶工作程序,在系統產生新的連接時,所述管理程序仍按 照所述平均分配原則把新連接提交給合適的工作程序處理,當 所有的工作程序處理的連接個數均達到所能處理的極限個數 夺系統將拒絕新的連接接入。 如申清專利範圍第7項所述之處理TCP請求之方法,其中所 述動態工作程序釋放步驟將在下述三種情況之一的情況下關 閉所生成之動態工作程序: 第一種情況,系統停止時若存在動態工作程序,則關閉該 動態工作程序; 17 1307231 個數::一其處理之連接 :以及 10、For example, in the system for processing a TCP request as described in the first item of the patent application, the dynamic operation program is closed in the case of the following three cases: In case, if there is a dynamic working program when the system stops, the dynamic working program is closed; in the second case, when the dynamic working program runs, the number of connected connections is 0, then the dynamic working program is closed; In two cases, when the system decides to use some dynamic working procedures according to the current security situation, the secret will be _ _ _ two miscellaneous processing connection = after processing to other working programs, close the dynamic working program. 4. In the case of the second case, the dynamic work program release module further includes the following modules: , is used to send the request to close the dynamic working program, ', alpha system, and 15 1307231 Zheng Wei. Office / month Guangxiu - decision module, the system determines whether the dynamic work program can be closed' The system allows to close 'and then close the dynamic worker. The system for processing a TCP request as described in claim 1 includes: a system parameter δ and an I module for determining a system parameter value by a performance test in a corresponding system environment. 6. The system for processing a TCP request as described in claim 5, wherein the system parameter values include: the number of the static working programs, the connection gate value processed by each working program, The upper limit of the number of dynamic work programs generated and the limit value of the connection bribe that each X material can handle. 7. A method for processing a TCP request, which is used to quickly process a TCp request, the method comprising the following steps: each sequence starting step, the rib in the New Qilin pre-frequency management program and a plurality of working programs, and the plural The attribute tag state work program of the work program; the wild-program assignment and generation step, the system (4) is connected to the user program at the user end, and the read track of each work (4) is used for the work process of the miscellaneous connection or Generating a new dynamic working program, the sky-socket asynchronous wheeling/output control step is used to enable the working program to simultaneously monitor the 0: drink connection and per-connection through the asynchronous input/rounding function. Status; and, 16 ^ 307231 a dynamic work program release step, is used to root the current network connection number and working status, _ generated dynamic work -S, such as patent application scope 7 The processing is taken. The step of assigning and generating the light sequence further includes the following steps: an average allocation step for checking that each of the sequences automatically generates a new dynamic working program; and the At system 'court restriction step' is used to When the threshold of the sorrowful work procedure generated by the door-to-supervised alarm step is reversed, the system* generates a new scooping work program, and when the system generates a new connection, the management program still follows The average allocation principle submits the new connection to the appropriate work process. When all the work processes process the number of connections to reach the limit that can be processed, the system will reject the new connection access. The method for processing a TCP request according to claim 7, wherein the dynamic worker release step closes the generated dynamic work program in one of the following three cases: In the first case, the system stops. If there is a dynamic working program, close the dynamic working program; 17 1307231 number:: the connection of its processing: and 10, 第-種情況’衫統根據當前功情況決定關某些動態 —敕序時,則系統把將要關閉之動態工作程序處理之連接交 4它工作辦處理後,關__二作程序。 、、申二專利mi第9項所述之處理TCp請求之方法,其中在所 述第二種情況下’關所述_工作程序之步驟進-步包括以 下步驟: * ^ -睛求步驟’係㈣將_所述雜工作程序之請求發 給系統;以及 决斷步驟’係由系統決定是否可以關閉該動態工作程 序’若系統允許關閉’則關閉該動態工作程序。 1如申明專利範圍第7項所述之處理Tcp請求之方法,其中包括: 、一系統參數設置步驟’係用以在相應系統環境下經過性能 測試確定系統參數值。 12、如申凊專利範圍第u項所述之處理Tcp請求之方法,其中所 述系統參數值包括:所述靜態工作程序之個數、每個工作程序 處理的連接個數之門檀值、生成的動態工作程序之個數的上限 值以及每個工作程序所能處理的連接個數之極限值。 18In the first case, the system decides to turn off certain dynamics according to the current power situation. When the system is in the process of ordering, the system will handle the connection of the dynamic working program to be closed. The method for processing a TCp request according to the second item of claim 2, wherein in the second case, the step of "closing the _work procedure" includes the following steps: * ^ - the step of seeking The system (4) sends a request for the heterogeneous work program to the system; and the decision step is determined by the system whether the dynamic work program can be closed, and if the system allows the shutdown, the dynamic work program is closed. 1 The method for processing a Tcp request as recited in claim 7, wherein: a system parameter setting step is used to determine a system parameter value through a performance test in a corresponding system environment. 12. The method of processing a Tcp request as described in claim 5, wherein the system parameter value comprises: a number of the static working program, a number of connections processed by each working program, and a threshold value. The upper limit of the number of generated dynamic work programs and the limit of the number of connections that can be processed by each work program. 18
TW095117546A 2006-05-17 2006-05-17 System and method for processing TCP request TW200744347A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW095117546A TW200744347A (en) 2006-05-17 2006-05-17 System and method for processing TCP request

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW095117546A TW200744347A (en) 2006-05-17 2006-05-17 System and method for processing TCP request

Publications (2)

Publication Number Publication Date
TW200744347A TW200744347A (en) 2007-12-01
TWI307231B true TWI307231B (en) 2009-03-01

Family

ID=45071560

Family Applications (1)

Application Number Title Priority Date Filing Date
TW095117546A TW200744347A (en) 2006-05-17 2006-05-17 System and method for processing TCP request

Country Status (1)

Country Link
TW (1) TW200744347A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI555366B (en) * 2013-11-04 2016-10-21 騰訊科技(深圳)有限公司 A protocol request processing method

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI555366B (en) * 2013-11-04 2016-10-21 騰訊科技(深圳)有限公司 A protocol request processing method
US10009444B2 (en) 2013-11-04 2018-06-26 Tencent Technology (Shenzhen) Company Limited Protocol request processing method and device using the same

Also Published As

Publication number Publication date
TW200744347A (en) 2007-12-01

Similar Documents

Publication Publication Date Title
TWI447650B (en) Interrupt distribution scheme
US8462369B2 (en) Hybrid image processing system for a single field of view having a plurality of inspection threads
CN102156532B (en) Computer and method that reduces power consumption while maintaining a specific function
US9262272B2 (en) Data center power adjustment
WO2019195969A1 (en) Data synchronization processing method and apparatus
TWI480728B (en) Squelch filtration to limit false wakeups
US20130047157A1 (en) Information processing apparatus and interrupt control method
CN102597912B (en) Coordinating device and application break events for platform power saving
CN111858228B (en) Method and system for accelerated kernel status monitoring in a storage device
EP2618257A2 (en) Scalable sockets
US20210326177A1 (en) Queue scaling based, at least, in part, on processing load
CN105138287A (en) Storage equipment, interruption control method and power supply time measuring method
WO2023098192A1 (en) Method and apparatus for processing abnormal power failure of solid state disk, and electronic device and medium
CN109117280A (en) The method that is communicated between electronic device and its limiting process, storage medium
JP4620768B2 (en) Control of a computer system having a processor including multiple cores
TWI307231B (en)
US11409670B2 (en) Managing lock coordinator rebalance in distributed file systems
Mhedheb et al. Energy-efficient task scheduling in data centers
JP4761229B2 (en) Operation management apparatus, operation management method and program
WO2021139113A1 (en) Method and apparatus for signing transaction data, computer device, and storage medium
CN103995789A (en) Direct memory access achievement system and method
KR20160061726A (en) Method for handling interrupts
CN113076189B (en) Data processing system with multiple data paths and virtual electronic device constructed using multiple data paths
Dong et al. Self-acting load balancing with parallel sub file migration for parallel file system
Tziritas et al. Server consolidation in cloud computing

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees