WO2005117461A1 - A intelligent data scheduling method of short message system - Google Patents

A intelligent data scheduling method of short message system Download PDF

Info

Publication number
WO2005117461A1
WO2005117461A1 PCT/CN2004/001303 CN2004001303W WO2005117461A1 WO 2005117461 A1 WO2005117461 A1 WO 2005117461A1 CN 2004001303 W CN2004001303 W CN 2004001303W WO 2005117461 A1 WO2005117461 A1 WO 2005117461A1
Authority
WO
WIPO (PCT)
Prior art keywords
short message
data
mode
message system
load
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2004/001303
Other languages
English (en)
French (fr)
Inventor
Kai Fu
Zhigui Yuan
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
ZTE Corp
Original Assignee
ZTE Corp
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 ZTE Corp filed Critical ZTE Corp
Publication of WO2005117461A1 publication Critical patent/WO2005117461A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/12Messaging; Mailboxes; Announcements
    • H04W4/14Short messaging services, e.g. short message services [SMS] or unstructured supplementary service data [USSD]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W28/00Network traffic management; Network resource management
    • H04W28/02Traffic management, e.g. flow control or congestion control
    • H04W28/10Flow control between communication endpoints

Definitions

  • the present invention relates to various types of short message systems in the field of communications, and in particular, to a method for scheduling data in a short message system. Background technique
  • Short message systems or similar products have been set up in existing CDMA networks, GSM networks, and fixed telephone networks. While bringing convenient services to the majority of users and bringing huge profits to operators, the requirements of short message systems are becoming more and more stringent. Now, the security and robustness of short message systems have been aligned with the main equipment such as switches. . It is a basic requirement to ensure the normal operation of the short message system in any case.
  • the basic architecture and data processing mode of the short message system determine that it has a lot of limitations in terms of resistance to traffic shocks and protection against database server failures, and the processing methods are relatively simple.
  • the commonly used overload method for anti-traffic shock will reject or delay the sending of a large number of normal messages, resulting in a low sending rate and response rate of the short message system, and under the long-term large traffic shock, if the accumulated Sending a message to a certain level may cause overflow and message loss until the entire system is down due to excessive load.
  • Chinese Patent Application No. 021239039.7 with a publication date of January 14, 2004 discloses a method for coping with a large traffic impact short message system.
  • the processing strategy of the method when the load increases is mainly to temporarily delay the transmission. However, this will cause a low response rate of the system; the method also discards short messages that failed to be sent by modifying the temporary errors of the short messages to permanent errors.
  • the higher the traffic pressure the more transmission failures will result. If the failure message is directly discarded without retransmission, the transmission success rate of the system will be greatly reduced.
  • the database server is the core component of the short message system. Once a failure occurs, it will have a fatal effect on the entire short message system. Therefore, its security needs to be enhanced.
  • system redundancy and disaster recovery technology is generally used to effectively control the failure rate of the entire system, but its implementation is relatively complicated and costly. Therefore, the current short message system urgently needs an efficient and easy method for intelligently scheduling data to enhance the robustness and security of the system, while ensuring the response rate and success rate of message sending. Summary of the invention
  • the present invention is proposed in view of the foregoing problems in the prior art, and its purpose is to provide a method for intelligently scheduling data in a short message system, which can enable the short message system to adopt a corresponding operating mode according to a change in the traffic volume to the data. Intelligent dispatch is performed to improve the ability of the short message system to resist the impact of traffic, and to ensure the sending rate and response rate of the short message system.
  • the invention provides a method for intelligently scheduling data in a short message system, selecting an operating mode of the system according to a load index of the current short message system, and scheduling the short message service data according to the selected operating mode.
  • the method for intelligently scheduling data in the short message system specifically includes the following steps:
  • the operation mode of the short message system is divided into a normal mode and a stress-resistant mode when the traffic pressure is too high, and the switching mode and switching conditions of the configuration modes are configured, as well as the data scheduling strategy corresponding to each mode.
  • the data generated by the business process is selectively discarded; the load indicator of the short message system is detected;
  • the data is scheduled according to the data scheduling strategy corresponding to the current operating mode.
  • the method for intelligently scheduling data further includes: dividing a short message service processing process into two independent processing processes of a short message initiation process and a short message termination process, and setting a termination suspension load threshold; among them,
  • the method further includes: in a normal mode, if the load indicator exceeds the termination call suspension load threshold, suspending the termination of the short message; If the load index is lower than the termination call suspension load threshold, termination processing is performed.
  • the method of the present invention provides multiple operation modes of the short message system, each mode corresponds to a different data scheduling strategy, and can be flexibly switched with each other, so that the short message system can use the corresponding operation according to the change of the traffic volume. Mode, thereby improving the anti-traffic capability of the system, enhancing the robustness, stability, and flexibility of the system. Since the existing sending strategy is not changed, the response rate of the system is also guaranteed.
  • the present invention can also divide the short message service processing process into two separate processes, initiation processing and end-call processing, and provides a system decompression method of suspending the end-call processing, which can exert the system's potential and reduce waves. Impact of Surge Traffic Impact on Short Message System. Further, the short message data that affects the success rate of sending is stored preferentially, so that the successful sending of the short message is preferentially guaranteed when the traffic pressure is high. In addition, because the present invention adopts a disaster tolerance mode, it can prevent the catastrophic impact of the short message system caused by the failure of the database server, and achieve disaster tolerance in a simple and easy way, saving system costs. BRIEF DESCRIPTION OF THE DRAWINGS
  • FIG. 1 is a schematic diagram of functional modules of a short message system using a method for intelligently scheduling data according to an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a short message service processing process
  • FIG. 3 is a schematic diagram showing a relationship between various operating modes of a short message system in a method for intelligently scheduling data according to an embodiment of the present invention
  • FIG. 4 is a service processing flowchart of a method for intelligently scheduling data according to an embodiment of the present invention in a stress-resistant mode
  • FIG. 5 is a flowchart of a method for intelligently scheduling data according to an embodiment of the present invention for selecting a data scheduling strategy in a compression-resistant mode
  • FIG. 6 is a service processing flowchart of a method for intelligently scheduling data in a disaster tolerance mode according to an embodiment of the present invention. detailed description
  • the basic idea of the present invention is to set multiple operating modes for the short message system, select the operating mode of the system according to the load index of the current short message system, and schedule the short message service data according to the selected operating mode.
  • the operation mode of the short message system is first divided into a normal mode and a pressure-resistant mode, wherein the normal mode is a basic operation mode of the prior art, and the pressure-resistant mode is defined as the short message system under traffic pressure. Big time operation mode. Then configure the switching mode and switching conditions between each operation mode, and the data scheduling strategy corresponding to each operation mode. In the anti-stress mode, the data generated by the short message service processing process is selectively discarded. Detection system Compare the load index with the above switching conditions; if the load index meets the switching conditions, then complete the switching between the operating modes according to the above switching method. For the data generated during the short message service processing, follow the current operation of the short message system The data scheduling strategy corresponding to the mode schedules it.
  • FIG. 1 is a functional module schematic diagram of a short message system adopting a method for intelligently scheduling data according to an embodiment of the present invention.
  • the short message system mainly includes a service processor 101, an operation and maintenance center 102 connected to the service processor 101, and Active database 103.
  • the service processor 101 is a core component for short message service processing, and various operation modes are implemented on the service processor 101.
  • the operation and maintenance center 102 is a man-machine interface for providing a man-machine interface for configuring an operation mode.
  • the main database 103 is a main data storage device for storing main data in the short message system, such as user information required for business processing, short message data generated by the business processing, and the like.
  • Normal mode It is the operating mode when the short message system is normal.
  • the business processor 101 is connected to the main database 103, and its business and data processing follow the existing standard processing flow; the system normally reads the user information required for business processing and saves the data generated by the business processing (such as Short message, etc.); In addition, the short message system completes the preservation of the data.
  • Anti-stress mode It is the operation mode when the traffic pressure of the short message system is too high.
  • the service processor 101 is connected to the main database 103, and the number generated by the short message service processing process is The data is selectively discarded.
  • the short message system only saves part or all of the data.
  • the data having no effect on the success rate of sending the short message is preferentially discarded.
  • the operation mode configuration of the short message system is mainly completed through the operation and maintenance center 102.
  • the related configuration items mainly include:
  • Configuration switching mode The switching mode between various operating modes can be manual switching or adaptive switching or a combination of both. If it is a manual switchover, the manual switchover is performed according to the field environmental indicators of the short message system (including real-time performance statistics, traffic flow, system alarm types, etc.); if it is an adaptive switchover, the service processor 101 according to the system Load indicators (such as traffic pressure and database service availability) complete the switching of the operating mode by themselves according to the configured switching strategy. The switching of various operating modes takes effect immediately without resetting the system.
  • the load index of the short message system adopts one or any combination of CPU usage rate, system read-write wait ratio, and cache usage rate; and the real-time acquisition of the system load index may use the commonly used detection CPU usage rate in the industry
  • the system read and write wait ratio (10 wait) application program interface (API) functions obtained.
  • API application program interface
  • the switching condition in the adaptive switching mode is: in the normal mode, if the load index exceeds the load mode load value for a set time, it automatically switches to the pressure mode; in the pressure mode If the load index is less than the load-resistance threshold of the compression mode for a set time, it will automatically switch to the normal mode.
  • the short message system For example, if the short message system is running in the normal mode and the load indicator of the service processor 101 exceeds the normal level for 10 minutes due to traffic storms and other reasons, the system automatically switches from the normal mode to the stress-resistant mode; after the traffic storm has passed, After the short message system detects that the load index returns to normal level and lasts for 10 minutes, it switches from the anti-stress mode to the normal mode.
  • the method described in this embodiment can make the short message system root According to the current traffic pressure and the available status of the database server, the appropriate operating mode and the corresponding data scheduling strategy are selected to process the data generated by the short message service, thereby achieving the purpose of enhancing the short message system's ability to resist traffic shocks.
  • the short message service processing can be divided into two independent processing processes: the initiation processing of the short message and the end processing of the short message.
  • the initiation process refers to the calling user sending a short message to the short message system;
  • the final call process refers to the short message system delivering the short message to the destination user.
  • Figure 2 shows a schematic diagram of the short message service processing process.
  • Call processing (thread) 210 After receiving the call request and processing, on the one hand, the complete short message is stored in the main database 103, and on the other hand, the brief information of the short message (including the short message identification number ID, the destination user number) Etc.) Insert into the short message buffer queue 220, and then send a response to the calling user that the call was successful. After the above operations are completed, the call processing 210 receives and processes the next call request. Call processing 210 No matter what the situation, unconditionally answer "11 ⁇ 2" call request and process it.
  • Termination processing (thread) 230 According to the principle of first in, first out, the clear information of the short message is retrieved from the short message buffer queue 220, and the corresponding short message is searched from the main database 103 for termination, and the termination processing After completion, the changed information is saved into the main database 103 again. After each termination, the termination of the next message is continued.
  • the service processing process of the short message is divided into a short message initiation process and a short message end call.
  • the terminal call processing thread 230 can determine whether to compare the system load indicator detected by the short message system in real time with the configured terminal call suspension load threshold to determine whether End call processing is performed to decompress the system. If the current system load indicator exceeds the termination call suspension load threshold, the termination call processing is suspended; once the system load indicator is lower than the termination call suspension load threshold, the termination call processing is continued.
  • the terminating thread can also perform terminating processing unconditionally. After performing the above operations, compare the system load index with the switching conditions, and then perform the subsequent steps.
  • the following configuration items need to be configured through the operation and maintenance center 102: whether to terminate the call unconditionally; and the threshold of the termination call suspension load when the call is not unconditionally terminated.
  • the termination call suspension load threshold is less than or equal to the compression mode load threshold.
  • the operation mode of the short message system further includes a disaster tolerance mode.
  • Disaster recovery mode is the operation mode when the main database of the short message system cannot provide services (such as system downtime, software upgrade, etc.).
  • the switching conditions for switching between the disaster tolerance mode and the above two operating modes are: Whether the primary database of the short message system fails: If the primary database fails in normal mode or stress mode, switch to disaster tolerance Mode; if the primary database returns to normal in the disaster recovery mode, switch back to the operating mode of the short message system before the primary database fails.
  • the business processor of the short message system is connected to the standby database and obtains the information required for business processing from the standby database.
  • the standby database 104 is a standby data storage device.
  • the service processor 101 is connected to the standby database 104 for short message service processing, and is to be used by the primary
  • the service processor 101 disconnects from the standby database 104 again, restores the connection with the primary database 103, and switches back to the operating mode of the short message system before the failure of the primary database 103.
  • the following configuration items need to be configured through the operation and maintenance center 102: a standby database type, a node location, and the like in the disaster recovery mode, and a data scheduling policy in the disaster recovery mode.
  • the data scheduling strategy in the disaster tolerance mode is to save only short messages to be resent.
  • the data scheduling strategy can also be set so that all short messages are not saved.
  • the standby database is regularly imported with the key of the main database Data is synchronized with data.
  • FIG. 3 is a schematic diagram showing a relationship between various operating modes of a short message system in a method for intelligently scheduling data according to an embodiment of the present invention.
  • a normal mode a stress-resistant mode, and a disaster tolerance mode can be adopted.
  • FIG. 4 is a service processing flowchart of a method for intelligently scheduling data according to an embodiment of the present invention in a stress-resistant mode.
  • step 410 is performed, and the short message system disconnects the standby database and connects to the main database.
  • step 420 the short message system performs normal service processing.
  • step 430 a type of the data operation is determined. If it is a data read request, in step 440, perform a normal data read operation, and then return to step 420 to continue the next round of business processing. If it is a data write request, in step 450, the data is discarded or discarded according to the currently selected data scheduling policy, and then returns to step 420 to continue the next round of business processing.
  • each load level is configured with a corresponding load index threshold and a corresponding data scheduling strategy.
  • the service processor needs to monitor the usage of the short message system such as the CPU, input / output interface 1/0, and buffers in real time, so as to determine the load level of the short message system based on the traffic volume.
  • FIG. 5 a flowchart of a method for intelligently scheduling data according to an embodiment of the present invention for selecting a data scheduling strategy in a stress-resistant mode.
  • the measured current load index is compared with the load index thresholds of each load level to determine the current load level of the short message system. In this embodiment, after comparison, it can be determined that the current load level is a high level .
  • the data scheduling strategy corresponding to the load level is selected in step 530, and the data scheduling strategy corresponding to each load level is: at a higher level At other times, only the successfully sent short messages are discarded (that is, not saved); at a high level, expired short messages, permanently failed short messages (more than 3 retransmission failures), and short message receipts are discarded; at a super high level When it is finished, all the processed short messages are further discarded, such as short messages waiting for retransmission. Therefore, in this embodiment, the data scheduling policy should adopt "discard successfully sent short messages, expired short messages, permanently failed short messages, and short message receipts".
  • the number of load levels and the corresponding data scheduling strategy in the anti-stress mode can be flexibly set. It should be noted that if the load index of the current short message system is smaller than the load index threshold of a higher level, but the short message system is still in a stress-resistant mode, the data scheduling strategy corresponding to the higher level is adopted.
  • the load levels in the above compression mode, the data scheduling strategy under each load level, and the load indicator thresholds for each load level need to be configured through the operation and maintenance center 102 in the short message system.
  • the anti-stress mode can be further divided into several modes corresponding to different traffic pressures, and these modes are configured with switching conditions and data scheduling policies.
  • this embodiment further divides different load levels in the compression mode, and uses different data scheduling strategies to decompress the short message system for different load levels.
  • the strategy is to follow As the system load level gradually increases, successful short messages, expired short messages, permanent failed short messages, and short message receipts are discarded until all processed short messages are completed.
  • the temporary response of the short message is not temporarily delayed to ensure the response rate of message transmission, and when the discarded data type is selected, the success rate of message transmission is preferentially guaranteed at any time.
  • FIG. 6 is a service processing flowchart of a method for intelligently scheduling data in a disaster tolerance mode according to an embodiment of the present invention.
  • the service processor disconnects from the primary database and connects to the standby database.
  • step 620 normal service processing is performed.
  • step 630 the type of the data operation is judged. If it is a data read request (such as reading user information for authentication), step 640 is performed from the standby database.
  • step 650 Read the necessary information for business processing, and then return to step 620 to continue the next round of business processing; if it is a data write request (such as saving a processed short message), in step 650, according to the data in the disaster recovery mode Scheduling strategy Save or discard the data, in this embodiment only save the short message to be resent, and then return to step
  • the disaster tolerance mode is adopted in this embodiment, the catastrophic impact of the short message system due to the failure of the database service can be prevented, the robustness, stability, and flexibility of the system are enhanced, and the disaster tolerance is achieved.
  • the method is simple and easy, saving costs.
  • the various operating modes proposed by the present invention as embodiments can be freely combined and changed according to application requirements.
  • the pressure-resistant mode can be further detailed Divided into ordinary traffic pressure mode and high-risk traffic pressure mode, these modes are also configured with switching conditions and data scheduling strategies;
  • the load indicator of the short message system can also use other indicators such as cache utilization, and weighting algorithms for multiple indicators In terms of data scheduling strategies, if the load continues to increase, all unnecessary short messages, such as timed short messages, can be further discarded.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Telephonic Communication Services (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Description

短消息系统中智能调度数据的方法
技术领域
本发明涉及通信领域的各类短消息系统, 尤其涉及短消息系统中数据 的调度方法。 背景技术
短消息作为电信行业最重要的增值业务, 地位日渐重要, 现有的 CDMA 网络、 GSM 网络和固定电话网中都设置了短消息系统或类似产品。 在为广 大用户带来便捷服务、 为运营商带来丰厚利润的同时, 各方对短消息系统 的要求也越来越严格, 现在短消息系统的安全性和健壮性已向交换机等主 设备看齐。 在任何情况下都保证短消息系统的正常运作是最基本的需求。
目前, 短消息系统的基本架构和数据处理模式决定了它在抗话务冲击 和防范数据库服务器故障等方面有很大局限, 处理手段较为单一。 普遍釆 用的用于抗话务冲击的过负荷方法会拒绝或緩发大量的正常消息, 导致短 消息系统发送率和响应率低下, 而且在长时间的大话务冲击下, 如果积累 的緩发消息达到一定量级, 则可能产生溢出并导致消息丢失, 直至整个系 统由于负荷过高而宕机。
例如公开日为 2004年 1月 14日的中国专利申请 02123903. 7中公开了 一种应对大话务冲击短消息系统的方法, 该方法在负载增大时的处理策略 主要是暂存緩发, 但这会导致系统的响应率低下; 该方法还通过将短消息 的临时错误修改为永久错误, 从而丟弃发送失败的短消息。 然而, 话务压 力越高, 导致的发送失败就越多, 如果直接丟弃失败消息而不重发, 会使 系统的发送成功率大幅降低。
数据库服务器是短消息系统的核心部件, 一旦发生故障会对整个短消 息系统产生致命影响, 因此也需要增强其安全性。 现有技术中一般采用系 统冗余容灾技术对整个系统进行 有效控制整个系统的故障发生率, 但其实现较为复杂并且成 ^艮高。 因此, 目前的短消息系统急需一种高效易行的智能调度数据的方法, 以增强系统的健壮性和安全性, 同时保证消息发送的响应率和成功率。 发明内容
本发明正是鉴于现有技术中的上述问题而提出的, 其目的在于提供一 种短消息系统中智能调度数据的方法, 可以使短消息系统根据业务量的变 化采用相应的运行模式以对数据进行智能调度, 从而提高短消息系统抗话 务冲击的能力, 并保证短消息系统的发送率和响应率。
本发明提供了一种短消息系统中智能调度数据的方法, 根据当前短消 息系统的负荷指标选择系统的运行模式, 并按照所选的运行模式对短消息 业务数据进行调度。
优选地, 本发明所述短消息系统中智能调度数据的方法, 具体包括以 下步骤:
将短消息系统的运行模式分为正常模式和话务压力过大时的抗压模 式,配置模式间的切换方式和切换条件, 以及各模式对应的数据调度策略, 在抗压模式下对短消息业务处理过程产生的数据有选择性地进行丢弃; 检测短消息系统的负荷指标;
将所述负荷指标与所述切换条件进行比较, 在满足所述切换条件时, 按配置的切换方式完成短消息系统运行模式间的切换; 以及
在短消息业务处理过程产生数据时, 按照当前运行模式所对应的数据 调度策略对数据进行调度。
优选地, 所述智能调度数据的方法还包括: 将短消息的业务处理过程 分为短消息的起呼处理和短消息的终呼处理两个独立的处理过程, 并设置 终呼暂停负荷阈值; 其中,
在将所述负荷指标与所述切换条件进行比较的步骤之前, 还包括: 在 正常模式下, 如果所述负荷指标超过所述终呼暂停负荷阈值, 则暂停短消 息的终呼处理; 如果所述负荷指标低于所述终呼暂停负荷阈值, 则进行终 呼处理。 由上可知, 本发明方法提供了短消息系统的多种运行模式, 各模式对 应于不同的数据调度策略, 并可以灵活地相互切换, 使短消息系统可以根 据业务量的变化釆用相应的运行模式, 从而提高了系统抗话务冲击能力, 增强了系统健壮性、 稳定性和灵活性, 由于没有改变现有的发送策略, 因 此也保证了系统的响应率。 另一方面, 本发明还可以将短消息业务处理过 程分为起呼处理和终呼处理两个独立的过程, 并提供了暂停终呼处理的系 统减压方式, 可以发挥系统潜能并减小浪涌式话务冲击对短消息系统的影 响。 进一步地, 通过优先保存影响发送成功率的短消息数据, 因而在话务 压力大时优先保证了短消息的成功发送。 另夕卜,本发明由于采用容灾模式, 可以防范短消息系统因数据库月 务器故障而导致的灾难性影响, 并以筒易 方式实现了容灾, 节约了系统成本。 附图说明
图 1是采用根据本发明的一个实施例的智能调度数据的方法的短消息 系统的功能模块示意图;
图 2是短消息业务处理过程的示意图;
图 3是根据本发明的一个实施例的智能调度数据的方法中短消息系统 各运行模式之间的关系示意图;
图 4是根据本发明的一个实施例的智能调度数据的方法在抗压模式下 的业务处理流程图;
图 5是根据本发明的一个实施例的智能调度数据的方法在抗压模式下 选择数据调度策略的流程图;
图 6是根据本发明的一个实施例的智能调度数据的方法在容灾模式下 的业务处理流程图。 具体实施方式
相信通过下面结合附图对本发明的优选实施例的详细说明, 可以更清 楚地了解本发明的上述和其它目的、 特征和优点。 本发明的基本思想是: 对短消息系统设置多种运行模式, 根据当前短 消息系统的负荷指标选择系统的运行模式, 并按照所选的运行模式对短消 息业务数据进行调度。
根据本发明的一个实施例 , 首先将短消息系统的运行模式分为正常模 式和抗压模式, 其中正常模式为现有技术的基本运行模式, 抗压模式定义 为短消息系统在话务压力过大时的运行模式。 然后配置各运 "模式间的切 换方式和切换条件, 以及各运行模式对应的数据调度策略, 其中在抗压模 式下, 对短消息业务处理过程产生的数据进行有选择性地丟弃。 检测系统 的负荷指标,将其与上述切换条件进行比较; 如果负荷指标满足切换条件, 则按照上述切换方式完成运行模式间的切换。 对于短消息业务处理过程中 产生的数据, 按照短消息系统的当前运行模式所对应的数据调度策略对其 进行调度。
图 1是采用根据本发明的一个实施例的智能调度数据的方法的短消息 系统的功能模块示意图, 该短消息系统主要包括业务处理机 101、 与该业 务处理机 101相连的操作维护中心 102和主用数据库 103。 其中业务处理 机 101是进行短消息业务处理的核心部件, 各种运行模式都在业务处理机 101上实现。 操作维护中心 102是人机接口, 用于提供配置运行模式的人 机界面。 主用数据库 103是主用的数据存储设备, 用于存放短信息系统中 的主要数据, 如业务处理所需的用户信息、 业务处理生成的短消息数据等 等。
以下对短消息系统的运行模式进行说明, 如上所述, 这些运行模式都 在业务处理机 101上实现。
正常模式: 是短消息系统无异常时的运行模式。 在该模式下, 业务处 理机 101连接主用数据库 103, 其业务和数据的处理均沿袭现有标准的处 理流程; 系统正常读取业务处理所需的用户信息, 保存业务处理产生的数 据(如短消息等) ; 此外短消息系统对数据进行完备的保存。
抗压模式: 是短消息系统话务压力过大时的运行模式。 在该模式下, 业务处理机 101连接到主用数据库 103, 对短消息业务处理过程产生的数 据有选择性地进行丟弃; 此外在该模式下, 短消息系统仅对数据进行部分 保存或全部不保存。 优选地, 为了优先保证短消息发送的成功率, 在丟弃 数据时, 优先丟弃对短消息发送成功率无影响的数据。
对短消息系统的运行模式进行配置主要通过操作维护中心 102完成, 相关的配置项主要包括:
1 )配置切换方式: 各运行模式间的切换方式可以是人工切换或者自适 应切换或者两者结合。 如果是人工切换, 则才艮据短消息系统的现场环境指 标(包括实时的性能统计、 话务流量、 系统告警类型等)进行手工切换; 如果是自适应切换, 则由业务处理机 101根据系统负荷指标(如话务压力 和数据库服务可用状况等)按照配置的切换策略自行完成运行模式的切换, 各种运行模式切换即时生效, 无需复位系统。 优选地, 短消息系统的负荷 指标采用的是 CPU使用率、 系统读写等待比以及緩存使用率中的一种或任 意組合; 而系统负荷指标的实时获取可以使用业界普遍使用的检测 CPU使 用率、 系统读写等待比 (10 wai t )的应用程序接口 (API ) 函数获得。 例 如, 在自适应切换方式下, 如果短消息系统处于正常模式, 则检测系统的 负荷指标,如果发现负荷指标满足切换条件, 则自动切换到抗压模式运行, 直到系统的负荷指标满足从抗压模式到正常模式的切换条件后, 再从抗压 模式切换回正常模式。
2 )配置切换条件: 例如对于自适应切换方式, 采用切换检测时长和切 换检测次数等。优选地, 自适应切换方式下的切换条件是: 在正常模式下, 如果所述负荷指标超过抗压模式负荷阐值且持续一设定时间, 则自动切换 到抗压模式; 在抗压模式下, 如果负荷指标小于抗压模式负荷阈值且持续 一设定时间, 则自动切换到正常模式。 例如, 短消息系统运行在正常模式 下,由于话务风暴等原因造成业务处理机 101的负荷指标持续 10分钟超过 正常水平, 则系统自动由正常模式切换到抗压模式; 等到话务风暴过后, 短消息系统检测到负荷指标回归到正常水平并持续 10分钟后,从抗压模式 切换回正常模式。
通过上述描述可以知道, 采用本实施例所述方法可以使短消息系统根 据当前的话务压力和数据库服务器可用状况等指标, 选择合适的运行模式 和对应的数据调度策略对短消息业务产生的数据进行处理, 从而达到增强 短消息系统抗话务冲击能力的目的。
为了充分利用系统的硬件资源, 提升应用程序的处理性能, 在业务处 理实施上可以将短消息业务处理分为两个彼此独立的处理过程: 短消息的 起呼处理和短消息的终呼处理, 其中, 起呼处理是指主叫用户发送短消息 到短消息系统; 终呼处理是指短消息系统将短消息投送到目的用户。 图 2 给出了短消息业务处理过程的示意图。 起呼处理(线程) 210接收起呼请 求并处理后, 一方面将完整的短消息保存到主用数据库 103中, 另一方面 将短消息的简明信息(包括短消息标识号 ID、 目的用户号码等)插入短消 息緩冲队列 220中, 然后向主叫用户发送起呼成功的响应, 上述操作完成 后, 起呼处理 210接收并处理下一个起呼请求。 起呼处理 210无论什么情 况均无条件地接》1½呼请求并处理。 终呼处理(线程) 230按先进先出的 原则从短消息緩冲队列 220中取出短消息的筒明信息, 并由此从主用数据 库 103中查找对应得短消息进行终呼, 终呼处理完成后, 将改动的信息再 次保存到主用数据库 103中。每个终呼完成后继续进行下一条消息的终呼。
在本发明的另一个实施例中, 在将短消息系统的运行模式分为正常模 式和抗压模式的同时, 将短消息的业务处理过程分为短消息的起呼处理和 短消息的终呼处理两个独立的处理过程, 并且配置了终呼暂停负荷阔值, 这样, 终呼处理线程 230可以根据短消息系统实时检测的系统负荷指标与 配置的终呼暂停负荷阈值的比较结果, 决定是否进行终呼处理, 从而为系 统减压。 如果当前系统负荷指标超过了终呼暂停负荷阈值, 则暂停终呼处 理; 一旦系统负荷指标低于终呼暂停负荷阈值, 则继续进行终呼处理。 当 然终呼线程也可以无条件进行终呼处理。 在执行完上述操作后, 再将系统 负荷指标与切换条件进行比较, 然后执行后续步骤。
在该实施例中, 需要通过操作维护中心 102配置下述配置项: 是否无 条件终呼; 非无条件终呼时的终呼暂停负荷阔值。 优选地, 终呼暂停负荷 阈值小于或等于抗压模式负荷阈值。 通过上述描述可知, 采用本实施例, 可以对系统的暂时过负荷(即系 统指标超过了正常水平, 但又没有达到切换条件)进行减压, 一方面发挥 了系统的潜能, 削平了浪涌式话务冲击对系统的冲击影响, 另一方面通过 与抗压模式相结合, 为短消息系统提供应对不同话务量的数据调度策略, 对于系统的业务处理能力和抗冲击能力有本质的提升。
在本发明的另一个实施例中, 为了防止短消息系统因为数据库的故障 而导致灾难性影响, 短消息系统的运行模式还包括容灾模式。 容灾模式是 在短消息系统的主用数据库无法提供服务(如系统宕机、 软件升级等) 时 的运行模式。 容灾模式与上述两个运行模式之间进行切换的切换条件是: 短消息系统的主用数据库是否发生故障: 如果在正常模式或抗压模式下, 主用数据库发生故障, 则切换到容灾模式; 如果在容灾模式下, 主用数据 库恢复正常, 则切换回主用数据库发生故障前短消息系统的运行模式。 在 容灾模式下, 短消息系统的业务处理机连接到备用数据库, 从备用数据库 获取业务处理所需的信息。 如图 1所示, 备用数据库 104是备用的数据存 殳备, 作为主用数据库 103的备份, 在容灾模式下, 业务处理机 101连 接到备用数据库 104上进行短消息业务处理, 待主用数据库 103恢复正常 后, 业务处理机 101再断开与备用数据库 104的连接, 恢复与主用数据库 103的连接, 并切换回主用数据库 103故障前短消息系统所处的运行模式。
在该实施例中, 需要通过操作维护中心 102配置下述配置项: 容灾模 式下的备用数据库类型、 节点位置等, 以及容灾模式下的数据调度策略。 在本实施例中, 容灾模式下的数据调度策略是仅保存待重发的短消息, 当 然也可以设定数据调度策略为所有短消息都不保存。
优选地, 在备用数据库中只保存业务处理所必须的数据, 对于业务处 理过程产生的数据, 只保存待重发的短消息或者全部不保存; 备用数据库 通过定时导入的方式与主用数据库的关键数据进行数据同步。
通过以上描述可知, 采用本实施例, 可以保证短消息系统在主用数据 库发生故障时仍可正常运行, 增强了系统的健壮性, 并且大大节约了备用 数据库的设备成本。 图 3给出了根据本发明的一个实施例的智能调度数据的方法中短消息 系统各运行模式之间的关系示意图, 其中正常模式、 抗压模式和容灾模式 三种模式之间都可以采用人工切换或者自适应切换的切换方式。
图 4是根据本发明的一个实施例的智能调度数据的方法在抗压模式下 的业务处理流程图。 当短消息系统切换到抗压模式下后, 如果短消息系统 是从容突模式切换过来的, 则执行步骤 410, 短消息系统断开备用数据库, 连接到主用数据库上。 在步骤 420, 短消息系统进行正常的业务处理。 当 在业务处理过程中发起数据操作请求时, 在步骤 430, 判断数据操作的类 型。 如果是数据读取请求, 则在步驟 440中, 进行正常的数据读取操作, 然后返回步骤 420, 继续下一轮的业务处理。 如果是数据写入请求, 则在 步驟 450中, 按照当前选择的数据调度策略^^或丟弃该数据, 然后返回 步骤 420, 继续下一轮的业务处理。
下面描述如何选择步骤 450中使用的数据调度策略。 在抗压模式中, 还可以进一步地划分有若干负荷级别, 每一负荷级别配置对应的负荷指标 阈值以及相应的数据调度策略。 当短消息系统 ^抗压模式后, 业务处理 机需实时监测短消息系统的如 CPU、输入输出接口 1/0、緩存等的使用情况, 从而判断话务量对短消息系统的负荷级别, 根据不同的负荷级别选择数据 调度策略。 如图 5所示的根据本发明的一个实施例的智能调度数据的方法 在抗压模式下选择数据调度策略的流程图, 在本实施例中, 假设抗压模式 中划分了三个负荷级别: 较高、 高和超高, 三个负荷级别的负荷指标阈值 为: 较高级别为 CPU使用率 -60%, 读写等待比 I0wai t=30%; 高级别为 CPU 使用率 =70%, 读写等待比 I0wai t=40%; 超高级别为 CPU使用率 =80%, 读写 等待比 I0wai t=50%。 在步骤 510, 检测短消息系统当前的负荷指标, 假设 测得的当前负荷指标为 CPU使用率 =70%, 读写等待比 I0wai t=45%。 在步驟 520 中, 将测得的当前负荷指标与各负荷级別的负荷指标阈值进行比较, 确定短消息系统当前的负荷级别, 在本实施例中, 经过比较可以确定当前 的负荷级别为高级别。 确定了负荷级别后, 在步骤 530中选择该负荷级别 所对应的数据调度策略, 而各负荷级别对应的数据调度策略是: 在较高级 别时, 只丢弃(即不保存)成功发送的短消息; 在高级别时, 进一步丟弃 过期短消息、 永久失败的短消息(重发失败 3次以上)和短消息回执; 在 超高级别时, 再进一步丢弃所有处理完成的短消息, 如等待重发的短消息 等。 因此本实施例中, 数据调度策略应采用 "丟弃成功发送的短消息、 过 期短消息、 永久失败的短消息和短消息回执 " 。
当然, 抗压模式下负荷级别的级数和对应的数据调度策略可以灵活地 设置。 需要说明的是, 如果当前短消息系统的负荷指标小于较高级别的负 荷指标阔值, 但短消息系统还处于抗压模式, 则采用较高级别所对应的数 据调度策略。
上述抗压模式下的负荷级别、 各负荷级别下的数据调度策略以及各负 荷级别的负荷指标阔值都需要通过短消息系统中的操作维护中心 102进行 配置。
此外, 抗压模式还可进一步分为若干对应于不同话务压力的模式, 并 对这些模式进行切换条件和数据调度策略的配置。
通过以上的描述可以看出, 本实施例在抗压模式中进一步划分了不同 的负荷级别, 并针对不同的负荷級别采用了不同的数据调度策略对短消息 系统进行减压, 其策略是随着系统负荷级别的逐步升高依次丟弃成功短消 息、 过期短消息、 永久失败短消息、 短消息回执直至所有处理完成的短消 息。 本实施例通过不对短消息进行暂存緩发, 保证了消息发送的响应率, 而在选择丢弃的数据类型时, 任何时候都优先保证消息发送的成功率。
图 6是根据本发明的一个实施例的智能调度数据的方法在容灾模式下 的业务处理流程图。 当短消息系统切换到容灾模式后, 在步骤 610, 业务 处理机断开与主用数据库的连接, 连接到备用数据库上。 然后在步骤 620, 进行正常的业务处理。 在业务处理过程中, 当发起数据操作请求时, 在步 骤 630中, 判断数据操作的类型, 如果是数据读取请求(如读取用户信息 进行鉴权), 则执行步驟 640, 从备用数据库中读取业务处理必要的信息, 然后返回步骤 620, 继续下一轮的业务处理; 如果是数据写入请求(如保 存处理过的短消息) , 则在步骤 650中, 根据容灾模式下的数据调度策略 保存或丢弃该数据, 在本实施例中仅保存待重发的短消息, 然后返回步骤
620, 继续下一轮的业务处理。
通过以上描述可以看出, 本实施例采用容灾模式后, 可以防范短消息 系统因数据库服务故障而导致的灾难性影响, 增强了系统的健壮性、 稳定 性和灵活性, 并且实现容灾的方式简单易行, 节约了成本。
应当理解的是, 本发明作为实施例提出的各种运行模式可以视应用需 要进行自由組合和变换,例如,可以只采用正常模式和抗压模式进行组合; 又如还可以将抗压模式进一步细分为普通话务压力模式和高危话务压力模 式, 对这些模式也配置切换条件和数据调度策略; 此外短消息系统的负荷 指标还可以采用緩存使用率等其它指标,对于多个指标可采用加权算法等; 在数据调度策略方面, 如果负荷持续增大, 还可以进一步丟弃所有非必要 的短消息, 如定时短消息等等。

Claims

权利要求书
1. 一种短消息系统中智能调度数据的方法, 才艮据当前短消息系统的 负荷指标选择系统的运行模式, 并按照所选的运行模式对短消息业务数据 进行调度。
2. 根据权利要求 1 所述的短消息系统中智能调度数据的方法, 其特 征在于, 包括以下步骤:
将短消息系统的运行模式分为正常模式和话务压力过大时的抗压模 式, 配置模式间的切换方式和切换条件, 以及各模式对应的数据调度策略, 在抗压模式下对短消息业务处理过程产生的数据有选择性地进行丟弃; 检测短消息系统的负荷指标;
将所述负荷指标与所述切换条件进行比较, 在满足所述切换条件时, 按配置的切换方式完成短消息系统运行模式间的切换; 以及
在短消息业务处理过程产生数据时, 按照当前运行模式所对应的数据调度 策略对数据进行调度。
3. 根据权利要求 2 所述的短消息系统中智能调度数据的方法, 其特 征在于, 所述智能调度数据的方法还包括: 将短消息的业务处理过程分为 短消息的起呼处理和短消息的终呼处理两个独立的处理过程, 并设置终呼 暂停负荷阈值; 其中,
在将所述负荷指标与所述切换条件进行比较的步骤之前, 还包括: 在 正常模式下, 如果所述负荷指标超过所述终呼暂停负荷阈值, 则暂停短消 息的终呼处理; 如果负荷指标低于终呼暂停负荷阈值, 则进行终呼处理。
4. 根据权利要求 2或 3所述的短消息系统中智能调度数据的方法, 其特征在于, 所述切换方式配置为人工切换或自适应切换或两者的结合。
5. 根据权利要求 2或 3所述的短消息系统中智能调度数据的方法, 其特征在于, 所述自适应切换方式下的切换条件是: 在正常模式下, 如果 所述负荷指标超过抗压模式负荷闹值且持续一设定时间, 则自动切换到抗 压模式; 在抗压模式下, 如果负荷指标小于抗压模式负荷阈值且持续一设 定时间, 则自动切换到正常模式。
6. 根据权利要求 2或 3所述的短消息系统中智能调度数据的方法, 其特征在于, 在所述抗压模式下对业务产生的数据进行选择性丟弃时, 优 先丢弃对短消息发送成功率无影响的数据。
7. 根据权利要求 6 所述的短消息系统中智能调度数据的方法, 其特 征在于, 所述抗压模式进一步分为若干负荷级别, 划分有若干负荷级别, 每一负荷级别配置对应的负荷指标阈值以及相应的数据调度策略; 所述方 法还包括:将测得的当前负荷指标与各负荷级别的负荷指标阈值进行比较, 确定短消息系统当前的负荷级别, 再选择对应的数据调度策略。
8. 根据权利要求 7 所述的短消息系统中智能调度数据的方法, 其特 征在于, 所述各负荷级别对应的数据调度策略应满足: 随着负荷级别的提 高, 按下述顺序调度数据: 丢弃发送成功的短消息; 丟弃过期短消息; 丟 弃永久失败短消息; 丢弃短消息回执; 丟弃所有处理完成的短消息。
9. 根据权利要求 2或 3所述的短消息系统中智能调度数据的方法, 其特征在于,所述负荷指标采用 CPU使用率、 系统读写等待比和緩存使用 率中的一种或任意组合。
10.根据权利要求 2或 3所述的短消息系统中智能调度数据的方法, 其特征在于, 所述系统的运行模式还包括在主用数据库故障时启用的容灾 模式; 在该模式下, 短消息系统连接到备用数据库进行业务处理, 等主用 数据库恢复正常后, 再切换回主用数据库发生故障前短消息系统的运行模 式, 断开与备用数据库的连接, 连接到主用数据库。
11.根据权利要求 10所述的短消息系统中智能调度数据的方法, 其特 征在于, 所述备用数据库只保存业务处理所必须的数据, 对于业务处理过 程产生的数据只保存待重发的短消息或全部不保存; 所述备用数据库通过 定时导入的方式与所述主用数据库的关键数据进行数据同步。
12.根据权利要求 2或 3所述的短消息系统中智能调度数据的方法, 其特征在于, 所述抗压模式进一步分为若干对应于不同话务压力的模式, 并对这些模式进行切换条件和数据调度策略的配置。
PCT/CN2004/001303 2004-05-25 2004-11-15 A intelligent data scheduling method of short message system Ceased WO2005117461A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200410042583.0 2004-05-25
CNB2004100425830A CN100341345C (zh) 2004-05-25 2004-05-25 一种短消息中心多模式数据调度处理方法

Publications (1)

Publication Number Publication Date
WO2005117461A1 true WO2005117461A1 (en) 2005-12-08

Family

ID=34601619

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2004/001303 Ceased WO2005117461A1 (en) 2004-05-25 2004-11-15 A intelligent data scheduling method of short message system

Country Status (2)

Country Link
CN (1) CN100341345C (zh)
WO (1) WO2005117461A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118095444A (zh) * 2024-04-23 2024-05-28 创新奇智(青岛)科技有限公司 大模型推理的优化方法、装置、电子设备及存储介质

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101588556B (zh) * 2008-05-23 2012-03-28 中兴通讯股份有限公司 一种短消息过负荷处理方法
CN101350660B (zh) * 2008-07-22 2012-08-29 京信通信系统(中国)有限公司 数字射频拉远系统及其备用载波信道控制方法
CN101674611A (zh) * 2009-09-16 2010-03-17 中兴通讯股份有限公司 一种短消息中心过负荷控制的方法和系统
CN102394940B (zh) * 2011-11-30 2014-06-25 中国工商银行股份有限公司 自适应数据传输量控制系统和方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1440596A (zh) * 2000-07-05 2003-09-03 艾利森电话股份有限公司 基于排队/信道使用的分组数据的功率控制算法
CN1467952A (zh) * 2002-07-08 2004-01-14 华为技术有限公司 一种无线移动网络中的短消息调度管理方法
CN1468010A (zh) * 2002-07-12 2004-01-14 华为技术有限公司 一种无线移动网络中的短消息调度管理方法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6449488B1 (en) * 1999-08-25 2002-09-10 Lucent Technologies Inc. Quality of service based CDMA broadcast scheduler

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1440596A (zh) * 2000-07-05 2003-09-03 艾利森电话股份有限公司 基于排队/信道使用的分组数据的功率控制算法
CN1467952A (zh) * 2002-07-08 2004-01-14 华为技术有限公司 一种无线移动网络中的短消息调度管理方法
CN1468010A (zh) * 2002-07-12 2004-01-14 华为技术有限公司 一种无线移动网络中的短消息调度管理方法

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118095444A (zh) * 2024-04-23 2024-05-28 创新奇智(青岛)科技有限公司 大模型推理的优化方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN1585517A (zh) 2005-02-23
CN100341345C (zh) 2007-10-03

Similar Documents

Publication Publication Date Title
CN101188527B (zh) 一种心跳检测方法和装置
CN112965823B (zh) 调用请求的控制方法、装置、电子设备和存储介质
CN101599987A (zh) 消息队列管理方法和装置
CN113467969A (zh) 一种处理消息堆积的方法
CN101543100A (zh) 一种处理容灾切换的方法、装置及系统
CN100464600C (zh) 告警处理方法和后台管理装置
CN105376785B (zh) 网络通信功能异常的处理方法、应用处理器及移动终端
CN113992609B (zh) 一种处理多链路业务数据乱序的方法及系统
CN112564990B (zh) 一种用于音频管理服务器切换的管理方法
CN102006314B (zh) 多服务器自适应任务调度方法及装置
CN114979250A (zh) 消息推送方法、装置和设备
WO2005117461A1 (en) A intelligent data scheduling method of short message system
CN102413191A (zh) 一种基于浏览器的web即时通讯方法
CN105517030B (zh) 网络通信功能异常的处理方法、调制解调器及移动终端
CN110636035B (zh) 一种通信方法、装置及可读存储介质
CN102004665A (zh) 工作流系统中任务并发处理方法
CN102143598B (zh) 信令处理方法和系统以及网络侧设备和终端
US9042396B2 (en) System and method for processing multimedia messaging service
CN100429948C (zh) 共路信令链路系统的过载控制方法
CN103684875A (zh) 过载保护方法以及设备过载保护装置
CN1330154C (zh) 一种交换系统中接续消息的处理方法
CN102263661A (zh) 一种用户会话的备份切换方法及系统
CN114286374B (zh) 拥塞等级检测及处理方法、系统、设备及存储介质
CN114173007B (zh) 一种通信方法、装置以及系统
CN101599848A (zh) 告警处理方法和装置

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase

Ref country code: DE

WWW Wipo information: withdrawn in national office

Ref document number: DE

WWE Wipo information: entry into national phase

Ref document number: 7629/DELNP/2006

Country of ref document: IN

122 Ep: pct application non-entry in european phase