CN108540557A - 一种基于动态限速的云应用负载调度方法 - Google Patents
一种基于动态限速的云应用负载调度方法 Download PDFInfo
- Publication number
- CN108540557A CN108540557A CN201810335819.1A CN201810335819A CN108540557A CN 108540557 A CN108540557 A CN 108540557A CN 201810335819 A CN201810335819 A CN 201810335819A CN 108540557 A CN108540557 A CN 108540557A
- Authority
- CN
- China
- Prior art keywords
- load
- speed limit
- server
- processing
- slo
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/101—Server selection for load balancing based on network conditions
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1097—Protocols in which an application is distributed across nodes in the network for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Computer And Data Communications (AREA)
Abstract
发明涉及一种基于动态限速的云应用负载调度方法。面向长时间运行的云应用,根据负载历史记录生成r‑b曲线以描述存储和网络利用率,基于动态规划为每类负载自动生成限速参数,并选择能够满足SLO要求的服务器。在保障处理负载的性能满足SLO的约束下,通过对自动化设置存储和网络限速参数,调度并整合负载以最小化处理负载的服务器数量,从而提高资源利用率并减少能耗。
Description
技术领域
本发明涉及一种基于动态限速的云应用负载调度方法,属于软件技术领域。
背景技术
在云计算环境中,云服务提供商为多个客户提供共享的计算、网络和存储资源以最大化资源利用率,降低总体能耗,从而减少数据中心的运营成本,同时保障良好性能,如满足请求处理延迟时间,以提升客户满意度。客户通常定义服务水平目标(SLO,ServiceLevel Object)以描述处理负载的性能要求,比如“80%的请求必须在100毫秒内完成”。那么,需要具有高效的负载调度方法,服务器在满足性能目标的前提下,能够处理多样化的负载。为了应对整合负载所带来的网络拥塞问题,云服务提供者和客户通常会达成限速协议,客户静态设定限速规则,服务提供商则进行相应的优化,以实现性能最大化。当前负载调度方法是,为客户预先保留一定数量的资源,或者以尽最大努力的方式处理负载。Silo(KeonJang, Justine Sherry, Hitesh Ballani, and Toby Moncaster. 2015. Silo:Predictable Message Latency in the Cloud. In ACM SIGCOMM. ACM, 435–448.)设计限速方案以保障网络延迟;pClock(Ajay Gulati, Arif Merchant, and Peter J.Varman. 2007. pClock: an arrival curve based approach for QoS guarantees inshared storage systems. In Proceedings of the 2007 ACM SIGMETRICSinternational conference on Measurement and modeling of computer systems.ACM, New York, NY, USA, 13–24.)设计限速方案以保障存储延迟;文献(Nohhyun Park,Irfan Ahmad, and David J. Lilja. 2012. Romano: Autonomous Storage ManagementUsing Performance Prediction in Multi-tenant Datacenters.In Proceedings ofthe Third ACM Symposium on Cloud Computing. ACM,New York, NY, USA, Article21, 14 pages.)采用性能预测的方法动态调整等待处理的负载。然而,面对大量负载,难以实现通过预留资源来满足处理延迟,尤其是,短期突发性负载会对处理延迟产生显著影响。同时,难以为不同的负载类型设置合理的限速参数。
发明内容
本发明的目的:面向长时间运行的云应用,在保障处理负载的性能满足SLO的约束下,将负载调度并整合到目标服务器,以最小化处理负载的服务器数量。
本发明的原理:刻画负载,描述其对处理延迟的影响,通过对存储和网络进行限速,并设置负载的优先级,在满足处理延迟的条件下,自动化的减少服务器数量。
本发明技术解决方案:一种基于动态限速的云应用负载调度方法,其特点在于实现步骤如下:
(1)生成r-b曲线以描述负载处理速率r与令牌桶容积b之间的关系。当请求到达时,令牌添加到令牌桶中,如果令牌桶中有足够的空间来添加令牌,即不超过令牌桶大小为b,则允许继续处理请求。否则,请求就会排队等待,直到令牌桶中有足够的空间。令牌以速率r不断地从桶中流出,空间逐渐变得可用。对于给定的r值,通过重放具有速率r以及无限大小令牌桶在任意时间点的执行轨迹,计算得到不需要排队的请求数量b。输入r与输出b构成<r,b>元组作为点,连接形成分段r-b曲线。对r值进行标准化处理(例如,网络流量除以网络带宽),那么,r=1.0表示负载占用了所有带宽资源。传输的数据量取决于请求类型(例如,读/写),分别生成不同的r-b曲线;
(2)选择限速参数:使用网络微积分方程计算,由于在服务器上排队而导致的处理延迟,对于优先级p的负载,处理延迟的上限为:
,
其中,<rj,bj>是负载j的限速设置,bj是负载j的令牌桶大小,rj是负载j的处理速率,pj是负载j的优先级,其高于或等于p,SLOp是与优先级p关联的SLO。进而可以得到下式:
,
使用分段线性凸函数r-b曲线,可以将bi表示为ri的函数,进而利用线性规划方法求解得到每个限速元组<rj,bj>,满足约束条件:
,
每个负载关联r-b曲线,当新的负载调度到该服务器时,动态重新计算现有负载共享该服务器的限速设置。
(3)选择处理负载的服务器:本发明通过线性规划求解,将负载分发给SLOs能够满足的服务器,采用首先匹配的策略。在通常情况下,大多数服务器几乎都是满载的,所以新负载不能分发给几乎满负载的服务器。因此,本发明提出了快速首次匹配方法,跟踪每个服务器上配置的速率总和,跳过将负载放到接近满负载的服务器,避免了不必要的运行线性规划计算过程。
本发明与现有技术相比具有如下优点:
(1)能够根据服务的实际资源使用状态,动态调整限速参数,即负载处理速度r和令牌桶容积b;
(2)在保障处理负载的性能满足SLO的约束下,将负载调度并整合到目标服务器,以最小化处理负载的服务器数量;
(3)能够满足在同一台服务器上,不同类型负载对于处理延迟的要求。
附图说明
图1为云应用负载调度系统架构。
具体实施方式
以下结合具体实施例和附图对本发明进行详细说明,如图1所示,本发明实施例方法流程:
本发明将阿里云弹性块存储云服务作为典型应用场景,客户连接到一个或多个块存储服务器实例,产生网络负载来访问存储在服务器上的数据。负载调度器由五个组件构成:
(1)r-b曲线生成器:根据处理负载的历史记录生成r-b曲线,描述负载的存储和网络利用率,并根据客户需求定义SLO;
(2)部署器:标识可以分发负载的候选服务器;
(3)优化器:为每个负载配置<r, b>限速参数,并决定在哪个服务器上放置负载来满足处理延迟要求;
(4)延迟检查器:确定负载的候选位置和<r, b>元组是否能够满足负载的SLO要求;
(5)实施器:配置适当的存储和网络限速,并将负载分配给服务器。
请求调度方法的执行流程如下:
(1)r-b曲线生成器根据负载类型的历史执行记录生成网络或内存的r-b曲线(即线性分段函数b=f(r)),并将用户定义的负载SLO要求,一同发送给部署器;
(2)部署器选择可以分发负载的服务器,并生成候选服务器列表,即存在较充足资源的服务器;
(3)优化器使用线性规划方法计算服务器上共存的各类负载的<r, b>元组参数,并将计算结果发送给延迟检查器:
,
其中,<rj,bj>是负载j的限速设置,bj是负载j的令牌桶大小,rj是负载j的处理速率,pj是负载j的优先级,其高于或等于p,SLOp是与优先级p关联的SLO满足约束条件:
。
(4)延迟检查器检测在候选服务器以及<r, b>元组是否能够满足用户所定义的SLO要求,如果能满足则将候选服务器以及<r, b>元组信息发送给实施器,否则重新发回部署器以生成新的方案;
(5)实施器将负载发送给选定的候选服务器,并设置<r, b>元组参数。
Claims (1)
1.方法特征在于实现步骤如下:
第一步,生成r-b曲线:根据处理负载的历史执行记录生成网络和内存的r-b曲线,其中,r是处理负载的速率,b是令牌桶容积,并定义负载的处理延迟要求;
第二步,选择候选服务器:选择较充足内存和网络资源的服务器,生成候选服务器列表;
第三步,计算参数r和b:使用线性规划方法计算服务器上共存的各类负载的<r,b>元组:,其中,rj和bj是负载类型j的参数,bj是负载j的令牌桶大小,rj是负载j的处理速率,p是用户定义的优先级,pj p是负载j的优先级, SLOp是与优先级p关联的处理延迟,满足约束条件:;
第四步,检测候选服务器以及限速参数r和b是否能够满足用户所定义的负载处理延迟要求,如果能满足则生成候选服务器以及限速参数r和b,否则重复第三步执行。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810335819.1A CN108540557A (zh) | 2018-04-16 | 2018-04-16 | 一种基于动态限速的云应用负载调度方法 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810335819.1A CN108540557A (zh) | 2018-04-16 | 2018-04-16 | 一种基于动态限速的云应用负载调度方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108540557A true CN108540557A (zh) | 2018-09-14 |
Family
ID=63480655
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810335819.1A Pending CN108540557A (zh) | 2018-04-16 | 2018-04-16 | 一种基于动态限速的云应用负载调度方法 |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108540557A (zh) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112242959A (zh) * | 2019-07-16 | 2021-01-19 | 中国移动通信集团浙江有限公司 | 微服务限流控制方法、装置、设备及计算机存储介质 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101174986A (zh) * | 2007-11-07 | 2008-05-07 | 中兴通讯股份有限公司 | 动态反馈负载分发方法和装置 |
CN102209107A (zh) * | 2011-05-11 | 2011-10-05 | 奇智软件(北京)有限公司 | 一种智能限速方法和装置、一种下载系统 |
CN103929366A (zh) * | 2014-04-02 | 2014-07-16 | 华为技术有限公司 | 流量控制方法、装置及设备 |
US20140365659A1 (en) * | 2013-06-05 | 2014-12-11 | Sap Ag | Load controller framework |
-
2018
- 2018-04-16 CN CN201810335819.1A patent/CN108540557A/zh active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101174986A (zh) * | 2007-11-07 | 2008-05-07 | 中兴通讯股份有限公司 | 动态反馈负载分发方法和装置 |
CN102209107A (zh) * | 2011-05-11 | 2011-10-05 | 奇智软件(北京)有限公司 | 一种智能限速方法和装置、一种下载系统 |
US20140365659A1 (en) * | 2013-06-05 | 2014-12-11 | Sap Ag | Load controller framework |
CN103929366A (zh) * | 2014-04-02 | 2014-07-16 | 华为技术有限公司 | 流量控制方法、装置及设备 |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112242959A (zh) * | 2019-07-16 | 2021-01-19 | 中国移动通信集团浙江有限公司 | 微服务限流控制方法、装置、设备及计算机存储介质 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20200314168A1 (en) | Distributed code execution involving a serverless computing infrastructure | |
US8930731B2 (en) | Reducing power consumption in data centers having nodes for hosting virtual machines | |
US8214843B2 (en) | Framework for distribution of computer workloads based on real-time energy costs | |
US20150067028A1 (en) | Message driven method and system for optimal management of dynamic production workflows in a distributed environment | |
US20160378570A1 (en) | Techniques for Offloading Computational Tasks between Nodes | |
US20160218950A1 (en) | Proactive identification of hotspots in a cloud computing environment | |
US20120323988A1 (en) | Task allocation in a computer network | |
US7925785B2 (en) | On-demand capacity management | |
US20140344398A1 (en) | Control systems and methods for cloud resource management | |
US8862729B2 (en) | Forecast-less service capacity management | |
US20150133214A1 (en) | Video encoding based on areas of interest | |
US20100251258A1 (en) | Recording medium having load balancing program recorded thereon, load balancing apparatus and method thereof | |
US20120324111A1 (en) | Task allocation in a computer network | |
WO2024016596A1 (zh) | 容器集群调度的方法、装置、设备及存储介质 | |
US9244718B2 (en) | Virtual machine resource allocation based on connection time coverage exceeding a minimum threshold | |
US20160292609A1 (en) | Optimization of workload placement | |
Stavrinides et al. | Cost‐aware cloud bursting in a fog‐cloud environment with real‐time workflow applications | |
CN113132456A (zh) | 一种基于截止时间感知的边云协同任务调度方法及系统 | |
CN112600761A (zh) | 一种资源分配的方法、装置及存储介质 | |
US20130227113A1 (en) | Managing virtualized networks based on node relationships | |
CN116225679A (zh) | 算力网络调度方法、装置及存储介质 | |
Deochake | Cloud cost optimization: A comprehensive review of strategies and case studies | |
CN108540557A (zh) | 一种基于动态限速的云应用负载调度方法 | |
US10979359B1 (en) | Polling resource management system | |
CN110308991A (zh) | 一种基于随机任务的数据中心节能优化方法及系统 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180914 |