評估效能和容量需求的社交環境 (SharePoint Server 2013)Estimate performance and capacity requirements for social environments (SharePoint Server 2013)

摘要:了解如何決定數目和類型所需的容量計劃的 「 我的網站和 SharePoint Server 2013 為基礎的社交運算入口網站的電腦。Summary: Learn how to determine the number and types of computers that you need for a capacity plan for a My Site and social computing portal based on SharePoint Server 2013.

若要建立效能和容量規劃 「 我的網站的企業內部網路與社交運算入口網站解決方案,本文會包含下列領域的相關資訊:To create a performance and capacity plan for an enterprise intranet My Site and social computing portal solution, this article contains information about the following areas:

  • 實驗室環境規格,如硬體、 伺服器陣列拓撲及伺服器陣列設定Lab environment specifications, such as hardware, farm topology, and farm configuration

  • 測試伺服器陣列工作負載與資料集是用來產生測試負載The test farm workload and dataset that was used to generate test load

  • 測試結果與分析︰ 示範與說明在特定刻度點的負載下,輸送量、 延遲及硬體需求的趨勢。Test results and analysis that demonstrate and explain trends in throughput, latency, and hardware demand under load at specific scale points.

使用本文中的資訊來瞭解下列概念:Use the information in this article to understand the following concepts:

  • 一般和尖峰負載下案例的特性Characteristics of the scenario under both normal and peak loads

  • 效能趨勢如何變更時向外延展伺服器陣列的伺服器How performance trends change when farm servers are scaled out

  • 如何估算適當您計劃的架構起點How to estimate an appropriate starting point for your planned architecture

  • 考慮當您計劃的資源在伺服器陣列的重要因素必須維持可接受的層級的尖峰負載下的效能Important factors to consider when you plan for the resources your farm will need to maintain acceptable levels of performance under peak load

環境簡介Introduction to this environment

企業通常會使用 SharePoint Server 2013 發佈我的網站及經過驗證的使用者存取內部網路網站上的社交運算入口網站。本文包含可協助您規劃要使用的電腦數目和類型的 SharePoint Server 2013 中發佈 「 我的網站與社交運算入口網站所需的電腦的容量與效能資料。Enterprises often use SharePoint Server 2013 to publish My Site and social computing portals that authenticated users access on an intranet site. This article contains capacity and performance data to help plan the number of computers to use and the types of computers that are required to publish My Site and social computing portals in SharePoint Server 2013.

其他指引說明如何向外延展 SharePoint Server 2013 企業 「 我的網站和社交運算入口網站解決方案中的伺服器。容量計劃會告知硬體,以最佳化您的解決方案的購買和系統設定的相關決策。Additional guidance explains how to scale out servers in a SharePoint Server 2013 enterprise My Site and social computing portal solution. Capacity planning informs decisions about hardware to purchase and system configurations that optimize your solution.

個別 SharePoint Server 2013 伺服器陣列是唯一的因為每個伺服器陣列會有不同的硬體、 使用者行為、 的已安裝的功能、 設定及許多其他因素而定的需求。因此,補充這份指導您自己環境中硬體上的其他測試。如果您計劃的設計和工作負載的格式類似於本文所述的環境,您可以使用本文來繪製結論有關如何擴充您的環境。Because individual SharePoint Server 2013 farms are unique, each farm has different requirements that depend on hardware, user behavior, the configuration of installed features, and many other factors. Therefore, supplement this guidance with additional testing on your own hardware in your own environment. If your planned design and workload resembles the environment described in this article, you can use this article to draw conclusions about how to scale your environment.

本文中的測試結果產生在測試實驗室中,來模擬高度控制的情況下為實際執行環境中使用工作負載、 資料集和架構。雖然小心已運用在設計這些測試,在測試實驗室的效能特性一些永遠不在實際執行環境的行為相同。這些測試結果不代表實際執行伺服器陣列的效能與容量特性。而是測試結果示範觀察到的輸送量、 延遲及硬體需求的趨勢。若要協助您規劃容量及管理自己的伺服器陣列使用觀察到的資料的分析。Test results in this article were produced in a test lab, using a workload, dataset, and architecture to simulate a production environment under highly controlled conditions. While great care was exercised in designing these tests, the performance characteristics of a test lab are never the same as the behavior of a production environment. These test results do not represent the performance and capacity characteristics of a production farm. Instead, the test results demonstrate observed trends in throughput, latency, and hardware demand. Use the analysis of the observed data to help you plan capacity and manage your own farm.

本文包括下列資訊︰This article includes the following:

  • 規格的硬體、 拓撲及設定Specifications, which include hardware, topology, and configuration

  • 工作量,包括需求的伺服器陣列、 使用者數目及使用狀況特性的分析The workload, which includes an analysis of the demand on the farm, the number of users, and usage characteristics

  • 資料集,例如資料庫大小和內容類型The dataset, such as database sizes and content types

  • 測試結果與分析向外延展網頁伺服器Test results and analysis to scale out web servers

閱讀本文之前,請閱讀下列文章,以確認您了解 SharePoint Server 2013 中的容量管理的重要概念。Before you read this article, read the following articles to make sure that you understand the key concepts behind capacity management in SharePoint Server 2013.

這些文章提供下列資訊︰These articles provide the following information:

  • 容量管理的建議方法The recommended approach to capacity management

  • 如何有效利用本文中的資訊How to make effective use of the information in this article

詞彙Glossary

下列清單包含之本文中找到的關鍵字詞的定義:The following list contains definitions for key terms found in this article:

  • RPS:每秒要求數目。RPS 為伺服器陣列或伺服器會收到一秒的要求數目。這是常見的伺服器與伺服器陣列負載度量單位。RPS: Requests per second. RPS is the number of requests that a farm or server receives in one second. This is a common measurement of server and farm load.

    重要

    請注意要求跟頁面載入。每個頁面包含數個元件,每一種會建立一或多個要求瀏覽器中載入頁面時。因此,一個載入頁面建立數個要求。不在 RPS 測量計算驗證檢查及事件通常用於微乎其微的資源。Note that requests differ from page loads. Each page contains several components, each of which creates one or more requests when a browser loads a page. Therefore, one page load creates several requests. Authentication checks and events that use insignificant resources typically are not counted in RPS measurements.

  • 綠色區域:綠色區域代表一組已定義下正常作業的情況,最多個預期每日的尖峰負載的負載特性。在這個範圍中運作的伺服器陣列應該能夠維持回應時間與介於參數可接受的延遲。Green Zone: Green Zone represents a defined set of load characteristics under normal operating conditions, up to expected daily peak loads. A farm that operates in this range should be able to sustain response times and latency that are within acceptable parameters.

    在此狀態下,伺服器可維持下列準則︰This is the state at which the server can maintain the following set of criteria:

    • 至少 75%的要求之伺服器端延遲為小於 0.5 秒。The server-side latency for at least 75 percent of the requests is less than 0.5 seconds.

    • 所有伺服器都維持平均 CPU 使用率小於 50%。All servers maintain an average CPU utilization of less than 50 percent.

    • 失敗率為小於 0.1%。Failure rate is less than 0.1 percent.

  • 紅色區域 (Max):紅色區域代表一組已定義的尖峰作業的情況下的負載特性。紅色區域在伺服器陣列的錯誤率非常高暫時性的資源需求愈多,它可以維持僅針對有限的段前失敗並發生其他效能和可靠性的問題。Red Zone (Max): Red Zone represents a defined set of load characteristics under peak operation conditions. At Red Zone, the farm experiences very high transient resource demands that it can sustain only for limited periods before failures and other performance and reliability issues occur.

    在此狀態下,伺服器可在有限的持續時間內維持下列準則︰This is the state at which the server can maintain the following set of criteria for a limited duration:

    • 至少 75%的要求伺服器端的延遲是少於 1 秒。Server-side latency for at least 75 percent of the requests is less than 1 second.

    • 平均資料庫伺服器 CPU 使用率是小於 80%。Average database server CPU utilization is less than 80 percent.

    • 失敗率為小於 0.1%。Failure rate is less than 0.1 percent.

概觀Overview

本節摘要說明我們延展方法,此實驗室環境和類似的案例研究環境中,與我們測試方法之間的關係。This section summarizes our scaling approach, the relationship between this lab environment and a similar case study environment, and our test methodology.

延展方法Scaling approach

我們建議您在特定順序我們後面的縮放比例我們的測試實驗室環境中對環境中向外的電腦。這個方法可讓您尋找您工作負載的最佳組態。We recommend that you scale the computers in your environment in the specific order we followed for scaling our test lab environment. This approach will enable you to find the best configuration for your workload.

我們會將效能測試循環分成三種工作負載類別。決定類別界限的主要參數是數字的使用者設定檔,設為在 10k、 100 K 和 500 位使用者設定檔測試。另一個參數是作用中使用者所執行的社交功能集相關的動作數目。同時具有設定檔的使用者數目及作用中使用者數,我們已執行測試以模擬就是類似實際部署的應用程式的使用方式。下表說明初始的資料集和作用中使用者數。We divided the performance test cycles into three workload categories. The primary parameter that determined the category boundary was number of user profiles, which was set at 10K, 100K and 500K user profile tests. Another parameter was the number of active users, who were carrying out actions related to the social set of features. With both the number of users with a profile and number of active users, we ran tests to simulate usage of the application that would be similar to actual deployments. The following table depicts the initial data set and the number of active users.

初始資料集Initial Data Set

實體Entity %的使用者具有此功能% of users with this feature 小型 (10 位使用者)Small (10K users) 中型 (100 位使用者)Medium (100K users) Large (500 位使用者)Large (500K users)
使用者設定檔的使用者數目Number of user profiles for users
100%100%
10K10K
100 K100K
500 K500K
已佈建 「 我的網站的數目Number of provisioned My Sites
100%100%
10K10K
100 K100K
500 K500K
使用者相片的使用者設定檔的數目Number of user profiles that have user photos
50%50%
5 K5K
50K50K
250 K250K
文章的使用者設定檔的數目Number of user profiles that have posts
10%10%
1 K1K
10K10K
50K50K
小組的數目Number of teams
1,8601,860
18,60018,600
93 K93K
每日的作用中使用者數Number of active users per day
10%10%
1 K1K
10K10K
50K50K
每小時的作用中使用者數Number of active users per hour
5%5%
500 個500
5 K5K
25 K25K

測試著重於下列主要案例:Testing focused on the following key scenarios:

  • 新聞摘要] 頁面上存取及其他動作News Feed page access and other actions

  • 設定檔頁面Profile page

  • 網站摘要] 頁面上存取及其他動作Site feed page access and other actions

  • Outlook Social Connector 活動摘要同步處理Outlook Social Connector Activity Feed Sync

  • 商務] 頁面上存取 OneDriveOneDrive for Business page access

  • OneDrive for Business 用戶端使用狀況OneDrive for Business client usage

若要模擬真實感化的部署案例、 已經有資料的資料庫上執行所有測試。資料集是樹狀目錄組織的平均每小組、 4-6 使用者與 3-4 層級深入的模型。若要產生這些數字,我們分析來自內部的社交網站的流量。下表說明我們用來建立初始的資料集的參數組。To simulate a realistic deployment scenario, all tests were run on a database that already had data. The dataset was a model of a tree organization with an average of 4-6 users per team, and 3-4 levels deep. To generate these numbers, we analyzed traffic from an internal social site. The following table describes the set of parameters that we used to build the initial data set.

初始資料庫的資料模型Data model for initial database

資料實體描述Data entity description 數字Number
在小組中使用者的平均數目Average number of users in on team
55
組織各層級的平均數目Average number of levels per organization
44
小組每 1000 位使用者的數目Number of teams per 1,000 users
186186
接在使用者的同事的平均數目Average number of colleagues a user follows
5050
使用者設定檔屬性的數目Number of User Profile properties
9393

下表說明一組參數就會導致資料填入的動作:The following table describes the set of parameters in terms of actions that would result in the data population:

使用狀況特性Usage characteristics

參數Parameter 數字或百分比Number or percentage
1-3 文章使用使用者的百分比Percentage of users with 1-3 posts
10%10%
文章每位使用者的平均數目Average number of posts per user
22
每篇文章的回覆的平均數目Average number of replies per post
22
會按讚的文章的百分比Percentage of posts that are Liked
15%15%
文章的連結的百分比Percentage of posts with links
5%5%
文章標籤的百分比Percentage of posts with tags
12%12%
與使用者提及文章的百分比Percentage of posts with user mentions
8%8%
具有附加的圖像的文章的百分比Percentage of posts with image attached
5%5%

若要建立每個規模測試,我們將套用下列巨集指令混合到以上的資料集和作用中使用者數:To create each of our scale tests, we applied the following action mix to the preceding data set and the number of active users:

使用者讀取動作User READ Actions

使用者動作User action %的使用者利用此巨集指令% of user taking this action 案例Scenario 功能或 URLFeature or URL
瀏覽至 [我的網站首頁Navigate to My Site home page
12%12%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
瀏覽至使用者的公用設定檔頁面Navigate to the user's public profile page
8%8%
設定檔Profile
設定檔頁面 (http://my/person.aspx?accountname=<別名>)Profile page (http://my/person.aspx?accountname=<alias>)
瀏覽至使用者的私人的設定檔頁面Navigate to the user's private profile page
4%4%
設定檔Profile
設定檔] 頁面 (http://my/person.aspx)Profile page (http://my/person.aspx)
自動同步處理活動摘要Automatic syncing of activity feed
32%32%
Outlook Social ConnectorOutlook Social Connector
none
瀏覽至 [我正在關注的人員] 頁面上Navigate to the People I'm following page
%33%
追蹤人員清單Follow People List
http://my/MyPeople.aspx
瀏覽至 [預設文件庫Navigate to the default document library
6%6%
OneDrive for BusinessOneDrive for Business
https://msft-my.spoppe.com/personal/<使用者>/文件https://msft-my.spoppe.com/personal/<user>/Documents
瀏覽至 [後面文件] 頁面Navigate to followed documents page
%33%
OneDrive for BusinessOneDrive for Business
https://msft-my.spoppe.com/personal/<使用者>/Social/FollowedContent.aspxhttps://msft-my.spoppe.com/personal/<user>/Social/FollowedContent.aspx
瀏覽至 [後面文件] 頁面Navigate to followed documents page
%33%
OneDrive for BusinessOneDrive for Business
https://msft-my.spoppe.com/personal/<使用者>/Social/FollowedContent.aspxhttps://msft-my.spoppe.com/personal/<user>/Social/FollowedContent.aspx
瀏覽至 [網站摘要] 頁面Navigate to the site feed page
8%8%
網站摘要Site Feed
網站摘要] 頁面 (https://<網域>/teams/<網站>/newsfeed.aspx_Site Feed page (https://<domain>/teams/<site>/newsfeed.aspx_
檢視所有回覆的執行緒上View all replies on a thread
1%1%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
檢視所有人摘要View Everyone feed
%33%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
檢視新聞摘要上的其他文章View more posts on the newsfeed
2%2%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
檢視 @mentions 頁面View the @mentions page
1%1%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
檢視新聞摘要 (行動電話)View newsfeed (Mobile)
1%1%
手機Mobile
行動裝置的代表性狀態傳輸 (REST) 通話Mobile Representational State Transfer (REST) Call
檢視分類的新聞摘要View categorized newsfeed
%33%
手機Mobile
行動裝置的其餘呼叫Mobile REST Call

使用者寫入的動作User WRITE Actions

使用者動作User action PercentagePercentage 案例Scenario 功能或 URLFeature or URL
建立根文章的摘要Create root post in the feed
0.5%0.5%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
像摘要中的文章Like a post in the feed
0.3%0.3%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
回覆摘要中的文章Reply to a post in the feed
0.7%0.7%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
在含有 @mention 摘要中建立貼文Create post in the feed with @mention
0.1%0.1%
新聞摘要Newsfeed
新聞摘要] 頁面 (http://my/default.aspx)Newsfeed page (http://my/default.aspx)
在 [網站摘要中建立根文章Create root post in the site feed
0.5%0.5%
網站摘要Site Feed
網站摘要] 頁面 (https://<網域>/teams/<網站>/newsfeed.aspx)Site feed page (https://<domain>/teams/<site>/newsfeed.aspx)
網站與 @mention 摘要中建立貼文Create post in the site feed with @mention
0.5%0.5%
網站摘要Site Feed
網站摘要] 頁面 (https://<網域>/teams/<網站>/newsfeed.aspx)Site feed page (https://<domain>/teams/<site>/newsfeed.aspx)
在 [網站摘要發文回覆Reply to a post in the site feed
0.15%0.15%
網站摘要Site Feed
網站摘要] 頁面 (https://<網域>/teams/<網站>/newsfeed.aspx)Site feed page (https://<domain>/teams/<site>/newsfeed.aspx)
在網站具有標記摘要中建立貼文Create post in the site feed with a tag
0.05%0.05%
網站摘要Site Feed
網站摘要] 頁面 (https://<網域>/teams/<網站>/newsfeed.aspx)Site feed page (https://<domain>/teams/<site>/newsfeed.aspx)

商務用戶端動作 OneDriveOneDrive for Business client actions

使用者動作User action PercentagePercentage 案例Scenario 功能或 URLFeature or URL
OneDrive for Business 初始同步處理OneDrive for Business initial sync
0.2%0.2%
OneDrive for BusinessOneDrive for Business
初始同步處理Initial Sync
OneDrive for Business 累加同步處理-下載檔案OneDrive for Business incremental sync - download a file
0.88%0.88%
OneDrive for BusinessOneDrive for Business
累加同步處理Incremental Sync
OneDrive for Business 累加同步處理-沒有變更OneDrive for Business incremental sync - no changes
8.1%8.1%
OneDrive for BusinessOneDrive for Business
累加同步處理Incremental Sync

測試方法Test Methodology

我們開始使用社交功能的基本 SharePoint Server 2013 伺服器陣列設定。我們套用至測試伺服器陣列的特性的社交負載,並增加負載直到我們觀察到層級的一般和最大伺服器容量。我們分析在每個這些的載入層級和向外延展伺服器陣列設定的多載角色新增的機器瓶頸。此除了減輕每個瓶頸並針對特定資料集所提供的伺服器延展性特性的檢視。我們的容量規劃重複此向外延展的程序提供的 SharePoint Server 2013 伺服器陣列的延展性特性代表性合併彙算的三種部署大小和指導方針。We started with a minimum SharePoint Server 2013 farm configuration for social features. We applied a characteristic social load to the test farm and increased the load until we observed levels of normal and maximum server capacity. We analyzed bottlenecks at each of these load levels and added machines of the overloaded role to scale out the farm configuration. This addition alleviated the bottlenecks in each case and provided a view of scalability characteristics of the server for a particular dataset. We repeated this scale-out process for three deployment sizes to provide representative summaries of a SharePoint Server 2013 farm's scalability characteristics and guidelines for capacity planning.

規格Specifications

本節提供硬體、 軟體、 拓撲及設定的實驗室環境的詳細的資訊。This section provides detailed information about the hardware, software, topology, and configuration of the lab environment.

重要

Al 網頁伺服器和應用程式伺服器測試實驗室中的已套用到虛擬化使用 HYPER-V 主機。資料庫伺服器已不在虛擬化。實體主機硬體和虛擬機器虛擬硬體詳細說明分開下列各節。Al web servers and application servers in the test lab were virtualized by using Hyper-V hosts. Database servers were not virtualized. The physical host hardware and virtual machine virtual hardware are detailed separately in the following sections.

硬體Hardware

下表列出此測試中所使用之電腦的硬體規格。在多個重複項目測試期間新增至伺服器陣列的前端網頁伺服器也符合與這些規格。The following table lists hardware specifications for the computers that were used in this test. Front-end web servers that were added to the server farm during multiple iterations of the test also complied with these specifications.

Hyper-V 主機Hyper-V Hosts

伺服器陣列包含三個同名設定 HYPER-V 主機,總共和每個主機執行一至四個虛擬機器。The farm includes a total of three identically configured Hyper-V hosts, and each host runs one to four virtual machines.

主機硬體 * * **Host hardware* Value
處理器Processor(s)
2 個四核心 2.27 GHz 處理器2 Quad-core 2.27 GHz processors
RAMRAM
64 GB64 GB
作業系統Operating System
Windows Server 2008 R2 SP1Windows Server 2008 R2 SP1
網路介面卡數目Number of network adapters
22
網路介面卡速度Network adapter speed
1 Gigabit1 Gigabit

虛擬網頁伺服器與應用程式伺服器Virtual web servers and application servers

在伺服器陣列具有一至八部虛擬網頁伺服器。其他專用虛擬伺服器執行分散式快取服務。The farm has from one to eight virtual web servers. An additional dedicated virtual server runs the Distributed Cache Service.

注意

在實際執行環境中,執行分散式快取服務的專用的伺服器通常被部署中的高度可用的組態。測試用途,我們使用單一專用的伺服器的分散式快取因為高可用性不是關鍵要素。In a production environment, dedicated servers that run the Distributed Cache Service are typically deployed in a highly available configuration. For test purposes, we used a single dedicated server for Distributed Cache because high availability was not a critical factor.

VM 硬體VM hardware 網頁伺服器Web servers
處理器Processors
4 個虛擬處理器4 virtual processors
RAMRAM
12 GB12 GB
作業系統Operating system
Windows Server 2008 R2 SP1Windows Server 2008 R2 SP1
SharePoint 磁碟機的大小Size of the SharePoint drive
建立在內容資料庫中且可供轉譯文件的空間。根據預設,可供轉譯文件的快取為 100 GB。我們不建議您增加可用的快取。100 GB
網路介面卡數目Number of network adapters
22
網路介面卡速度Network adapter speed
1 Gigabit1 Gigabit
驗證Authentication
Windows NTLMWindows NTLM
負載平衡器類型Load balancer type
F5 Big IPF5 Big IP
在本機上執行的服務Services running locally
Microsoft SharePoint Foundation Web 應用程式、 Microsoft SharePoint Foundation 內送電子郵件、 Microsoft SharePoint Foundation 工作流程計時器服務、 受管理的中繼資料 Web 服務、 使用者設定檔服務Microsoft SharePoint Foundation Web Application, Microsoft SharePoint Foundation Incoming E-Mail, Microsoft SharePoint Foundation Workflow Timer Service, Managed Metadata Web Service, User Profile Service
VM 硬體VM hardware 快取Cache
處理器Processors
4 個虛擬處理器4 virtual processors
RAMRAM
12 GB12 GB
作業系統Operating system
Windows Server 2008 R2 SP1Windows Server 2008 R2 SP1
SharePoint 磁碟機的大小Size of the SharePoint drive
建立在內容資料庫中且可供轉譯文件的空間。根據預設,可供轉譯文件的快取為 100 GB。我們不建議您增加可用的快取。100 GB
網路介面卡數目Number of network adapters
22
網路介面卡速度Network adapter speed
1 Gigabit1 Gigabit
驗證Authentication
Windows NTLMWindows NTLM
在本機上執行的服務Services running locally
分散式快取、 Microsoft SharePoint Foundation 工作流程計時器服務Distributed Cache, Microsoft SharePoint Foundation Workflow Timer Service
VM 硬體VM hardware 搜尋查詢元件Search query component
處理器Processors
4 個虛擬處理器4 virtual processors
RAMRAM
12 GB12 GB
作業系統Operating system
Windows Server 2008 R2 SP1Windows Server 2008 R2 SP1
網路介面卡數目Number of network adapters
22
網路介面卡速度Network adapter speed
1 Gigabit1 Gigabit
驗證Authentication
Windows NTLMWindows NTLM
在本機上執行的服務Services running locally
Microsoft SharePoint Foundation Web 應用程式、 Microsoft SharePoint Foundation 內送電子郵件、 Microsoft SharePoint Foundation 工作流程計時器服務、 搜尋查詢及網站設定服務、 SharePoint Server 搜尋Microsoft SharePoint Foundation Web Application, Microsoft SharePoint Foundation Incoming E-Mail, Microsoft SharePoint Foundation Workflow Timer Service, Search Query and Site Settings Service, SharePoint Server Search
VM 硬體VM Hardware 搜尋索引元件Search index component
處理器Processors
4 個虛擬處理器4 virtual processors
RAMRAM
12 GB12 GB
作業系統Operating system
Windows Server 2008 R2 SP1Windows Server 2008 R2 SP1
網路介面卡數目Number of network adapters
22
網路介面卡速度Network adapter speed
1 Gigabit1 Gigabit
驗證Authentication
Windows NTLMWindows NTLM
在本機上執行的服務Services running locally
Microsoft SharePoint Foundation Web 應用程式、 Microsoft SharePoint Foundation 內送電子郵件、 Microsoft SharePoint Foundation 工作流程計時器服務、 SharePoint Server 搜尋Microsoft SharePoint Foundation Web Application, Microsoft SharePoint Foundation Incoming E-Mail, Microsoft SharePoint Foundation Workflow Timer Service, SharePoint Server Search

資料庫伺服器Database servers

一部實體資料庫伺服器執行具有 SharePoint 資料庫的預設 SQL Server 執行個體。本文不會追蹤記錄資料庫。One physical database server runs the default SQL Server instance that has the SharePoint databases. This article does not track the logging database.

注意

如果您啟用流量報告,我們建議您將儲存在記錄資料庫上個別的邏輯單位編號 (LUN)。大型部署與某些中型部署可能需要專用的記錄資料庫伺服器以容納大量的記錄事件所產生的處理器需求。> 在此實驗室環境中,記錄受到限制,且記錄資料庫儲存在 SQL Server 的個別執行個體。If you enable usage reporting, we recommend that you store the logging database on a separate Logical Unit Number (LUN). Large deployments and some medium deployments might require a dedicated logging database server to accommodate the demand on the processor that a high volume of logging events generates. > In this lab environment, logging was constrained, and the logging database was stored in a separate instance of SQL Server.

資料庫伺服器-預設執行個體Database Server - Default Instance

處理器Processors
2 個四核心 3.3 GHz 處理器2 Quad-core 3.3 GHz processors
RAMRAM
32 GB32 GB
作業系統Operating system
Windows Server 2008 R2 SP1Windows Server 2008 R2 SP1
儲存體與幾何Storage and geometry
直接連接儲存裝置 (DAS)Direct Attached Storage (DAS)
內部與 6 x 300 GB 15krpm 磁碟陣列Internal array with 6 x 300 GB 15krpm disk
外部與 15 x 450 GB 15krpm 磁碟陣列External array with 15 x 450 GB 15krpm disk
50 x 內容資料 (外部 RAID10,2 x 3 磁針每個 300 GB)50 x content data (external RAID10, 2x3 spindles 300 GB each)
50 x 內容記錄檔 (內部 RAID10 2x2 磁針每個 300 GB)50 x content logs (internal RAID10, 2x2 spindle 300 GB each)
1 x 暫時資料 (內部 RAID10,2x2 磁針每個 300 GB)1 x temp data (internal RAID10, 2x2 spindles 300 GB each)
1 x 暫時記錄 (內部 RAID10,2x2 磁針每個 300 GB)1 x temp log (internal RAID10, 2x2 spindles 300 GB each)
網路介面卡數目Number of network adapters
11
網路介面卡速度Network adapter speed
1 Gigabit1 Gigabit
驗證Authentication
Windows NTLMWindows NTLM
軟體版本Software version
SQL Server 2008 R2SQL Server 2008 R2

拓撲Topology

下表說明此實驗室環境的拓撲:The following table shows the topology for this lab environment:

實驗室環境的拓撲Lab environment topology

角色Role 小型部署 (10 位使用者)Small deployment (10k users) 中型部署 (100 位使用者)Medium deployment (100K users) 大型部署 (500 位使用者)Large deployment (500K users)
網頁伺服器Web server
2-42-4
4-84-8
88
快取Cache
11
1-21-2
33
SQL ServerSQL Server
11
1-21-2
22

測試程序Test Process

重要

測試僅模型一般營業時間使用一般的社交運算入口網站。我們並未在一天/晚上循環產生使用者產生流量考量循環的變更。我們測試過等設定檔同步處理和人員搜尋編目需要單獨具有相同的測試工作負載來判斷其效果的大幅資源的計時器工作。> 測試聚焦社交作業,例如新聞、 社交標記及讀取人員設定檔。測試混合包含少量的一般共同作業流量更妥善地模擬在實際執行環境。我們預期這些結果有助於設計專用於 「 我的網站及社交功能的個別入口網站。> 測試混合不包含來自搜尋內容編目的流量。>The tests only model normal business-hour usage on a typical social computing portal. We did not consider the cyclical changes in user-generated traffic that day/night cycles produce. We tested Timer jobs such as Profile Synchronization and People Search Crawl, which require significant resources, independently with the same test workload to determine their effect. > The tests focus on social operations, such as newsfeeds, social tagging, and reading people profiles. The test mix includes a small amount of typical collaboration traffic to better simulate a production environment. We expect these results will help to design a separate portal that is dedicated to My Sites and social features. > The test mix does not include traffic from the Search Content Crawl. >

我們執行測試針對小型、 中型及大型部署的社交功能。若要設定伺服器硬體,我們在最小大小的最小設定啟動並填入資料集的縮放比例方法] 區段中所述的測試資料庫。We conducted tests against small, medium, and large deployments for the social features. To configure server hardware, we started at minimum configurations for the smallest size and populated the test database with the dataset as described in the Scaling approach section.

我們使用 Visual Studio Team System (VSTS) 來模擬的工作量及套用特性社交負載主導對伺服器的小型負載第一次。我們統一緩慢增加此負載和直到我們觀察到最大 RPS 記錄在所有伺服器角色上的效能評量。這 was 辨識伺服器陣列上的套用負載增加其中因為沒有增加所造成的狀態傳遞 RPS 輸出伺服器瓶頸條件約束。We used Visual Studio Team System (VSTS) to simulate a workload and apply a characteristic social load, driving a small load against the server at first. We uniformly increased this load slowly and recorded performance metrics on all server roles until we observed maximum RPS. This was recognizable as the state where an increase of the applied load on the farm resulted in no increase in delivered RPS output because of server bottleneck constraints.

從這些記錄的評量,我們會定義綠色區域與紅色區域的狀態,它表示在特定的電腦設定的虛擬機器伺服器一般和滿載狀態。我們再套用在綠色區域與紅色區域層級來分析穩定狀態效能評量在這些載入穩定負載。提供此伺服器每種拓撲設定這些主要負載的情況下的虛擬機器伺服器健康狀況與效能的表示法。From these recorded metrics, we defined green zone and red zone states, which represent the normal and fully loaded states of the VM server at a given computer configuration. We then applied a steady load at both green zone and red zone levels to analyze steady-state performance metrics at these loads. This provided a server health and performance representation of the VM server under these key load conditions for each topology configuration.

我們瞭解綠色和紅色負載特性和每種拓撲的縮放比例的曲線上之後,我們會識別限制 RPS 的縮放比例瓶頸。但如果是社交工作量,這通常是小型資料集的網頁伺服器 CPU。大型資料集,我們也會觀察分散式快取節點上的記憶體壓力。我們要在每個案例中移除瓶頸,並繼續向外延展的程序設定新增虛擬伺服器的多載角色。我們再重複效能趨勢和綠色和紅色區域定義在每個設定的大小以其符合性的分析直到我們達到特定部署大小的需求。After we understood the green and red load characteristics and the scaling curve for each topology, we identified the scaling bottleneck that limited RPS. In the case of social workload, this was typically web server CPU for small datasets. For larger datasets, we also observed memory pressure on the Distributed Cache nodes. We added virtual servers of the overloaded role to the configuration to remove the bottlenecks in each case and continue the scale-out process. We then repeated the analysis of performance trends and their conformity to green and red zone definitions at each configuration size until we achieved requirements for a specific deployment size.

我們瞭解每個部署大小之後,我們重新設定最小的下一個較大的間距設定測試伺服器陣列、 填入 [縮放比例方法] 區段中所述的資料集和重複分析/向外延展程序循環和為單位的每一個資料集大小的向外延展特性。After we understood each deployment size, we reconfigured the test farm to the smallest configuration of the next larger size, populated the dataset as described in the Scaling approach section, and repeated the analysis/scale-out process cycle, and measured scale-out characteristics of each dataset size.

結果與分析Results and analysis

本節說明這三種部署大小的測量的結果。尤其是它會顯示向外延展伺服器陣列新增網頁伺服器的影響綠色和紅色區域的 RPS、 延遲及平均 CPU 使用率。This section shows the measured results for the three deployment sizes. Specifically, it shows how scaling out the server farm by adding web servers affects green and red zone RPS, latency, and average CPU usage.

下列趨勢跨所有三個部署大小是一致的:The following trends were consistent across all three deployment sizes:

  • 綠色和紅色區域的 RPS 以線性方式增加虛擬網頁伺服器的數目。Both red and green zone RPS increases linearly with the number of virtual web servers.

  • 跨所有已測試設定的主要瓶頸是網頁伺服器 CPU。The primary bottleneck across all tested configurations was the web server CPU.

  • 紅色區域在稍微的延遲增加為我們新增網頁伺服器,並增加負載。這是由 SQL Server 及分散式快取服務 (這執行測試伺服器陣列中的所有網頁伺服器) 上新增壓力所造成。At red zone, latency increases slightly as we add web servers and increase load. This is caused by added pressure on SQL Server and the Distributed Cache service (which is running on all web servers in the test farm).

  • 此外,在 SQL Server 與分散式快取的電腦上的平均 CPU 使用量增加網頁伺服器增加數。這是在 SQL Server 與分散式快取服務快取的額外負載所造成。Additionally, average CPU usage on SQL Server and Distributed Cache computers increases as the number of web servers increases. This is caused by additional caching load on the on SQL Server and the Distributed Cache service.

  • 綠色區域的延遲會維持單層為 web 伺服器個數增加。這是因為在網頁伺服器都不負荷過大時於綠色區域的載入層級。Green zone latency remains fairly flat as the number of web servers increases. This is because the web servers are not overburdened at green zone load levels.

小規模結果Small Scale Results

下圖顯示如何增加網頁伺服器的數目會影響 RPS 對綠色與紅色區域。The following graph shows how increasing the number of web servers affects RPS for both green and red zones.

這個螢幕擷取畫面顯示在有 10000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的 RPS 產生何種影響。

下圖顯示如何增加網頁伺服器的數目會影響這兩個綠色和紅色區域負載層級的延遲。The following graph shows how increasing the number of web servers affects latency for both green and red zone load levels.

這個螢幕擷取畫面顯示在有 10000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的延遲產生何種影響。

下圖顯示如何增加網頁伺服器的數目會影響對這兩個綠色和紅色區域負載層級的平均 CPU 使用量。The following graph shows how increasing the number of web servers affects average CPU usage for both green and red zone load levels.

這個螢幕擷取畫面顯示在有 10000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的 CPU 使用量產生何種影響。

中型規模結果Medium Scale Results

下圖顯示如何增加網頁伺服器的數目會影響 RPS 對綠色與紅色區域。The following graph shows how increasing the number of web servers affects RPS for both green and red zones.

這個螢幕擷取畫面顯示在有 100000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的 RPS 產生何種影響。

下圖顯示如何增加網頁伺服器的數目會影響這兩個綠色和紅色區域負載層級的延遲。The following graph shows how increasing the number of web servers affects latency for both green and red zone load levels.

這個螢幕擷取畫面顯示在有 100000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的延遲產生何種影響。

下圖顯示如何增加網頁伺服器的數目會影響對這兩個綠色和紅色區域負載層級的平均 CPU 使用量。The following graph shows how increasing the number of web servers affects average CPU usage for both green and red zone load levels.

這個螢幕擷取畫面顯示在有 100000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的 CPU 使用量產生何種影響。

大型結果Large Scale Results

下圖顯示如何增加網頁伺服器的數目會影響 RPS 對綠色與紅色區域。The following graph shows how increasing the number of web servers affects RPS for both green and red zones.

這個螢幕擷取畫面顯示在有 500000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的 RPS 產生何種影響。

下圖顯示如何增加網頁伺服器的數目會影響這兩個綠色和紅色區域負載層級的延遲。The following graph shows how increasing the number of web servers affects latency for both green and red zone load levels.

這個螢幕擷取畫面顯示在有 500000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的延遲產生何種影響。

下圖顯示如何增加網頁伺服器的數目會影響對這兩個綠色和紅色區域負載層級的平均 CPU 使用量。The following graph shows how increasing the number of web servers affects average CPU usage for both green and red zone load levels.

這個螢幕擷取畫面顯示在有 500000 位使用者的案例中,增加前端網頁伺服器的數目會對綠色與紅色區域的 CPU 使用量產生何種影響。

網頁伺服器數目增加時,會發生下列事件:As the number of web servers increases, the following events occur:

  • 平均 CPU 使用量增加 SQL Server 與分散式快取節點因為這些共用資源上新增負擔。Average CPU usage increases for SQL Server and Distributed Cache nodes because of added burden on these shared resources.

  • 平均網頁伺服器 CPU 使用量紅色區域在稍微減少因瓶頸稍微移位至 SQL Server 與分散式快取的電腦。Average web server CPU usage at red zone slightly decreases because of bottleneck shifting slightly to SQL Server and Distributed Cache computers.

  • 因為伺服器會保留不在建議的載入層級平均網頁伺服器於綠色區域的 CPU 使用率維持不變。Average web server CPU usage at green zone remains constant because servers are kept at recommended load levels.

建議Recommendations

成功的 SharePoint Server 2013 社交部署是由效能衡量取決於下列因素:A successful SharePoint Server 2013 social deployment as measured by performance depends on the following factors:

  • 您要支援的作用中使用者的數目The number of active users who you want to support

  • 預期的交易混合的讀取和寫入作業The expected transaction mix of read and write operations

  • 負載分散於伺服器陣列中伺服器的方式How the load is distributed across the farm servers

預期的作用中使用者數是一個的關鍵因素決定您應該規劃拓撲中的伺服器的數目。作用中使用者數也會判斷組成主控的各種服務所需的社交案例跨伺服器已啟用。The expected number of active users is one key factor to determine the number of servers that you should plan to have in the topology. The number of active users also determines the makeup of hosting of the various services that are required to be enabled for the social scenario across the servers.

但是我們測試用典型資料集並套用您可能會預期在實際客戶部署中的負載複雜性,每個部署是唯一的。容量規劃投入比應考慮預期的使用狀況特性、 功能設定和硬體資源的可用性。一些因素可以影響或變更的容量數字的重大的方式如下:Though our testing used a typical dataset and applied the load complexity that you might expect in a real-world customer deployment, every deployment is unique. Your capacity planning effort should consider expected usage characteristics, feature configuration, and hardware resource availability. Some factors that can have an affect or change the capacity numbers in a significant way are as follows:

  • 增加的電子郵件流量模式可能會增加 Outlook Social Connector 所產生的負載。A pattern of increased email usage might increase the load that the Outlook Social Connector generates.

  • 大幅增加百分比的交易混合寫入動作 (例如增加的標記或 @mention) 可能會增加資料庫伺服器上的負載。A significant increase in the percentage of write actions (for example, an increase in tagging or @mention) in the transaction mix might increase the load on the database server.

  • 您可以新增或移除的網頁伺服器以平衡 CPU 負載網頁伺服器、 SQL Server 與分散式快取節點之間。You can add or remove web servers to balance CPU load between web servers, SQL Server, and Distributed Cache nodes.

謹慎遵循以達最佳效能標準的 SharePoint Server 2013 設定指導。特別針對社交交易重要的考量如下所示:Carefully follow standard SharePoint Server 2013 configuration guidance for optimal performance. Considerations that matter specifically for social transactions are as follows:

  • 不同的實體磁碟的設定檔資料庫-因為對設定檔資料庫可能造成社交交易的粗磁碟使用量我們建議您保留設定檔資料庫上執行 SQL Server 的伺服器上的實體磁碟一組自己。Separate physical disks for Profile DB - Because of the heavy disk usage that social transactions can have on Profile DB, we recommend that you keep Profile DB on its own set of physical disks on the server that runs SQL Server.

  • User Profile service 應用程式的記憶體需求的 User Profile service 應用程式位於前端網頁伺服器與嚴重依賴其於記憶體中快取。請確定前端網頁伺服器具有足夠的 RAM 快取許多要求的資料。最小值的建議 RAM 為 12 GB 每部前端網頁伺服器。Memory requirements for User Profile service application - The User Profile service application is located on front-end web servers and relies heavily on its in-memory cache. Make sure that front-end web servers have enough RAM to cache many requests for data. Minimum recommended RAM is 12 GB per front-end web server.

  • 分散式快取伺服器的記憶體需求-社交功能、 微型部落格特別是,而定嚴重不足與完善的分散式快取儲存。雖然此快取要重新填入這些電腦上的記憶體不足情況可能會降低 SharePoint 伺服器陣列的容量。因此建議您將主機分散式快取使用至少 12 GB 的 RAM,並向外延展伺服器設定為所需根據使用者總數部署中。Memory requirements for Distributed Cache servers- Social features, microblogging in particular, depend heavily on sufficient and robust Distributed Cache storage. Low memory situations on these computers can degrade the capacity of the SharePoint farm while this cache is being repopulated. Therefore we recommend that you configure servers that host the Distributed Cache to use at least 12 GB of RAM, and scaled out as needed based on total number of users in the deployment.

SharePoint Server 2013 社交部署會強制佈建針對每個想要使用社交功能的使用者個人網站。規劃建立個人網站集合層級的內容資料庫的成長。如需如何擴充個人網站集合的詳細資訊,請參閱SharePoint 2013 的軟體界限及限制The SharePoint Server 2013 social deployment makes it mandatory to provision a personal site for every user who wants to use social features. Plan the growth of the creation of personal site collections at the level of the content database . For more information about how to scale personal site collections, see Software boundaries and limits for SharePoint 2013.

另請參閱See also

概念Concepts

規劃 SharePoint Server 2013 中規劃效能Performance planning in SharePoint Server 2013

其他資源Other Resources

SharePoint 2013 的軟體界限及限制Software boundaries and limits for SharePoint 2013