The document summarizes performance testing of a Microsoft Office Communications Server (OCS) Enterprise Edition Consolidated Configuration deployed on an HP BladeSystem. Key findings from tests with 30,000 simulated users include:
1) Tests were conducted with user logon rates from 1 to 36 users per second to analyze the impact on server performance.
2) Two different storage solutions were tested for the SQL Server: an HP StorageWorks SB40c storage blade and an HP StorageWorks MSA1500 modular smart array.
3) Key server performance counters were monitored, including CPU utilization, queue and sproc latency, and SQL page life expectancy, to evaluate system health during user ramp-up and steady state.
Upgrading key components in servers can significantly improve performance and capacity. The report tested upgrading components in Dell PowerEdge servers. It found that upgrading to newer Dell PowerEdge R720 servers with Intel Xeon and SSD components, Windows Server 2012, and 10GbE networking supported 4.5 times as many VMs as the older Dell PowerEdge R710 servers. Each individual component upgrade, such as the processor, storage, or network card, increased performance. But upgrading all components together maximized the number of supported VMs, database transactions, and email users.
Dell 3-2-1 Reference Configurations: High available and scalable performance ...Principled Technologies
Dell 3-2-1 Reference Configurations consisting of Dell PowerEdge R720 servers, Dell Force10 S4810P switches, and Dell EqualLogic PS6110XV storage provide a range of configuration options so that you can select the one that is right for your business needs. Additionally, each configuration is easily scalable to accommodate for business growth.
The 3-2-1 base configuration provides a robust solution for any business needing to support up to 500 users on multiple business applications. Our tests demonstrate this solution supporting 500 users per VM for four VMs including two database VMs, one email VM, and one collaboration application VM.
If the number of users you need to support exceeds the user count the base configuration provides, adding another Dell PowerEdge R720 server to make a 4-2-1 configuration can support up to 1,000 users on multiple business applications while maintaining high availability. Our tests demonstrate this solution supporting 1,000 users per VM in the same four VMs as the 3-2-1 solution. Additionally, these configurations are so easy to scale upward that you can complete the process of adding a server in a matter of two hours or less from start to finish.
If you require support for an even higher number of users, adding a server and storage array to the base configuration, to make a 4-2-2 configuration allows for up to 1,500 users on multiple business applications. Our tests demonstrated this when we ran 1,500 users against each of the four VMs. Better yet, adding a Dell EqualLogic PS6110XV storage array to an existing 4-2-1 configuration is a cinch, taking under 30 minutes start to finish.
All of these Dell Reference Configurations give your business the reliability of highly available hardware configurations, greatly reducing downtime resulting from any hardware malfunctions.
By choosing proven Dell architecture, you avoid the hassles of putting your infrastructure together piece by piece, reducing the potential for error and providing you with a sturdy solution that is easily scalable to fit your present and future needs.
Exchange Server 2013 Preview brings several new features and improvements, including:
1. Support for a multigenerational workforce through enhanced search capabilities and easier contact merging.
2. A refreshed user interface for Outlook 2013 Preview and Outlook Web App.
3. Greater integration with Microsoft SharePoint 2013 Preview and Lync 2013 Preview through new site mailboxes and improved eDiscovery capabilities.
Exchange 2013 introduces a new server role architecture with two main building blocks - the Database Availability Group (DAG) and the Client Access server role. The DAG allows for multiple Mailbox servers to host copies of mailbox databases and provide failover capabilities. The Client Access role is a load balanced front end that routes clients to the appropriate Mailbox server based on the active database copy. This new architecture aims to simplify deployment and administration while improving hardware efficiency and cross-version interoperability compared to previous versions of Exchange.
A primer that I put together for my Network Engineering team to help them understand Exchange. This goes into detail on how MAPI, and other exchange stack protocols work across the network.
Exchange Server 2013 introduced new server roles that simplified the architecture. The Client Access server role became a thin, stateless proxy server that routes client protocol requests. The Mailbox server role consolidated functionality into a single server that processes, renders, and stores Exchange data. This evolution improved hardware efficiency, simplified deployments, and enabled cross-version interoperability. The Front End Transport service acts as a load-balanced ingress/egress point that determines the optimal Mailbox server for delivering messages to avoid unnecessary hops.
The document provides an overview of new features and roles in Windows Server 2008 and 2008 R2 compared to Windows Server 2003. Some key points:
- Windows Server 2008 introduced virtualization capabilities like Hyper-V and new roles like Read-Only Domain Controllers (RODC) to improve security.
- Server Core installation provides a minimal environment optimized for specific server roles like DNS or file services to reduce maintenance and attacks.
- New features in 2008 R2 include Active Directory Recycle Bin, PowerShell 2.0, and administrative tools.
- RODCs allow domain controllers to be deployed in less secure locations read-only, improving authentication speeds and access to resources from branches.
benefits of SQL Server 2008 R2 Enterprise EditionTobias Koprowski
This document contains information about a SQL Server 2008 R2 launch event, including details about the speaker. It provides the speaker's biography, listing their 12 years of experience in IT, focus areas including high availability and security, and certifications. It also lists the speaker's involvement in Microsoft programs, user groups, publishing, and technical support roles.
Upgrading key components in servers can significantly improve performance and capacity. The report tested upgrading components in Dell PowerEdge servers. It found that upgrading to newer Dell PowerEdge R720 servers with Intel Xeon and SSD components, Windows Server 2012, and 10GbE networking supported 4.5 times as many VMs as the older Dell PowerEdge R710 servers. Each individual component upgrade, such as the processor, storage, or network card, increased performance. But upgrading all components together maximized the number of supported VMs, database transactions, and email users.
Dell 3-2-1 Reference Configurations: High available and scalable performance ...Principled Technologies
Dell 3-2-1 Reference Configurations consisting of Dell PowerEdge R720 servers, Dell Force10 S4810P switches, and Dell EqualLogic PS6110XV storage provide a range of configuration options so that you can select the one that is right for your business needs. Additionally, each configuration is easily scalable to accommodate for business growth.
The 3-2-1 base configuration provides a robust solution for any business needing to support up to 500 users on multiple business applications. Our tests demonstrate this solution supporting 500 users per VM for four VMs including two database VMs, one email VM, and one collaboration application VM.
If the number of users you need to support exceeds the user count the base configuration provides, adding another Dell PowerEdge R720 server to make a 4-2-1 configuration can support up to 1,000 users on multiple business applications while maintaining high availability. Our tests demonstrate this solution supporting 1,000 users per VM in the same four VMs as the 3-2-1 solution. Additionally, these configurations are so easy to scale upward that you can complete the process of adding a server in a matter of two hours or less from start to finish.
If you require support for an even higher number of users, adding a server and storage array to the base configuration, to make a 4-2-2 configuration allows for up to 1,500 users on multiple business applications. Our tests demonstrated this when we ran 1,500 users against each of the four VMs. Better yet, adding a Dell EqualLogic PS6110XV storage array to an existing 4-2-1 configuration is a cinch, taking under 30 minutes start to finish.
All of these Dell Reference Configurations give your business the reliability of highly available hardware configurations, greatly reducing downtime resulting from any hardware malfunctions.
By choosing proven Dell architecture, you avoid the hassles of putting your infrastructure together piece by piece, reducing the potential for error and providing you with a sturdy solution that is easily scalable to fit your present and future needs.
Exchange Server 2013 Preview brings several new features and improvements, including:
1. Support for a multigenerational workforce through enhanced search capabilities and easier contact merging.
2. A refreshed user interface for Outlook 2013 Preview and Outlook Web App.
3. Greater integration with Microsoft SharePoint 2013 Preview and Lync 2013 Preview through new site mailboxes and improved eDiscovery capabilities.
Exchange 2013 introduces a new server role architecture with two main building blocks - the Database Availability Group (DAG) and the Client Access server role. The DAG allows for multiple Mailbox servers to host copies of mailbox databases and provide failover capabilities. The Client Access role is a load balanced front end that routes clients to the appropriate Mailbox server based on the active database copy. This new architecture aims to simplify deployment and administration while improving hardware efficiency and cross-version interoperability compared to previous versions of Exchange.
A primer that I put together for my Network Engineering team to help them understand Exchange. This goes into detail on how MAPI, and other exchange stack protocols work across the network.
Exchange Server 2013 introduced new server roles that simplified the architecture. The Client Access server role became a thin, stateless proxy server that routes client protocol requests. The Mailbox server role consolidated functionality into a single server that processes, renders, and stores Exchange data. This evolution improved hardware efficiency, simplified deployments, and enabled cross-version interoperability. The Front End Transport service acts as a load-balanced ingress/egress point that determines the optimal Mailbox server for delivering messages to avoid unnecessary hops.
The document provides an overview of new features and roles in Windows Server 2008 and 2008 R2 compared to Windows Server 2003. Some key points:
- Windows Server 2008 introduced virtualization capabilities like Hyper-V and new roles like Read-Only Domain Controllers (RODC) to improve security.
- Server Core installation provides a minimal environment optimized for specific server roles like DNS or file services to reduce maintenance and attacks.
- New features in 2008 R2 include Active Directory Recycle Bin, PowerShell 2.0, and administrative tools.
- RODCs allow domain controllers to be deployed in less secure locations read-only, improving authentication speeds and access to resources from branches.
benefits of SQL Server 2008 R2 Enterprise EditionTobias Koprowski
This document contains information about a SQL Server 2008 R2 launch event, including details about the speaker. It provides the speaker's biography, listing their 12 years of experience in IT, focus areas including high availability and security, and certifications. It also lists the speaker's involvement in Microsoft programs, user groups, publishing, and technical support roles.
IRJET- Load Balancing Cluster based on Linux Virtual ServerIRJET Journal
1) The document discusses load balancing techniques for distributed server clusters using Linux Virtual Server (LVS). LVS allows building a virtual server from a cluster of real servers to improve scalability and availability.
2) It presents an LVS cluster architecture with multiple front-ends that can also act as back-ends to remove bottlenecks and balance loads more effectively as requests increase.
3) The proposed scheduling algorithm distributes requests equally among servers considering their loads, allowing servers to redirect requests to less busy backends to avoid overloading.
Dell 3-2-1 Reference Configurations: Scalable performance and simplicity in s...Principled Technologies
Dell 3-2-1 Reference Configurations provide a range of virtualized infrastructure solutions to meet your business’s needs today and in the future. Easy to deploy, manage, and upgrade, these robust solutions can grow as your business does, all the while reducing the likelihood of extended downtime due to their highly available architecture.
The Microsoft MSCOM Ops team tested SQL Server 2008 and Windows Server 2008's improvements to wide area network (WAN) replication performance. They found push and pull replication between data centers 3,000 miles apart was up to 100 times faster with the new versions compared to SQL Server 2005 and Windows Server 2003. Based on this, they replicated "live" content from their MSDN database in Washington to Dublin, Ireland, reducing European page load times by 33%. The team plans to expand use of WAN replication after demonstrating its feasibility for a real application.
This document discusses performance tuning and optimization in client/server computing. It covers tuning performance at the server, client, database, and network levels. Some key points include:
- Upgrading server hardware, using multiple network cards, and high-performance file systems can improve server performance. Offloading processing to specialized servers also helps.
- Client performance is impacted by hardware, operating systems that support multitasking, and efficient applications.
- Database performance relies on normalized database design, efficient indexing, and optimized query design using WHERE clauses to limit returned records.
- Performance tuning follows steps like identifying bottlenecks, modifying systems, and remeasuring impact while performance optimization employs techniques to improve network usage and data retrieval speeds.
This document provides instructions for configuring Lotus Notes to journal email to Cryoserver instead of a local Notes server for archiving purposes. It describes enabling basic journaling in Notes, setting an SMTP route to Cryoserver, and configuring LDAP integration to allow Cryoserver to validate logins and expand distribution lists using information from the Notes LDAP directory. Some known issues are also listed, such as the inability to view BCC recipients when journaling to Cryoserver instead of a local Notes server.
This document outlines the evolution of Microsoft Exchange server from version 4.0 to 2013. It discusses key features and changes introduced in each new version, such as support for Outlook, integration with Active Directory, improved web access, mobile device support, continuous replication for high availability, role-based access control and new server roles. Each version built upon the previous one to provide enhanced email, calendaring and collaboration capabilities.
The document describes reports from Tideway Foundation used in virtualization and consolidation projects. It allows identification of systems by hardware, OS, power usage. Dependency mapping prioritizes systems by business impact. Transition reports link physical and virtual systems and compare builds. Lifecycle management tracks deployments and alerts on changes.
Exchange 2013 is the latest version of Microsoft Exchange server software. Key new features include web-based administration, managed mailboxes that use separate worker processes, data loss prevention capabilities, modernized public folders, support for multiple databases per disk, and enhanced high availability. Exchange 2013 also reduces server roles and improves features such as lagged copies, offline Outlook Web Access, and anti-malware protection compared to previous versions.
Component upgrades from Intel and Dell can increase VM density and boost perf...Principled Technologies
The document summarizes an experiment conducted by Principled Technologies that tested the performance improvements from upgrading server components. They found that upgrading from a Dell PowerEdge R720 to a Dell PowerEdge R730 server, along with upgrading the processor, operating system, storage drives and network cards, increased the number of supported VMs by 67% and database performance by 60%. Upgrading all components maximized performance benefits.
This document discusses IBM SmartCloud Notes mail routing. It provides an overview of IBM SmartCloud Notes, including email hygiene, email retention, and configurations for hybrid and hosted environments. It then describes the routing topology and server roles. The rest of the document discusses various mail flow scenarios for hosted-only, hybrid, and typical hybrid environments. It also addresses some common issues, cases, and best practices. Useful URLs are provided at the end.
This document provides an overview of monitoring and maintaining Microsoft Exchange Server 2010. It discusses the importance of performance monitoring and tools for monitoring Exchange servers. It also outlines processes for deploying software and hardware updates, developing a troubleshooting methodology, and troubleshooting issues with mailbox servers and client access servers. The document concludes with a lab scenario for hands-on exercises in monitoring Exchange and troubleshooting database availability and client access server issues.
This white paper presents a solution to test performance and analyze the results for web services that are deployed on the webMethods Integration Server using Apache JMeter.
Technical Report_Sercer 2012 R2 - Adeeb RajaAdeeb Raja
This technical training report provides guidance on installing and configuring Windows Server 2012 R2. It discusses the key features of Windows Server 2012 R2 including the installation options, user interface, Task Manager, IP Address Management, Active Directory, Hyper-V, ReFS, and IIS 8.0. It then covers installing Windows Server 2012 R2, configuring the local server, adding roles and features, managing Active Directory users and computers, and configuring Group Policy management. The report also provides information on using Hyper-V including virtual hard disks, storage quality of service, live migrations, and creating virtual machines. Finally, it discusses domain joining Windows 7 Ultimate to an Active Directory domain.
Array Networks & Microsoft Exchange Server 2010 Array Networks
Array Networks Enables Highly Optimized Microsoft
Exchange Server 2010 Services.
Array Networks has worked closely with
Microsoft to create a joint solution which
accelerates, secures and optimizes the
delivery of Exchange services. This joint
solution enables organizations to gain
the greatest value from their Exchange
investment and to deliver business critical
Exchange services. The Array solution also
addresses the problem of securely accessing
Exchange email from anywhere on any
device.
Attention Session en Anglais, animée par Scott Schnoll, Senior Content Developer à Microsoft Corp. Cette session vous expliquera comment architecturer une migration vers Exchange 2013 depuis d'ancienne versions d'Exchange avec comme speaker Scott Schnoll, Gourou Exchange en provenance directe de Redmond. La coexistence de serveurs de versions différentes sera aussi abordée.
Speaker : Scott Schnoll (Microsoft)
This document discusses securing Microsoft Exchange Server 2010. It covers configuring role-based access control (RBAC) to define permissions for administrators and users. It also discusses securing server roles and internet access. Demonstrations show how to configure custom RBAC role groups, the Threat Management Gateway for Outlook Web App, and a lab exercise for securing an Exchange deployment.
The document summarizes a health check for Microsoft SQL Server that assesses efficiency and effectiveness. The check evaluates how fully the SQL Server products have been utilized and considers issues like hardware resources, database tuning, and staff skills. It addresses these issues over 1 to 5 days to provide recommendations around performance, stability, and availability.
The document discusses Microsoft's roadmap for Windows Server releases, including plans for Windows Server 2008 R2. It notes that R2 releases enable refinement of features and adaptation to changing technologies between major releases. Windows Server 2008 R2 will focus on virtualization, management improvements, integration with Windows 7, scalability, reliability, and web functionality. The release is intended to build on Windows Server 2008 and deliver targeted innovations on a predictable two-year cycle.
This presentation was prepared as part of a group consultation assignment in the psychiatric mental-health nurse practitioner program at University of Tennessee Health Sciences Center.
This document summarizes the economic impacts of the recession on the African American community. It finds that African American homeownership, income, employment, and wealth were disproportionately impacted. Unemployment rates for African Americans remain twice as high as whites. While all groups saw declines in income and wealth during the recession, the drops were largest for people of color who rely more on housing for wealth. The summary concludes that significant recovery in the African American housing market will require non-traditional approaches to address issues like blight, foreclosures, and access to credit.
This document provides an overview of foundational theories and frameworks in the instructional technology field. It discusses key concepts like learning and instruction, and summarizes several important psychological foundation theories including constructivism, cognitive information processing theory, Gagne's theory of instruction, and situated learning theory. It also covers motivational theories and models such as achievement motivation, the time-continuum model, and the ARCS model. Finally, it discusses different learning environments including integrated learning systems, distance learning, advance organizers, and learning communities.
Case Transfluent, Julien Fourgeaud, VP Mobile (in English)TouNethanke
This document discusses rapid language translation services for social media, web, and apps provided by Transfluent. Key points include:
- Transfluent provides near real-time translations of social media posts and comments to reach global audiences.
- Case study of music band Thirty Seconds to Mars seeing success translating to Chinese and 10 other languages on Facebook.
- Transfluent's machine translation is supplemented by human translators and has affordable per-word pricing.
- The company is seeing growing traction and translations since opening a US office in January 2013.
IRJET- Load Balancing Cluster based on Linux Virtual ServerIRJET Journal
1) The document discusses load balancing techniques for distributed server clusters using Linux Virtual Server (LVS). LVS allows building a virtual server from a cluster of real servers to improve scalability and availability.
2) It presents an LVS cluster architecture with multiple front-ends that can also act as back-ends to remove bottlenecks and balance loads more effectively as requests increase.
3) The proposed scheduling algorithm distributes requests equally among servers considering their loads, allowing servers to redirect requests to less busy backends to avoid overloading.
Dell 3-2-1 Reference Configurations: Scalable performance and simplicity in s...Principled Technologies
Dell 3-2-1 Reference Configurations provide a range of virtualized infrastructure solutions to meet your business’s needs today and in the future. Easy to deploy, manage, and upgrade, these robust solutions can grow as your business does, all the while reducing the likelihood of extended downtime due to their highly available architecture.
The Microsoft MSCOM Ops team tested SQL Server 2008 and Windows Server 2008's improvements to wide area network (WAN) replication performance. They found push and pull replication between data centers 3,000 miles apart was up to 100 times faster with the new versions compared to SQL Server 2005 and Windows Server 2003. Based on this, they replicated "live" content from their MSDN database in Washington to Dublin, Ireland, reducing European page load times by 33%. The team plans to expand use of WAN replication after demonstrating its feasibility for a real application.
This document discusses performance tuning and optimization in client/server computing. It covers tuning performance at the server, client, database, and network levels. Some key points include:
- Upgrading server hardware, using multiple network cards, and high-performance file systems can improve server performance. Offloading processing to specialized servers also helps.
- Client performance is impacted by hardware, operating systems that support multitasking, and efficient applications.
- Database performance relies on normalized database design, efficient indexing, and optimized query design using WHERE clauses to limit returned records.
- Performance tuning follows steps like identifying bottlenecks, modifying systems, and remeasuring impact while performance optimization employs techniques to improve network usage and data retrieval speeds.
This document provides instructions for configuring Lotus Notes to journal email to Cryoserver instead of a local Notes server for archiving purposes. It describes enabling basic journaling in Notes, setting an SMTP route to Cryoserver, and configuring LDAP integration to allow Cryoserver to validate logins and expand distribution lists using information from the Notes LDAP directory. Some known issues are also listed, such as the inability to view BCC recipients when journaling to Cryoserver instead of a local Notes server.
This document outlines the evolution of Microsoft Exchange server from version 4.0 to 2013. It discusses key features and changes introduced in each new version, such as support for Outlook, integration with Active Directory, improved web access, mobile device support, continuous replication for high availability, role-based access control and new server roles. Each version built upon the previous one to provide enhanced email, calendaring and collaboration capabilities.
The document describes reports from Tideway Foundation used in virtualization and consolidation projects. It allows identification of systems by hardware, OS, power usage. Dependency mapping prioritizes systems by business impact. Transition reports link physical and virtual systems and compare builds. Lifecycle management tracks deployments and alerts on changes.
Exchange 2013 is the latest version of Microsoft Exchange server software. Key new features include web-based administration, managed mailboxes that use separate worker processes, data loss prevention capabilities, modernized public folders, support for multiple databases per disk, and enhanced high availability. Exchange 2013 also reduces server roles and improves features such as lagged copies, offline Outlook Web Access, and anti-malware protection compared to previous versions.
Component upgrades from Intel and Dell can increase VM density and boost perf...Principled Technologies
The document summarizes an experiment conducted by Principled Technologies that tested the performance improvements from upgrading server components. They found that upgrading from a Dell PowerEdge R720 to a Dell PowerEdge R730 server, along with upgrading the processor, operating system, storage drives and network cards, increased the number of supported VMs by 67% and database performance by 60%. Upgrading all components maximized performance benefits.
This document discusses IBM SmartCloud Notes mail routing. It provides an overview of IBM SmartCloud Notes, including email hygiene, email retention, and configurations for hybrid and hosted environments. It then describes the routing topology and server roles. The rest of the document discusses various mail flow scenarios for hosted-only, hybrid, and typical hybrid environments. It also addresses some common issues, cases, and best practices. Useful URLs are provided at the end.
This document provides an overview of monitoring and maintaining Microsoft Exchange Server 2010. It discusses the importance of performance monitoring and tools for monitoring Exchange servers. It also outlines processes for deploying software and hardware updates, developing a troubleshooting methodology, and troubleshooting issues with mailbox servers and client access servers. The document concludes with a lab scenario for hands-on exercises in monitoring Exchange and troubleshooting database availability and client access server issues.
This white paper presents a solution to test performance and analyze the results for web services that are deployed on the webMethods Integration Server using Apache JMeter.
Technical Report_Sercer 2012 R2 - Adeeb RajaAdeeb Raja
This technical training report provides guidance on installing and configuring Windows Server 2012 R2. It discusses the key features of Windows Server 2012 R2 including the installation options, user interface, Task Manager, IP Address Management, Active Directory, Hyper-V, ReFS, and IIS 8.0. It then covers installing Windows Server 2012 R2, configuring the local server, adding roles and features, managing Active Directory users and computers, and configuring Group Policy management. The report also provides information on using Hyper-V including virtual hard disks, storage quality of service, live migrations, and creating virtual machines. Finally, it discusses domain joining Windows 7 Ultimate to an Active Directory domain.
Array Networks & Microsoft Exchange Server 2010 Array Networks
Array Networks Enables Highly Optimized Microsoft
Exchange Server 2010 Services.
Array Networks has worked closely with
Microsoft to create a joint solution which
accelerates, secures and optimizes the
delivery of Exchange services. This joint
solution enables organizations to gain
the greatest value from their Exchange
investment and to deliver business critical
Exchange services. The Array solution also
addresses the problem of securely accessing
Exchange email from anywhere on any
device.
Attention Session en Anglais, animée par Scott Schnoll, Senior Content Developer à Microsoft Corp. Cette session vous expliquera comment architecturer une migration vers Exchange 2013 depuis d'ancienne versions d'Exchange avec comme speaker Scott Schnoll, Gourou Exchange en provenance directe de Redmond. La coexistence de serveurs de versions différentes sera aussi abordée.
Speaker : Scott Schnoll (Microsoft)
This document discusses securing Microsoft Exchange Server 2010. It covers configuring role-based access control (RBAC) to define permissions for administrators and users. It also discusses securing server roles and internet access. Demonstrations show how to configure custom RBAC role groups, the Threat Management Gateway for Outlook Web App, and a lab exercise for securing an Exchange deployment.
The document summarizes a health check for Microsoft SQL Server that assesses efficiency and effectiveness. The check evaluates how fully the SQL Server products have been utilized and considers issues like hardware resources, database tuning, and staff skills. It addresses these issues over 1 to 5 days to provide recommendations around performance, stability, and availability.
The document discusses Microsoft's roadmap for Windows Server releases, including plans for Windows Server 2008 R2. It notes that R2 releases enable refinement of features and adaptation to changing technologies between major releases. Windows Server 2008 R2 will focus on virtualization, management improvements, integration with Windows 7, scalability, reliability, and web functionality. The release is intended to build on Windows Server 2008 and deliver targeted innovations on a predictable two-year cycle.
This presentation was prepared as part of a group consultation assignment in the psychiatric mental-health nurse practitioner program at University of Tennessee Health Sciences Center.
This document summarizes the economic impacts of the recession on the African American community. It finds that African American homeownership, income, employment, and wealth were disproportionately impacted. Unemployment rates for African Americans remain twice as high as whites. While all groups saw declines in income and wealth during the recession, the drops were largest for people of color who rely more on housing for wealth. The summary concludes that significant recovery in the African American housing market will require non-traditional approaches to address issues like blight, foreclosures, and access to credit.
This document provides an overview of foundational theories and frameworks in the instructional technology field. It discusses key concepts like learning and instruction, and summarizes several important psychological foundation theories including constructivism, cognitive information processing theory, Gagne's theory of instruction, and situated learning theory. It also covers motivational theories and models such as achievement motivation, the time-continuum model, and the ARCS model. Finally, it discusses different learning environments including integrated learning systems, distance learning, advance organizers, and learning communities.
Case Transfluent, Julien Fourgeaud, VP Mobile (in English)TouNethanke
This document discusses rapid language translation services for social media, web, and apps provided by Transfluent. Key points include:
- Transfluent provides near real-time translations of social media posts and comments to reach global audiences.
- Case study of music band Thirty Seconds to Mars seeing success translating to Chinese and 10 other languages on Facebook.
- Transfluent's machine translation is supplemented by human translators and has affordable per-word pricing.
- The company is seeing growing traction and translations since opening a US office in January 2013.
This document discusses numerical discretisation methods for solving transport equations using finite volume methods on polyhedral meshes. It begins by defining discretisation as representing differential equations with algebraic expressions, usually in matrix form. It then discusses representing fields on computational meshes, integrating operators over cells, and representing spatial and temporal variation to create second-order accurate discretisations. The document focuses on a methodology for discretising operators like divergence, gradient, and time derivatives to build the discretised transport equation on polyhedral meshes in a way that generalizes to different cell types.
This white paper summarizes the results of benchmark testing of Microsoft BizTalk Server 2006 and SQL Server 2005 running on Unisys ES7000/one Enterprise Servers. The testing achieved unprecedented throughput levels, with the Latency Application scenario reaching 1,156 orchestrations per second. This level of performance far surpassed any previously recorded by Microsoft for BizTalk Server. The results demonstrate the scalability of BizTalk Server on the ES7000 platform and that further optimizations could allow even higher performance.
Whats new in Enterprise 5.0 Product SuiteMicro Focus
This document summarizes new features across Micro Focus's Enterprise Product Suite version 5.0, including .NET Core support, Amazon Web Services Quick Start, COBOL formatting, code analysis views, Enterprise Server scale out architecture, common web administration, Application Workflow Manager improvements, AppMaster Builder data view changes, CICS and IMS support enhancements, COBOL and PL/I language additions, debugging upgrades, and more. Key areas of focus include multi-system administration of Enterprise Server, integration of mainframe workloads on modern platforms, and development productivity aids.
The document discusses load testing software called Traffic Simulator. It allows tracking traffic on a production system without loading it, reproducing the traffic on another version of SQL Server or hardware. This allows comparing performance between configurations. Traffic Simulator records traffic, analyzes it, and replays it on a test system. It also provides reports to identify differences in query speeds, errors or data between the original and test systems. This helps evaluate the impact of changes like upgrading SQL Server versions or hardware.
DBA, LEVEL III TTLM Monitoring and Administering Database.docxseifusisay06
The document provides information about monitoring, administering, and tuning a SQL Server database, including:
1) Steps for installing and configuring SQL Server.
2) The importance of database monitoring to track performance and ensure availability.
3) Tools that can be used for database monitoring and performance tuning.
4) Activities involved in database maintenance and the different editions of SQL Server 2008.
5) Methods for installing SQL Server, including local, unattended, and remote installations.
This document provides an overview of Microsoft's Windows clustering technologies, including Network Load Balancing (NLB), Component Load Balancing (CLB), and Server Cluster. It discusses how these technologies can be used together to provide high availability, reliability, and scalability for applications and services. NLB is used for load balancing web servers, CLB balances application components, and Server Cluster provides failover for databases and backend services. Clusters can be organized as farms of identical servers or packs that share partitioned data. Technologies can scale out by adding servers or scale up individual servers. Server Cluster nodes can operate actively or passively.
This document discusses the challenges cloud providers face in managing the performance of enterprise applications deployed in the cloud. It outlines how queuing models can be used to analyze application performance, identify bottlenecks, determine optimal resource allocation, and ensure performance meets SLAs. The key points are:
1) Cloud providers must monitor application workloads, characterize transactions and usage patterns, and plan capacity based on changing demands.
2) Queuing models can simulate application behavior under different workloads and help size resources needed to meet performance targets.
3) Both hardware and software bottlenecks must be identified and addressed, as insufficient tuning parameters can impact performance more than hardware capacity.
This document describes MedicaLogic's process for performing scalability testing on their Electronic Medical Records software over three generations to support an increasing number of users. The first generation used 24 PCs and 3 servers to simulate up to 60 users. Problems included difficulty automating the software. The second generation used Compuware's QALoad to simulate up to 400 users by capturing SQL statements. The third generation added more powerful computers and a UNIX server to simulate over 1,000 users, with ongoing work to improve the testing tools and environment.
Odca interop across_clouds_standard units of measurement for iaa_sSeanscs
This document proposes standard units of measure (SUoM) for cloud computing services. It focuses on infrastructure as a service (IaaS). Quantitative units are proposed for compute (using SPECvirt benchmarks), storage (TB and IOPS), and network (Gbps and latency/jitter). Qualitative measures differentiate service levels as bronze, silver, gold, and platinum. The SUoM aims to enable easier comparison of cloud services between providers and with internal capabilities.
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology.
This document discusses techniques for optimizing the performance of PeopleSoft applications. It covers tuning several aspects within a PeopleSoft environment, including server performance, web server performance, Tuxedo performance management, application performance, and database performance. Some key recommendations include implementing a methodology to monitor resource consumption without utilizing critical resources, ensuring load balancing strategies are sound, measuring historical patterns of server resource utilization, capturing key performance metrics for Tuxedo, and focusing on tuning high-resource consuming SQL statements and indexes.
Scale and Load Testing of Micro-ServiceIRJET Journal
1) The document discusses testing of microservices using load and scale testing frameworks. It describes using Locust for load testing and analyzing response times as the number of requests per second and pods are increased.
2) APIs were tested using Pytest with over 50 test cases covering different responses and statuses. Test results were generated as Allure reports.
3) New Relic was used to monitor application performance under increasing load and identify bottlenecks. Testing showed the system could handle 2000 requests per second per container before response times increased significantly.
CS 301 Computer ArchitectureStudent # 1 EID 09Kingdom of .docxfaithxdunce63732
This document summarizes the results of simulations run to analyze the performance of different processor configurations with varying levels of instruction-level parallelism. The key findings are:
1) For processors with significant memory latency, there is little performance difference between simple in-order and more complex out-of-order designs, as memory latency dominates execution time.
2) Supporting just two concurrently pending instructions provides most of the benefit of more complex out-of-order execution, while greatly reducing hardware complexity.
3) As the mismatch between processor and memory system performance increases, all designs see similar performance, regardless of the level of instruction-level parallelism exploited.
The technical level of this document is 300.
This article requires knowledge about Microsoft Performance Monitor, Lync Monitoring, Hardware SPEC’s and Consulting.
Building the Test User Scenario, requires full understand of each Lync feature and how user will utilize those feature. Please also carefully understand the Lync PSTN Gateway Simulator. The Simulator must be provide with correct Dial-Plans, Voice Policy and Voice Routes.
You also need to understand the Lync User Provisioning Tool and how to design the Stress Test Scenarios.
Note:
This document is neither a sizing nor a configuration guide. You should use this document only for your environment planning’s purposes and security considerations. In lager environments you should spend some time to evaluate the optimal path of your Lync deployment.
Table of Content:
Introduction 4
The Calculation Process (I can recommend to you): 6
Build the Stress Test Lab 7
The Stress Test Process: 9
The Validation Process: 11
User Provisioning Tool 12
Setup User Provisioning Tool 13
User Creation 13
Contacts Creation 14
Distribution List Creation 15
Location Info Service Config 16
Run Configuration Scripts 18
Stress Test Simulations 20
User Profile Generator 20
Common Configuration 21
General Scenario 23
Voice Scenario 34
Reach Scenario 49
Mobility Scenario 52
Summary (Important User Load definition) 54
Table of Figures 58
Author: Thomas Poett MVP, Managing Consultant Microsoft Unified Communication
Extended ECM - Guide to Cloud Services - Whitepaper.pdfTarek Ashmar
The document provides an overview of OpenText's Extended ECM Cloud Edition services. It describes the key components, metrics, service level policies around availability, recovery objectives, and regions. It also outlines the deployment options, onboarding process, system updates, configuration capabilities, and request fulfillment. Finally, it covers data policies around lifecycle, encryption, authentication, and certifications around security and compliance.
Microsoft SQL High Availability and ScalingJustin Whyte
This is a white paper that I wrote that explores the different options for Microsoft SQL Server High Availability. I explain the pros and cons of each technology, as well as important considerations like RTO and RPO.
The document provides sizing guidelines for deploying Oracle Identity Manager 11gR2PS1 based on testing. It discusses factors that impact hardware sizing like the number of users, provisioning operations, reconciliation data volume and frequency, and certification campaigns. The guidelines help estimate the computing resources and database size needed to support expected usage.
Configure an Integrated Exchange, Lync, and SharePoint Test LabVinh Nguyen
This document provides step-by-step instructions to configure an integrated test lab containing Exchange Server 2013, Lync Server 2013, and SharePoint Server 2013. The test lab consists of 5 servers - one domain controller, one SQL server, one SharePoint server, one Exchange server, and one Lync server. It also includes 2 client computers. The document outlines 9 steps to install and configure the servers and clients, and then integrate the Exchange, Lync and SharePoint servers by configuring server-to-server trust relationships.
Realize better value and performance migrating from Azure Database for Postgr...Principled Technologies
It’s time to move your databases from Azure Database for PostgreSQL – Single Server instances now, before March 2025 when Microsoft Azure retires the service. Our hands-on testing shows that moving to Azure Database for PostgreSQL – Flexible Server is a simple process that can offer significantly better database performance and value. With as much as 4.71 times the NOPM on HammerDB and up to 3.88 times the performance per dollar compared to Single Server, moving to Azure Database for PostgreSQL – Flexible Sever with AMD EPYC processors can give you higher performing databases for lower relative cost.
IRJET - Efficient Load Balancing in a Distributed EnvironmentIRJET Journal
This document discusses load balancing algorithms for distributed computing environments. It begins by defining load balancing and describing its importance in distributed systems for optimizing resource utilization and system performance. Several static and dynamic load balancing algorithms are then summarized, including round robin, random, min-min, and max-min algorithms. The document also outlines key issues in load balancing, advantages, metrics for evaluating algorithms, and provides more detailed descriptions of 13 load balancing algorithms.
Similar to Performance report for Microsoft Office Communications Server ... (20)
This paper proposes an adaptive energy management policy for wireless video streaming between a battery-powered client and server. It models the energy consumption of the server and client based on factors like CPU frequency, transmission power, and channel bandwidth. The paper formulates an optimization problem to assign optimal energy to each video frame. This maximizes system lifetime while meeting a minimum video quality requirement. Experimental results show the proposed policy increases overall system lifetime by 20% on average.
Microsoft PowerPoint - WirelessCluster_PresVideoguy
This document analyzes delays in unicast video streaming over IEEE 802.11 WLAN networks. It describes conducting an experiment using a testbed with a Darwin Streaming Server and WLAN probe to capture packets. The analysis found that video bitrate variations, packetization scheme, bandwidth load, and frame-based nature of video all impacted mean delay. Bursts of packets from video frames caused per-packet delay to increase in a sawtooth pattern. Increasing uplink load was also found to affect delay variations.
Proxy Cache Management for Fine-Grained Scalable Video StreamingVideoguy
This document proposes a novel video caching framework that uses MPEG-4 Fine-Grained Scalable (FGS) video with post-encoding rate control to achieve low-cost and fine-grained rate adaptation. The framework allows clients to have heterogeneous bandwidths and enables adaptive control of backbone bandwidth consumption. It examines issues in caching FGS videos, such as determining the optimal portion to cache (in terms of length and rate) and optimal streaming rate to clients. Simulation results show it significantly reduces transmission costs compared to non-adaptive caching while providing flexible utility to heterogeneous clients with low computational overhead.
The document compares Microsoft Windows Media and the Adobe Flash Platform for streaming media. It discusses key differences like user experience, workflows, and playback reach. Flash offers more flexibility in creative expression, richer interactions, and wider device playback than Windows Media. It also has a 98% install base, making it easier for viewers to watch streams without extra software. The document outlines workflows for experience design, programming, broadcasting, production, and more using Flash tools versus Microsoft alternatives.
Free-riding Resilient Video Streaming in Peer-to-Peer NetworksVideoguy
This document summarizes a PhD thesis about free-riding resilient video streaming in peer-to-peer networks. The thesis contains research on two approaches: tree-based live streaming and swarm-based video-on-demand. For tree-based live streaming, the thesis presents the Orchard algorithm for constructing and maintaining trees to distribute video in a peer-to-peer network. It analyzes attacks on Orchard like free-riding and evaluates Orchard's performance under different conditions through experiments. For swarm-based video-on-demand, the thesis introduces the Give-to-Get approach for distributing video files and compares it to other peer-to-peer protocols. It evaluates Give-to-Get's performance in experiments
BT has developed Fastnets technology to improve video streaming. It avoids start-up delays and picture freezing during congestion. Fastnets streams multiple encoded versions of the video at different data rates and seamlessly switches between them based on available bandwidth to maintain quality without pausing. This allows for near-instant start times and reduces bandwidth usage by up to 30%. Fastnets provides a high-quality video streaming solution for both mobile and IPTV applications.
This document summarizes recent research on video streaming over Bluetooth networks. It discusses three key areas: intermediate protocols, quality of service (QoS) control, and media compression. For intermediate protocols, it evaluates streaming via HCI, L2CAP, and IP layers and their tradeoffs. For QoS control, it describes how error control mechanisms like link layer FEC, retransmission, and error concealment can improve video quality over Bluetooth. It also discusses congestion control. For media compression, it notes the importance of compression to achieve efficiency over limited Bluetooth bandwidths.
The document discusses video streaming, including definitions and concepts. It covers topics such as the difference between streaming and downloading, common streaming categories like live and on-demand, protocols used for streaming like RTSP and RTP, and the development process for creating streaming video including content planning, capturing, editing, encoding, and integrating with servers.
Inlet Technologies offers a live video streaming solution called Spinnaker that uses Intel Xeon processors with quad-core technology. Spinnaker can encode live video streams into multiple formats and resolutions simultaneously. This allows content to be delivered optimally to various devices. Spinnaker is a flexible, scalable solution that can increase broadcast capacity cost-effectively while maintaining high video quality.
Considerations for Creating Streamed Video Content over 3G ...Videoguy
The document discusses considerations for creating video content that can be streamed over mobile networks with restricted bandwidth like 3G-324M. It covers topics like video basics, codecs, profiles and levels, video streaming techniques, guidelines for authoring mobile-friendly content, and tools for analyzing video streams. The goal is to help content creators optimize video quality for low-bandwidth mobile viewing.
ADVANCES IN CHANNEL-ADAPTIVE VIDEO STREAMINGVideoguy
This document summarizes recent advances in channel-adaptive video streaming. It reviews adaptive media playout at the client to reduce latency, rate-distortion optimized packet scheduling to determine the best packet to send, and channel-adaptive packet dependency control to improve error robustness and reduce latency. It also discusses challenges for wireless video streaming and different wireless streaming architectures.
Impact of FEC Overhead on Scalable Video StreamingVideoguy
The document discusses the impact of forward error correction (FEC) overhead on scalable video streaming. It aims to address uncertainty about the benefits of FEC and provide insight into how FEC overhead affects scalable video performance. The motivation section explains that FEC is often used for streaming to overcome packet loss without retransmission. However, previous studies have reported conflicting results on the benefits of FEC. The background section provides details on media-independent FEC schemes.
The document proposes a cost-effective solution for video streaming and rich media applications using Vela's RapidAccess video server combined with iQstor's iQ1200 SATA storage system. The integrated encoding, decoding and video serving capabilities of RapidAccess are paired with the scalable storage and virtualization features of the iQ1200 SATA storage array to provide a robust yet affordable infrastructure for applications such as video on demand, corporate training and distance learning.
This document provides information on streaming video into Second Life, including:
- The basic prerequisites for streaming video include being the landowner, using QuickTime format videos, and having the video hosted on a web server.
- There are three main ways to stream video: establishing movie playback, streaming live video, and broadcasting from Second Life.
- Streaming live video or broadcasting involves using software like QuickTime Broadcaster or Windows Media Encoder to capture the video stream and send it to a hosting server, then entering that URL in Second Life.
XStream Live 2 is a live video encoding and streaming software that allows users to broadcast high quality HD video at low bitrates. It supports various video formats and streaming servers. The software provides high quality H.264 encoding with proprietary technology. It is designed for live event streaming, IPTV, and other video distribution uses.
The document provides instructions for setting up a homemade videoconference streaming solution using Windows Media software. The solution involves installing Windows Media Encoder and Administrator on a server and configuring the software to receive a video stream from a videoconferencing terminal. The streaming server then broadcasts the stream in real-time to clients who can view it using media player software. The solution provides a low-cost way to stream videoconferences but has limitations such as only supporting one conference stream at a time.
This document describes iStream Live 2 software for live streaming video to iPhones and iPads. It allows streaming of SD or HD video over HTTP from a variety of video sources. Key features include support for all major CDNs, encoding of H.264 video and AAC audio for high quality at low bitrates, and integration with existing Windows streaming systems. It provides better quality streaming than other encoders at lower bandwidth requirements.
Glow: Video streaming training guide - FirefoxVideoguy
This document provides a guide to using Glow video streaming. It includes tutorials on setting up video streaming by adding the Video Streaming Management web part, uploading video clips, viewing clips, editing clip information, and deleting clips. The guide also discusses how video streaming can be used to support learning and teaching, such as adding videos to lessons.
Performance report for Microsoft Office Communications Server ...
1. Performance report for Microsoft Office Communications
Server Consolidated Configuration on HP BladeSystem
Introduction......................................................................................................................................... 2
Test methodology................................................................................................................................. 4
Test topology....................................................................................................................................... 5
SQL Server storage subsystem ........................................................................................................... 7
Test results .......................................................................................................................................... 8
Conclusions ...................................................................................................................................... 13
For more information.......................................................................................................................... 14
2. Introduction
Microsoft® Office Communications Server (OCS) 2007 is an extensive update to Live
Communications Server (LCS) 2005. In addition to the instant messaging and presence features of
LCS, OCS provides Enterprise Voice with integrated voice over IP (VoIP) and telephony integration,
audio/video conferencing, on premise web conferencing and federation enhancements.
OCS servers can host a core set of services including: Instant Messaging (IM) Conferencing Server,
Web Conferencing Server, Telephony Conferencing Server, Audio/Video Conferencing server,
Internet Information Services (IIS), and the Focus, which helps manage meeting participants. Database
services, provided by Microsoft SQL Server, are also required.
There is also a set of OCS server roles in addition to those core services to provide incremental
functionality such as access for users over the Internet, archiving and call detail reporting, telephony
integration, public IM connectivity, and access through a browser based client. These server roles are
typically deployed on servers separate from the core services described above.
OCS can be deployed in three primary configurations. The base configuration is OCS Standard
Edition (SE) which includes all core services and SQL Server 2005 Express Edition on a single server.
With all functionality on a single server, Standard Edition provides no fault tolerance and is not
recommended for deployments over 5000 users. Figure 1 depicts a Standard Edition deployment.
Figure 1. Microsoft Office Communications Server Standard Edition
The second method of deployment is OCS Enterprise Edition (EE) in a Consolidated Configuration.
In this configuration a pool of servers is created where all core services are on each of the servers
and a hardware load balancer distributes traffic among the members of the pool. Microsoft SQL
Server 2005 is on a separate server for back end database services. Figure 2 depicts an Enterprise
Edition Consolidated Configuration. Some advantages of a Consolidated Configuration are that the
SQL Server can be scaled for performance independently of other roles and can also be clustered
for increased availability and maintenance. Another benefit is that the number of OCS servers can
be scaled to meet each customer’s needs. Multiple OCS servers behind a hardware load balancer
provide redundancy for maintenance and high availability. The Consolidated Configuration is the
simpler of the two Enterprise Edition configurations and is generally recommended for up to
30,000 users.
2
3. Figure 2. Microsoft Office Communications Server Enterprise Edition Consolidated Configuration
The third method of deployment is OCS Enterprise Edition in an Expanded Configuration. This
configuration breaks out IIS, web conferencing and audio/video conferencing on to separate pools of
servers. In this configuration those functions can be scaled separately to meet specific needs. Figure 3
depicts an Enterprise Edition Expanded Configuration. As with a Consolidated Configuration, the SQL
Server is separate and can be scaled and clustered to address those needs. By breaking out the
server functionality in to separate server pools, each type of functionality can be scaled independently
as needed. The Enterprise Edition Expanded Configuration is recommended for up to 125,000 users.
Figure 3. Microsoft Office Communications Enterprise Edition Expanded Configuration
Figures 1, 2 and 3 are from “Microsoft Office Communications Server 2007 Technical Overview”.
For more information about these topologies please visit the Microsoft Office Communications Server
Technical Library located here: http://technet.microsoft.com/en-us/library/bb676082.aspx
Each of the topologies shown above do not include the additional servers to provide services outside
of the core services such as edge servers for Internet access, mediation servers for telephony
integration, archiving and call detail record services, etc.
The focus of this report is performance testing of the core services of an OCS Enterprise Edition
Consolidated Configuration as shown in Figure 2 above and as deployed on an HP BladeSystem and
HP StorageWorks solution. As noted above, one requirement for any solution deploying more than
one server in a pool is a hardware load balancer. In this solution a F5 BIG-IP Local Traffic Manager is
used.
3
4. Test methodology
One key element in testing any solution is the user profile. OCS can be utilized in many different
ways, from presence to IM, audio & video conferencing to web conferencing with a mix of internal
and external users, with authenticated and anonymous users. It is impossible to model a user profile to
satisfy every organization’s planned or actual usage.
This document uses a single profile for each of the users of a certain type, as described in the list
below for the 30,000 users under test. While this will not meet every organization’s needs, it provides
a reference point from which to plan an OCS deployment. This profile distribution assumes that an
organization of this size will not deploy all services at the same time to all users, but will take an
incremental approach, where presence and IM are rolled out to most users and VoIP and Audio &
Video conferencing are rolled out to fewer users.
The users are categorized as follows:
• 25,500 (85%) – Presence and IM
• 3,000 (10%) – Presence and VoIP
• 1,500 (5%) – Presence and Audio & Video (AV) conferencing
The presence profile is defined to include the number of times a user’s presence status is changed per
hour. In the profile under test, this is defined as 6 times per hour. The VoIP profile includes
conversations per hour and duration. In the test profile, these are defined as 6 calls per hour at 3
minutes each. The IM and AV conferencing profiles include conferences per user per day and
maximum length, which are defined as 48 per day at 5 minutes each for IM and 2 conferences per
day at 15 minutes each for AV conferencing.
In this test environment 6 load drivers are used to simulate the user activity. Four load drivers are used
for presence and IM activity, and 1 each is used for VoIP and AV conferencing. The load drivers are
configured with a user per second logon rate to vary the rate at which users are logged in to the
environment.
With a static profile, one of the parameters to vary is the user logon rate. This is important because
user productivity and help desk resources can be negatively impacted if the OCS environment cannot
handle peak logon activity. In these tests, the logon rate is varied from 1 user per second per load
driver, up to 6 users per second for each. This means that the cumulative logon rate varies from 6
users per second to 36 users per second.
One important point to note is that since the VoIP and AV Conferencing load drivers have different
numbers of users, they will complete the user logon process before the 4 that are running the presence
and IM load. While the initial logon rate may be up to 36 users per second, that rate will drop as
load generators complete the logon process at different times.
A key element in SQL Server performance is the database cache. The memory in the SQL server
should be sized so that the entire database can be held in RAM which reduces response time since
RAM is much faster than physical disk.
To analyze this aspect of the solution, tests were performed in 2 stages. Before the first stage, the SQL
and OCS servers were rebooted to reset any cache elements. Then a test of each user logon rate was
performed. At the conclusion of that test, the load generators were stopped and the database was
idle, but not rebooted. For the second stage of the test, the load generators were started, requiring a
new logon for each user, but the database was primed from the previous test. This methodology
shows the difference between a cold and a warm database cache.
The simulation performed in this document includes the general topology as show previously in Figure
2. It should be noted that this testing does not include access from the Internet, telephony integration,
or other additional services above those provided by the pool of servers under test.
4
5. As with any business critical application, OCS should be tested in a lab or proof of concept
environment before being deployed throughout an organization. The purpose of this report is to
summarize testing of a 30,000 user OCS Enterprise Edition Consolidated Configuration as deployed
on HP BladeSystem and StorageWorks. The main areas of focus are SQL Server and OCS Server
performance primarily concerning CPU, memory, and disk resources.
Important
As with any laboratory testing, the performance metrics quoted in this
paper are idealized. In a production environment, these metrics may be
impacted by a variety of factors.
HP recommends proof-of-concept testing in a non-production environment
using the actual target deployment and user profile as a matter of best
practice for all application deployments. Testing the actual target
application in a test/staging environment identical to, but isolated from, the
production environment is the most effective way to estimate systems
behavior.
Test topology
The document “Microsoft Office Communications Server 2007 Planning Guide”, from the Microsoft
Technical Library noted above, states that a Consolidated Configuration pool can support up to
30,000 users with four servers in the pool. The specifications from that guide for the OCS servers and
the SQL server are noted below.
OCS Consolidated SQL Server
Configuration pool member (up to 50,000 users)
CPU 2 x Dual-Core 2.6GHz+ 2 x Dual-Core 2.6GHz+
CPU Cache 1MB L2 2MB L2
Memory 4GB 8GB
Network 1Gb 1Gb
2 x 18GB – boot/system
1 x 36GB 15K – database log
Disks 2 x 18GB – boot/system
1 x 36GB 15K – database log
8 x 36GB 15K – database files
For this test, the pool of 4 servers was divided with two servers having AMD™ processors, and two
servers having Intel® processors
Figure 4 shows the architecture of the OCS EE Consolidated Configuration under test. For
configuration details for the F5 BIG-IP Local Traffic Manager (LTM), please visit the F5 website at:
http://www.f5.com/pdf/deployment-guides/microsoft-ocs-ltm94-dg.pdf
For OCS Enterprise Edition installation and configuration details please refer to “Office
Communications Server 2007 Enterprise Edition Deployment Guide” at the OCS Technical Library
previously mentioned.
5
6. Figure 4. OCS Enterprise Edition Consolidated Configuration architecture under test
HP ProLiant BL460c and BL465c servers were used for the OCS pool. These servers each have a total
of 4 processor cores – two Dual-Core Intel Xeon® 5160 3.0GHz processors or two Dual-Core AMD
Opteron™ 2220 2.8GHz processors, respectively. All four servers used 4GB of RAM and a single
gigabit Ethernet interface.
The SQL Server was a BL465c also with two dual-core 2.8GHz processors and 8GB of RAM. Each
server used a RAID1 mirror pair for the operating system. The details of the SQL Server storage
subsystem will be described later in this document.
The HP BladeSystem c7000 Enclosure was fully configured with six power supplies and 10 fans. For
Ethernet and Fibre Channel connectivity, two HP GbE2c Ethernet Blade Switch for c-Class
BladeSystem and two Brocade 4Gb SAN Switch for HP c-Class were used.
For more information about deploying OCS on HP BladeSystem, please read this document
“Designing a Microsoft Office Communication Server 2007 Architecture with HP BladeSystem”
located here: http://www.hp.com/solutions/microsoft/uc
6
7. SQL Server storage subsystem
The SQL Server storage subsystem is critical to the overall performance of the OCS solution. In this
report two different storage subsystems were tested. The BL465c SQL Server was first paired with an
HP StorageWorks SB40c storage blade to provide back end database services. The SB40c uses the
HP Smart Array P400 serial attached SCSI (SAS) storage controller for the six 72GB 15K RPM Small
Form Factor (SFF) SAS disks in the storage blade. The P400 has 2 channels with 2 disks on one
channel and the other 4 disks on the other channel. The SB40c was configured so that the 2 disks for
the database logs were on one channel, and the other 4 disks supported the database files on the
other channel. Figure 5 shows that configuration in the Array Configuration Utility (ACU). Note the
‘Port 1I’ and ‘Port 2I’ notation which indicates through which of the 2 internal ports those disks are
controlled. Logical drives from each array are presented to Windows® as a RAID1+0 volume as
shown in Figure 5.
Figure 5. SB40c Storage Blade configuration, logical and physical view
In this series of tests, firmware version 4.12 and driver version 6.8.0.64 were used for the Smart
Array P400 controller. The Microsoft storport.sys driver was updated to version 5.2.3790.4021 with
KB932755.
For another series of tests, the SQL Server was paired with an HP StorageWorks 1500cs Modular
Smart Array (MSA1500). This Fibre Channel array offers the option of clustering the SQL Server for
high availability and also offers a larger number of disk drives for increased performance. We’ll
compare the results of these two storage solutions later in this document.
The MSA1500 array was configured with eight 146GB 15K parallel SCSI disk with 2 disk drives for
log files and 6 for database files. These disks were configured as shown in Figure 6. The MSA1500
was configured with firmware version 7.00 which supports active/active controllers. For more
information on implementing active/active MSA1500 controllers, please refer to this document: “HP
StorageWorks 1000/1500 Modular Smart Array application note Migrating to active/active
controllers in Windows environments” at
http://h20000.www2.hp.com/bc/docs/support/SupportManual/c00841968/c00841968.pdf.
The physical and logical configurations of the array are outlined in Figure 6.
7
8. Figure 6. MSA1500cs configuration, logical and physical view
The Emulex LPe1105-HP 4Gb FC HBA for HP c-Class BladeSystem was used to connect the BL465c
server to the Fibre Channel SAN with driver version 7-2.00a12 and the default HBA parameters.
Test results
For each test run the performance can be viewed in two parts, the ramp up and the steady state. The
ramp up is defined as the period during which the load generators are logging in users and
performing the first several minutes of the user profile. Steady State is defined as the period when the
system is in relative equilibrium with limited variation in performance elements such as CPU utilization,
disk IO, etc.
At the OCS Team Blog site, Microsoft posted an article “How can I tell if my server is healthy in less
than 10 counters?” http://communicationsserverteam.com/archive/2007/09/10/9.aspx This article
outlines the following counters, the applicable server and the range to watch.
Performance Counter Server Role Range
Processor% Processor Time OCS Server less than 80%
less than 100msec at
LC:USrv – 00 – DBStoreUsrv – 002 – Queue Latency (msec) OCS Server
steady state
less than 100msec at
LC:USrv – 00 – DBStoreUsrv – 004– Sproc Latency (msec) OCS Server
steady state
LC:SIP - 07 - Load ManagementSIP - 000 - Average Holding Time
OCS Server Less than 6 seconds
For Incoming Messages
LC:SIP – 01 – PeersFlow Controlled Connections counter OCS Server zero during steady state
LC:SIP – 04 – Responses SIP – 051 – Local 503 Responses/sec OCS Server zero during steady state
LC:SIP – 04 – Responses SIP – 053 – Local 504 Responses/sec OCS Server zero during steady state
LC:SIP – 01 – PeersSIP – 017 - Sends Outstanding OCS Server zero during steady state
8
9. Performance Counter Server Role Range
MSSQL Buffer ManagerPage life expectancy (seconds) SQL Server greater than 3600
Processor% Processor Time SQL Server less than 80%
The blog article also mentions SQL Server disk latency as a critical factor, but it does not name it as
one of the 10 counters and does not outline the counter thresholds. We’ll look more at SQL Server
disk performance later in this section.
During the logon process, one of the key factors that drives SQL Server load is the user logon rate,
measured in users per second. In this report the following user logon rates are examined: 6, 18, 30,
and 36. Keep in mind that in this test, no users are logged in at the start of the test, and all are
logged in at the end of the logon period. Production environments will not likely see this pattern. In
many organizations users may stay logged in and mark their presence as ‘Away’ or ‘Busy’ until they
return for their next work day or they may remain logged in on their mobile device.
The table below provides a high level summary of the test results for the user logon rates mentioned
above and for the steady state of each test run.
SQL Server CPU% OCS Servers CPU% Storage Storage cache
Logon rate ramp up/ ramp up/ array used setting
Exceptions
users per second steady state steady state Read/Write
Ratio
6 20/30 5/20 SB40c 50/50 None
18 40/30 5/20 SB40c 50/50 None
30 55/30 7/20 SB40c 50/50 None
36 65/30 7/20 SB40c 50/50 None
36 70/30 7/20 MSA1500 50/50 None
The exception column represents any exceptions against the counters from the Microsoft Blog article.
In each of these test runs there were no exceptions against those counters.
9
10. In the test runs with 18 or more users per second, there were SQL Server CPU spikes of an additional
20-40%, but these spikes were brief and did not impact any of the key performance counters on the
SQL or OCS servers. An example SQL server ramp up and transition to steady state is shown in
Figure 7 below. The hourly CPU spikes are caused by the load generators updating their calendar
status every hour.
Figure 7. SQL and OCS server ramp up and steady state CPU utilization
10
11. Every 5 minutes the SQL Server flushes the transaction logs to the database resulting in a spike in
writes/second to the database disk. If the storage subsystem cannot handle these spikes, then the
queue and sproc latencies can be impacted such that communications between the OCS servers and
clients can be throttled. Since there is a spike every 5 minutes, analyzing averages over a long period
of time will not be useful in finding bottlenecks. Figure 8 is an example of one set of such disk IO on
the SQL Server.
Figure 8. SQL Server write IO
The blue horizontal line is the disk write IO to the log drive and is fairly constant during this sample
and during steady state of a test run. The green line represents the disk write IO to the database disk.
Each of the spikes are at 5 minutes intervals and occur when the logs are flushed to the database
disk. Both the green and blue lines use the scale on the left of the graph. The red and the pink lines
represent disk write latency and use the scale on the right of the graph. Log disk write latency peaks
at about 21ms and database disk latency peaks at 142 ms, but the latencies are near zero for much
of the time.
In looking at another of the test runs, of the 2331 data samples for database disk writes per second,
only 190 of the samples had values above zero with a maximum of 1230 writes/second. Of the
samples that were non zero, the average was 462, compared to an average of 32 writes/second for
the entire sample set. Similarly with write latency for that sample set, the average of the entire set is
2.5ms, but the average of the non-zero samples is 30ms with a maximum of 180ms.
11
12. With properly sized memory for the SQL Server, the database should reside completely in memory
during steady state which reduces the number of reads from the database disk and improves response
time since memory access is much faster than disk access. Figure 9 shows this on a newly rebooted
SQL server such that the database cache has not been primed, or read in to memory.
Figure 9. SQL Server database read IO with 8GB memory
This sample includes the ramp up and transition to steady state. The disk read IO peaks when the test
starts, levels off a bit during ramp up, and increases in the 10:10:00 time frame as the database
activity shifts from user logon to user activity. The chart also shows a steadily increasing database
page life expectancy, which means that database pages are expected to be held in memory longer
and reducing access time to that data.
12
13. Figure 10 shows the same configuration, but with 4GB of memory in the SQL Server. Rather than
tapering off to near zero, the database reads hover near 50 per second and peak at over 250. This
figure also shows that the database page life expectance never goes over 1100, which should not be
below 3600 during steady state. This is an example of a SQL Server with not enough RAM.
Figure 10. SQL Server database read IO with 4GB memory
As the solution was designed and deployed, these test results show that none of the disk, CPU or
memory subsystems were stressed to the point of impacting OCS Server functionality for this user
profile. Reducing the memory on the SQL server to 4GB shows stress that could lead to a poor user
experience, as could reducing other system resources.
Conclusions
This testing shows that deploying an OCS Consolidated Configuration as defined on HP BladeSystem
and StorageWorks for 30,000 users with this profile should provide a very positive user experience.
This is a solution that businesses can deploy and grow in to. There are server and CPU resources to
spare to allow for growth as users become more accustom to the features and expand the way they
use the system.
13