Sql server consolidation and virtualization


Published on

Presented in front of the Bulgaria's SQL and BI User Group

Published in: Technology
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • Hardware resources. Each consolidation approach has a different effect on hardware resource usage. Consider whether 32-bit or 64-bit hardware will provide the best performance and scalability for your solution. The general trend is towards the use of 64-bit hardware. Workload isolation. While the goal of consolidation is to reduce the number of servers as much as possible, you should profile the workloads performed by different database applications, and consider whether they should be isolated from one another for performance or other reasons. Application compatibility. Some applications may have specific requirements for data access protocols, collation, or other configuration options that might be incompatible with other applications. Manageability. Consolidation generally simplifies management by centralizing the resources that require management. However, you must consider the effect of maintenance tasks, such as backups and index management, on the consolidated system. You must also consider the security implications of consolidation, because operators with administrative rights in one database system might not require administrative rights in others. Availability.Your consolidation strategy might affect your choice of a high-availability solution. For example, database mirroring provides failover protection at the individual database level and server clustering protects the entire instance. Additionally, database mirroring provides for lesser data loss in case of failures.. Security. In some cases, it may be prudent to isolate database applications for security reasons. For example, one application might require that a feature be enabled in the database server, such as common language runtime (CLR) integration or Database Mail. If you isolate this database you can enable the required functionality while retaining a minimal surface area for other database applications.
  • For complete isolation at the operating system level, you can use a virtualization technology such as Microsoft Windows Hyper-V to host multiple virtual database servers on a single physical serverUsing virtualization, you achieve a good level of isolation between databases with different workloads, security requirements, manageability requirements, or compatibility requirements. The benefits - minimize the number of servers and licenses required and simplify the network infrastructure.Infrastructure simplification. In the worst cases, SQL Server sprawl results in multiple underused servers spread all over an enterprise. Consolidating databases through virtualization will reduce the number of physical database servers, with 120 to 200 database instances being hosted on one HP Enterprise DBC Appliance. After consolidation, the new physical servers will run at higher usage levels, and the original physical assets can be decommissioned and removed from the network.Sustainable IT Calculations of power usage indicate that consolidating 400 SQL Server instances onto three HP Enterprise DBC Appliances reduces energy costs by 86 percent and cut cooling costs by up to 93 percentBusiness agility. DBC does not just cut costs; it can also make organizations more agile and responsive. Virtual machine–based provisioning of database instances means that business units can request the resources to create a database and have that instance approved, built, and operational that same day.Use standardization to minimize administrative overhead.Implement virtual machine templates.Manage the consolidation environment effectively.Control virtual server sprawl.Choose the right licensing arrangements.Minimize governance issues
  • Dynamic disks cons fragmentation on physical layerhttp://www.aidanfinn.com/?p=11390CPU LimitationsHyperV – 4vCPUs
  • - Slightly higher processor utilization is required to achieve the same throughput with virtualization due to minimal additional processor overhead.- Native instances and virtual instances achieve the same level of scalability with 11 to 13 percent processor overhead.
  • The results of this test are consistent with previous tests and are summarized as follows:Slightly higher processor utilization is required to achieve the same response time with virtualization due to minimal additional processor overhead.Native instances and virtual instances achieve the same level of scalability with 11 to 13 percent processor overhead.The scalability suffers when I/O becomes a bottleneck.
  • Virtual memory management requires multiple layers of address translation between the guest VMs and the hypervisor that manages the VMs and the physical server. Until recently, the burden of managing the translation layers was performed by the hypervisor in software using page-table shadowing that introduces significant non productive processing overhead. The latest servers, not only provide hardware-assisted virtualization through their on-board processors, they also perform virtual memory management and other functions that were traditionally fulfilled by the hypervisor. This capability substantially benefits the workload because it improves overall system performance. For example, the latest processors from AMD® (Advanced Micro Devices) or Intel® both support hardware-enabled SLAT. AMD describes its SLAT technology as AMD Virtualization™ (AMD-V™) Rapid Virtualization Indexing (RVI) or nested paging tables (NPT). Intel describes their SLAT technology as Intel Virtualization - Extended Page Tables (VT-EPT).See also http://sites.amd.com/us/business/it-solutions/usage-models/virtualization/Pages/amd-v.aspx. See also http://www.intel.com/technology/virtualization/technology.htm?iid=tech_vt+tech.
  • The results of this test are summarized below:There is increased throughput due to consolidation using virtualization.When the OLTP workload runs using Hyper-V (solid orange line) with SLAT enabled (no over-commit), it demonstrates fairly linear scaling. When the same OLTP workload runs using Hyper-V (dotted red line) without SLAT, throughput begins to wane after three VMs are added.Depending on the workload, SLAT can be used with CPU over-commit (dotted green line). However, CPU over-commit can cause a decrease in performance as more VMs are added.
  • For this test, the SQL Server workload was initially run using a Hewlett-Packard ProLiant DL785 server using the Quad Core AMD Opteron™ 8358 processor (G5), formerly called Shanghai, 2.4 GHz with 32 logical processors and 128 GB RAM. Then the processor was upgraded to the latest Six Core AMD Opteron™ 2435 processor (G6), formerly called Istanbul, 2.6 GHz with 48 logical processors and 128 GB RAM. The SQL Server workload was run using the updated processor, and the results were compared.Note that the upgrade from the Shanghai processor to Istanbul was performed simply by swapping out the processor as both processors are drop-in compatible.The features and benefits of the Istanbul processor include:Six cores per socket. Offers improved performance and performance/watt, as compared to prior generations, for multi-threaded environments such as virtualization, database, and Web serving. HyperTransport™ Technology Assist (HT Assist). Reduces cache probe traffic between processors that can result in faster queries in 4-way and 8-way servers. This can increase performance for cache sensitive applications such as database, virtualization, and compute intensive applications. HyperTransport™ 3.0 Technology (HT3). Increases the interconnect rate from 2 GT/s with HT1 up to a maximum 4.8 GT/s (gigatransfers per second) with HT3. This helps to improve overall system balance and scalability for scale out computing environments such as high performance computing (HPC), database, and Web serving. AMD-P suite of power management features. These features are built into AMD Opteron processors, including AMD PowerNow!™ Technology, AMD CoolCore™ Technology, and Dual Dynamic Power Management™. These innovations help reduce energy use and cost. AMD-Vi. Supports I/O level virtualization that enables the VM to have direct control of the devices. In order to use this feature the SR5690/SR5670 chipsets are required. AMD-Vi improves performance through the direct assignment of devices to a guest operating system. In addition, AMD-Vi improves isolation to help improve VM security.See also http://www.amd.com/us/products/server/processors/six-core-opteron/Pages/six-core-processors-with-amd-chipset.aspx.See also http://www.amd.com/us/products/server/processors/six-core-opteron/Pages/six-core-opteron-key-architectural-features.aspx.See also http://www.amd.com/us/products/server/processors/six-core-opteron/Pages/six-core-opteron-product-brief.aspx.
  • Data Compression and Backup CompressionConsolidated systems must often support very large volumes of data, which leads to increased costs, especially when storage is handled through expensive SANs. Large data volumes can also reduce the performance of a system because much of its workload can involve moving data. One of the biggest bottlenecks on a database system is I/O, but because compression minimizes the I/O for a given volume of data, performance is improved. SQL Server 2008 Enterprise Edition supports data compression, which enables you to store your data more effectively, reduces the cost of storage on a consolidated system, and significantly improves performance for large I/O-bound workloads. The amount of disk capacity that you can save can be significant. A recent report by Forrester Consulting (Total Economic Impact of SQLServer 2008 Upgrade) found that for one organization:“The data compression feature in SQL Server 2008 [Enterprise Edition] has significantly reduced the amount of storage required. On average, the customer is reducing storage space by 50%. It was able to free up one terabyte (TB) of storage space for the existing 2005 SQL Server databases. Additionally, the amount of storage added in future years will be reduced from 500GB to 250GB per year. The result is that no storage needs to be added to support the SQL Server databases for the next three years.” Compression also helps in reducing the cost of business continuity. For disaster recovery purposes, one or multiple copies of the data is backed up on disk, tapes and other storage mechanisms, which directly translates to costs in terms of storage hardware. Backup compression addresses these concerns on both fronts: the size of the backup is reduced substantially, depending on the content of the data being compressed, and the overall time to back up and restore is similarly reduced. However, backup compression comes at the cost of CPU usage, so this might have an impact on your online workload.SecurityConsolidated databases bring additional security concerns, such as ensuring that users and administrators can access the data to which they require access, while restricting unauthorized access to other data on the same server. SQL Server 2008 Enterprise Edition has new and improved features that help to keep your consolidated systems secure:•        Transparent data encryption (TDE). Use TDE to encrypt and de-encrypt entire databases automatically. The process is transparent to applications, which do not require modification to take advantage of TDE, with the result that the development effort required to achieve greatly improved levels of security is virtually eliminated. •        Extensible key management. SQL Server’s support for enterprise key management enables you to integrate SQL Server encryption into your enterprise cryptographic infrastructure.•        Levels of isolation. You can isolate applications at three levels:o        Database. Control access to individual database on a user-by-user basis.o        Server. Control access to individual instances by using logins. In SQL Server or Windows integrated authentication.o        Operating System (physical or virtual). Control access to the operating systems that host SQL Server instances by using local Windows or Active Directory® accounts.•        SQL Server Audit. A key security requirement in an enterprise is the ability to audit activity in databases automatically and to be able to generate consistent audit reports for all database server instances. Consolidation simplifies this process, because there are fewer servers to manage. SQL Server 2008 Enterprise Edition includes extensive auditing capabilities that enable you to log activity at both the server and database level, including changes to the data and schema. You can define auditing filters to specify the actions to audit based on actions, objects, and security principals, and specify where the auditing information should be recorded. After you configure auditing on one instance of SQL Server 2008, you can apply these settings to other instances on the server. You can consolidate the audit results for your entire organization in one place, use Reporting Services to create customized reports, and use Analysis Services to analyze the centralized audit data.Scalability and PerformanceConsolidation makes use of existing hardware as well as, in some cases, new hardware. The decision is based on workload types and future workload projections. A scalable server such as SQL Server 2008 will mitigate some risk by helping organizations make decisions about hardware with the confidence that SQL Server can scale when required. The other big challenge with consolidation is resource contention. Resource prioritization will help application business needs. SQL Server 2008 Enterprise Edition provides some key new features to meet these challenges:•        Resource Governor. Resource Governor enables organizations to define resource limits and set priorities for different workloads. This enables the differentiation of workloads within a single instance on a consolidated server and the control of memory and processor resources to prevent runaway queries or large workloads from adversely affecting the performance of mission-critical workloads. This ability to proactively control use of resources provides more predictable performance of consolidated data services.•        Scaling up consolidated servers. SQL Server Enterprise Edition also provides optimal performance and scalability by providing several features to improve concurrency and prevent locks caused by one process blocking other processes on the server. The locking system in SQL Server 2008 provides greater concurrency and improved lock escalation compared to previous releases. You can organize tables and indexes into partitions on a disk and switch in and out of these partitions individually to reduce contention and improve overall performance. Furthermore, the release of Windows Server 2008 R2 includes support for 256 logical CPUs, offering SQL Server 2008 the ability to scale to unprecedented levels. For more information, see: http://www.microsoft.com/presspass/features/2008/nov08/11-06winserverr2.mspx•        Data and backup compression. As discussed earlier, these technologies can help you save a great deal of storage space, which makes it possible to scale up without purchasing more storage.ManageabilityServer consolidation helps in the centralization of administrative functions. SQL Server 2008 provides a suite of tools to manage, administer, monitor, and troubleshoot consolidated data systems. Centralizing management with SQL Server Management StudioSQL Server Management Studio (SSMS) is a great tool for managing a consolidated database environment. SSMS can be used to manage multiple native or virtual SQL Server databases and instances on local or remote servers. SSMS is backward compatible, so you can use it to manage SQL Server 2005 and SQL Server 2000 too. The same consistent SMMS interface is used for managing SQL Server Database Engine instances, SQL Server Analysis Services, SQL Server Integration Services, and SQL Server Reporting Services.Enforcing configuration policy compliance with policy-based managementOrganizations increasingly want to enforce a set of policies to follow regulatory or best-practice requirements. Previously this has involved configuring many tables, databases, and instances, and then periodically checking compliance against these policies. In SQL Server 2008, you can consolidate configuration management by using policy-based management to apply policies to multiple servers, databases, tables, and other targets in the enterprise. Policy-based management enables you to:•        Ensure compliance with policies for system configuration.•        Monitor or prevent changes to the system by authoring policies for the desired configuration.•        Scale management across multiple servers.•        Reduce total cost of ownership by simplifying administration tasks.You can proactively enforce policies by using triggers, or on a schedule by using SQL Server Agent. Additionally, you can use ad-hoc execution to check or configure target objects against policies in real time.Centralizing and consolidating monitoring using Data CollectorSQL Server generates valuable data to assist in performance tuning and troubleshooting. Although this is extremely useful, it can also be time-consuming to use multiple tools to identify problems. SQL Server 2008 includes Management Data Warehouse and a suite of tools to help you troubleshoot, tune, and monitor the state of one or more instances of SQL Server 2008. The centralized management of data collection, storage, and reporting of the various performance metrics can help in consolidation planning and ongoing operations.
  • The DL980 is the high-end and flag ship of our ProLiant server line– it’s highly resourced server with 8 Intel Xeon processors supporting up to 64 cores two terabytes of memory up to 16 PCI-Express 2.0 IO slots features the PREMA Architecture unique to the DL980 and provides greater scalability and resiliency. Optimized for data intensive enterprise workloads, the DL980 is designed with balanced scaling, self-healing resiliency and breakthrough efficiencies.Let’s talk about a little but about this PREMA architecture which represents the design direction for our scale-up x86 designs, As you start to add more and more processors into the system there is infrastructure at the system level – that we need to implement – this is what we call the PREMA architecture – for our 8w systems to enable higher levels of resiliency and scalability. Its based on HP decades of experience in delivering mission critical enterprise class servers.. So we’ve basically leveraged some our Superdome scaling and resiliency know-how and engineered into this server.Balanced scaling is really an important concept when you talk about server design. Intel has done a great job in maintaining the moore’s law climb on the cpu – but there are other resources like memory and IO in the system that are required to match those CPU increases. Balanced scaling is about how we have matched the IO, memory resources with the cpu in order to scale in without bottlenecking. Workloads like transaction processing, decision support and virtualization is not a cpu bound workload – it’s an IO bound workload – it’s a memory bound workload and you need those components to match the cpu capability out of Intel.Our balanced scaling design features Smart CPU caching which has resulted in some world record performance results. Smart CPU caching is an exclusive DL980 feature, is an attribute of the PREMA Architecture and is leveraged from our Superdome SMP experience. This feature enables more efficient scale-up processing capacity resulting in near linear performance It utilizes a node controller design and minimizes the administration and inter-processor traffic that can grow exponentially as you add processors. Bottom line - and this has resulted world record benchmark results and in more efficient scaling than competitive systems.With self healing resiliency capability of the DL980 G7 severs, our customers will see a 200% boost in system availability when compared with the previous generation 8 sckt DL785 server. This is in part is due to another PREMA feature called redundant system path. Resilient system fabric is also based on Superdome’s Crossbar fabric. Resilient system fabric connects the 8 processors and provides for much higher interconnect bandwidth and lower data error rates for improved resiliency.Breakthrough efficiency has been a hallmark of the ProLiantProLiant heritage of management and power efficiency and bottom line our customers can manage their 2 socket 4 socket and 8 socket servers with all the same ProLiant tools and features. Management and power efficiency continues to be a core ProLiant competency as evidenced by a number of new innovations in power and systems management such as the highly efficient Platinum-rated power supplies in the DL980, new iLO3 for remote management and then tools like Insight control that reduces operational costs.New iLO3 remote management solution eases administration burdensThe first Platinum level power supply (listed by the industry group 80 Plus) with 94% efficiency providing $20-$80 savings per power supply per year versus competitive productsTurbo-charged performance – 800x Faster Remote Console: Customers will experience remote console performance that is equal to the performance of KVM and software-based remote management solutions and over 8x the performance of the current iLO 2 management processor. Insight Control enables a ProLiant server to be deployed 12x times faster compared to doing it manually using vendor provided CDs or DVDsCommon Power Supplies, Hard Drives and Memory allows self service customers to reduce spare parts inventoriesInsight Control/Insight Dynamics - Reduce operational expenses by $48,300 per 100 users over 3 years by enabling HP Insight ControlManaging health proactively has enabled HP ProLiant customers to reduce unplanned downtime by up to 77%.*** Additional Information ***Self-healing resiliency Maximizes application uptime with a 200% boost in server availability(Note 200% boost in availability is based on System crash rates comparison between the DL980 G7 to the DL785 G5. System crash rate is determined by availability features such as hot-swap components, redundant paths, ECC, and tolerant links such as QPI.Resilient system fabric - HP PREMA Architecture extends the advanced reliability of the Intel Xeon processor 7500/6500 series in the ProLiant DL980 G7 with a resilient XNC fabric. This interconnect fabric provides higher interconnect bandwidth to improve performance and scaling, and availability features consistent with the QPI fabric.All fabrics in the system (Intel's SMI & QPI and HP's XNC system fabric) have many resiliency features to keep your system running in the presence of errorsBalanced scaling – delivering world record performance to handle exponential data growth by eliminating system bottlenecks enabled by HP PREMA Smart CPU cachingBreakthrough efficiency – ROI of 2 months for with HP-only innovations like Thermal Logic, iLO3, Insight controlXNC (node controller) Fabric—Connects HP node controllers to each other improving scalability and performanceHP XNC fabric has resiliency features that extends the functionality of Intel's QPI and SMI links to XNC system fabric.The XNC fabric also has redundant connections, but a complete link failure will cause a system crash. The redundancy allows the system to recover from a complete link failure by re-boot to route around the failed link. This allows the system to be used while waiting for a service call; which allows the service to be delayed until a convenient time.
  • MARKET NEEDS:Radically lower the cost for managing and maintaining compute infrastructures Integrated management and monitoringElastic capacity provisioningAutomated load balancing and servicingSelf Service deployment and monitoringTransparent declarative high availabilityMetering and chargeback capabilitiesAnother exciting announcement is the development of a Private Cloud appliance.Customers today have Private Cloud reference architectures for Hyper-V, and have been asking us for more. They have asked us to include the database, and to help them deploy it faster with HP hardware.When customers think about Private Cloud, they think about a solution that can provide resources on demand, that IT or users can self-provision, that has minimal downtime, can integrate with existing technology, and can help them in the path to metered use/chargeback.HP and Microsoft are developing a database private cloud appliance, to consolidate and manage data-centric solutions. We have many customers doing this today, and with the upcoming consolidation reference architecture (under the brand “Fast Track”), and the appliance (for customers that would rather have it all the HP hardware and Microsoft software installed and deployed in a few days), we expect to provide an even faster path to value, with higher levels of performance and availability.SOLUTION:Key Points:Complete – The DBC appliance includes hardware, software, and support – everything you need to identify, migrate, and manage physical servers to create a consolidated, virtualized datacenter or private cloud. The hardware and software and pre-tuned and configured, and customers benefit from a single point of support, making the DBC a comprehensive all-in-one solution.Optimized – You can consolidate 200 database servers in a single DBC rack. The DBC hardware is optimized for database workloads, and the appliance has been designed for optimal performance and energy efficiency. Additionally, the centralized monitoring and control of virtual machines that the appliance offers – along with dynamic allocation of compute and storage resources helps you optimize your database servers for the workloads you need to handle at any given time.Agile – Because the BDW is provided as an appliance, customers can easily implement it within their environment. This is MUCH faster than a bespoke solution, or even a Fast Track implementation would take. And as the customer’s requirements grow, there’s a straightforward migration path from the base configuration to the full configuration. The move to a private cloud increases IT agility (through dynamic pooling of hardware resources) and organizational agility (through self-service infrastructure provisioning and business unit financial accountability based on resource utilization)
  • The HP Enterprise Database Consolidation Appliance Optimized for SQL Server 2008 R2 is a virtualized appliance tuned for consolidating hundreds to thousands of tier 2 and tier 3 SQL Server databases. Built on Microsoft best practices and HP’s architectural guidance with HP-installed Converged Infrastructure components and on-site deployment, the appliance can be operational quickly. The appliance is designed to help ITand database administrators quickly provision SQL Servers with predictable performance, simplify SQL infrastructure management, and control database sprawl. The benefits of consolidating SQL instances onto the HP Enterprise Datbase Consolidation Appliance are:Complete virtualized system that improves resource utilization, reduce space requirements, and lower power consumption,Integrated management stack to centralize control and manage the infrastructure,Self-service Portal simplifies on-demand SQL provisioning and meters usage for chargeback to departments, Integrated tools that easily discover SQL instances, size and migrate the databases onto the virtual machines running on the appliance 
  • The HP Enterprise Database Consolidation Appliance is a rack of HP ProLiant BL465 blade servers and HP P2000 arrays, and storage. Integrated with the hardware is the management stack that includes: Windows operating system, management software to monitor and manage the virtualized resources, provision SQL instances, discover, size, and migrate SQL instances, as well as alert managers of hardware and software events before they fail  The appliance is a set of pre-tuned hardware integrated with management software. Microsoft Windows 2008 R2 Datacenter OEM and HP tools are pre-installed and licensed. The Microsoft Database Consolidation 2012software package is also pre-installed but licensed separately. The software package is a bundle of all the Microsoft management software and custom tools needed to manage the appliance (see details below).
  • [Edit] Add a small picture
  • [Edit] Add a small picture
  • Sql server consolidation and virtualization

    1. 1. SQL Server Consolidationand VirtualizationIvan DonevTechnology Lead MSSQL, ES, HPMCITP DBA, Development, BusinessIntelligenceITIL v.3 CertifiedBlog: sqlbg.wordpress.com©2011 Hewlett-Packard Development Company, L.P.The information contained herein is subject to change withoutnotice
    2. 2. Agenda 1 Consolidation 2 Virtualization 3 Market 4 SucceedDefine consolidation SQL Server and What products do we How to start aApproaches and Virtualization have on the market consolidation projectguidance Tools we can use when consolidating 2
    3. 3. CONSOLIDATION “…combining of various units into more efficient and stable larger units…” “… improved cost efficiency from higher utilization of resources, standardization and improved manageability of the IT environment as well as focus on the “GREEN”…”3
    4. 4. Why consolidate• Forces Driving consolidation – Reduced TCO – Optimized administration and licensing costs• Key Enablers – Software – Hardware – Virtualization4
    5. 5. Approaches• Multiple databases into a single database• Multiple databases into a single instance• Multiple instances into single physical server• Multiple virtual machines into single physical server5 HP Confidential
    6. 6. Consolidation guidanceConsiderations• Hardware resources• Workload isolation – OLTP – OLAP• Application compatibility• Manageability and availability• Security – Risks and vulnerabilities• Resource management
    7. 7. Consolidation guidanceGeneral advices• Only Consolidate Homogenous Workloads to a Single Windows Instance• Consolidate only non-Mission Critical Workloads first• Avoid enhancements to the application/database during consolidation• Drive standardization in the enterprise• Do not consolidate applications with EOL < 12 months• Consolidation candidates – Many servers in same physical location and Older / less efficient hardware – Relatively high read rate – Low current CPU utilization / demands – Similar / compatible collation8
    9. 9. SQL Server Virtualization• Complete isolation on OS level• Live migrationVMotion• Good isolation level on: – Workloads – Security – Manageability – Compatibility• Infrastructure simplification• Business agility10
    10. 10. Recommendations and BP with SQL Serverand Virtualization• Storage and IO – Standard storage performance and IO Best practices apply – Dynamic disks are not recommended – Pass-through vs. Fixed Size VHD (VMDK)• Memory – Do not overcommit• CPU – CPU overcommit for higher densities – Limitations• Sizing – Keep it standard (predefined VM “sizes”) – Changing sizes of resources require VM downtime11
    11. 11. Hyper-V I/O and disk latency
    12. 12. Hyper-V CPU Multi-Instance performance13
    13. 13. Hyper-V CPU Multi-Instance performancewith over-commit14
    14. 14. Virtual memory and SLAT The Virtual / Process view The Physical / real view Virtual Machine 1 Physical Memory Pages Virtual Machine 3 Virtual Machine 1 Hyper Visor Operating System15
    15. 15. Hyper-V SQL Server consolidation withhardware-assisted SLAT16
    16. 16. Hyper-V and advanced processors17 HP Confidential
    17. 17. MARKET What is out there that can help us©2011 Hewlett-Packard Development Company, L.P.The information contained herein is subject to change without notice
    18. 18. What is on the marketSQL Server 2008 R2• Data and backup compression• Security and auditing – TDE• Scalability and performance – Resource Governor – Improved lock management – Supported CPUs up to 256 – Compression• Manageability – CMS – PBM – MDW19
    19. 19. What is on the marketHP ProLiant DL980• Balanced Scaling – World record performance enabled by HP PREMA Architecture – Up to 8 Intel® Xeon® E7 Family and 7500 series processors – Up to 10 cores and 30MB L3 cache per processor – Up to 128 DDR3 DIMM slots; 4.0TB max memory with 32GB DIMMS – Up to 16 slots, PCI-E 2.0 or PCI-X – Quad-Port 1GbE embedded NIC upgradable to Dual Port 10GbE• Self-healing Resiliency – Resilient system fabric to maximize uptime – Up to 10X improved in memory error crashes with DDDC – 20+ Resiliency features added with E7 family• Breakthrough Efficiency – Consolidate multiple resource intensive legacy servers on to a single system – HP-only innovations like Thermal Logic, Sea of Sensors, and Dynamic Power Capping for ROI that competitors can‟t touch – iLO3 & Insight Control for tomorrow‟s datacenter capabilities today20
    20. 20. What is on the marketHP Database Consolidation Appliance• Market needs – On demand infrastructure and data based solutions – Self service and shared solutions – Delivering database solutions on demand – Migrate „live‟ older infrastructure, with no downtime – Metered Use• VM Based Solution• Optimized for database workloads• Some of the benefits of private cloud: – Elasticity – Self service – Resource pooling – Control, chargeback21
    21. 21. How does it work? P2V VM Server Windows Server, Hyper-V, System Center #1 VM Server SQL Server EE 2005 #2 VM SQL Server SE 2008 VM Server #3 VMM VM SQL Server SE 2008 R2 VM Newly created VMs. Can reuse existing „IT template‟ VMs. Room to grow Bank of VMs VM1 VM2 VM322
    22. 22. DBC Solution Software Provides base operating system and Active Directory services, as well as the Hyper-V role to host virtual machines Provides database storage for configuration databases used by System Center components and MAP Toolkit Provides configuration and patch management services for the appliance Provides monitoring services for guests and hosts Provides a centralized management interface for virtual machines, physical to virtual migration services, and a library for virtual machine templates and resources Provides incident tracking and resolution services for the virtualized datacenterSystem Center Virtual Machine Manager Provides self-service provisioning of virtual infrastructure, and resource usage-based reporting for accurate business unit charge-backSelf-Service Portal 2.0 SP1 Provides server and software discovery and reporting services to help youMicrosoft Assessment and Planning Toolkitplan migrations from physical servers to virtual machines Insight Control Provides server monitoring and management services for HP hardware 23 HP Confidential
    23. 23. Case studyVblock0 Costs Compared with HP Enterprise DBC Appliance CostsItem VMware Vblock0 HP Enterprise DBC ApplianceHardware (full rack) $300,000 $800,000VMware Enterprise $139,808 $0Windows Server 2008 R2 $95,968 $0SQL Server 2008 R2 purchase $879,840 $439,920Windows Server SA $23,992 $0SQL Server 2008 R2 SA $219,960 $109,980Management software $100,000 (estimate) $140,000Management software SA $35,000Total $1,759,568 $1,524,90024
    24. 24. SUCCESS©2011 Hewlett-Packard Development Company, L.P.The information contained herein is subject to change without notice
    25. 25. How to start a consolidation projectBrent Ozar‟s article• Summary – Inventory of consolidation instances – Categorize – Ask right questions to right persons – Choose consolidation method – Build and test the new environment – Plan migration method – Implement and measure for success• Brent Ozars How to start a consolidation project26 Picture source: http://www.simplewealth.com.au/wp-content/uploads/2010/10/consolidate3.jpg
    26. 26. Tools that can help• Microsoft Assessment and Planning (MAP) Toolkit• Microsoft Consolidation Planning Tool for SQL Server• SQL Server Migration Tool• Microsoft SQL Server 2008 Upgrade Advisor• Virtual Machine Servicing Tool (VMST) 3.0
    27. 27. What‟s left untoldAdditional resources• Brent Ozar‟s Resources – SAN – Virtualization• SQL CAT on SSAS Consolidation Practices• Quest‟s SQL Server Pedia on Consolidation• SQL OS Team on consolidation, virtualization and private cloud• VMWare‟s documentation on virtualization and SQL BPs• High Performance SQL Server Workloads on Hyper-V• SQL Master Academy – Workshop: SQL Server Enterprise Infrastructure Optimization28
    28. 28. Q&A
    29. 29. THANK YOUJoin the HP DBA family at jobs.hp.com