Microsoft Server Virtualization and Private Cloud

Editor's Notes

• #10: Windows Server 2012 Server Virtualization has many new features and enhancements at its core. To show you some of these and their value to your organization lets walk through a number of them Hyper-V Network Virtualization Hyper‑V Network Virtualization extends the concept of server virtualization to permit multiple virtual networks, potentially with overlapping IP addresses, to be deployed on the same physical network. With Hyper‑V Network Virtualization, you can set policies that isolate traffic in your dedicated virtual network independently of the physical infrastructure for fully secure and isolated multi-tenancy. Network Virtualization also provides IP Portability, and the ability for you to move Virtual Machines across physical subnets without changing your address space. You VM’s can keep there IP address whether moving across servers, racks, buildings, geographies or even to the cloud – no more need to reconfigure complex VLANs or adjust your address space to suite the destination environment. Shared-nothing Live Migration The ultimate in flexible VM mobility – moving a running VM without downtime to another Hyper-V host with no clustering, or no shared storage – just a network connection. Move VM’s between clusters without the need to stop, export and import. Move VM’s between Hyper-V servers anywhere in your organization, and even migrate your VM’s to service provides or the cloud without any downtime. Massive Scale Windows Server 2012 Hyper-V now support massive new scale improvements including support for 64 virtual CPU’s and 1 TB for Hyper-V Guests. And up to 320 Logical Processors and 4TB Memory support for Hyper-V hosts. Custer Enhancements Windows Serve 2012 Hyper-V has many new enchantments for clustering including Support for guest clustering via Fibre Channel Enhanced Live migrations to use more available network bandwidth which dramatically increases the performance of Live Migration and enables concurrent Live Migrations with no limits. Massive Scale. will now support up to 64 nodes and up to 4,000 virtual. Encrypted cluster volumes. Hyper-V application monitoring, where Hyper-V and Failover Clustering work together to bring higher availability to workloads that do not support clustering. Virtual machine failover prioritization. Inbox live migration queuing where Administrators can now perform large multi-select actions to queue live migrations of multiple virtual machines with ease and efficiency. And Affinity (and anti-affinity) virtual machine rules. Hyper-V Replica Asynchronous, application-consistent virtual machine replication is built in to Windows Server 2012. It permits replication of Hyper‑V virtual machines between two locations for business continuity and failure recovery. Hyper‑V Replica works with any server vendor, any network vendor, and any storage vendor. Hardware offloading Windows Server 2012 Hyper-V takes advantage of many new hardware offloading features that are supported in Windows Server 2012 including support for SAN-based ODX (offloaded data transfer), IP Offloading, and SR-IOV (through support for Single Root I/O Virtualization networking devices). Ultimately increasing the performance of your virtualized systems. Virtual Fibre Channel Virtual Fibre Channel lets virtual machines connect directly to Fibre Channel–based storage and presents up to 4 virtual Fibre Channel host bus adapter (HBA) ports in the guest operating system that runs in the virtual machine. This provides unmediated access to a SAN from you guests and Hardware-based I/O paths to the Windows software virtual hard disk stack. Guest NUMA Support Windows Server 2012 Hyper‑V now supports NUMA in a virtual machine. NUMA refers to a computer architecture in multiprocessor systems in which the required time for a processor to access memory depends on the memory’s location relative to the processor. With NUMA, a processor can access local memory (memory attached directly to the processor) faster than it can access remote memory (memory that is local to another processor in the system). Modern operating systems and high-performance applications such as SQL Server have developed optimizations to recognize the system’s NUMA topology and consider NUMA when they schedule threads or allocate memory to increase performance. Runtime memory configuration The Dynamic Memory improvements to Hyper‑V in Windows Server 2012 help you reach higher consolidation numbers with improved reliability of Hyper‑V operations. You can make memory configuration changes for your virtual machines without shutting down the virtual machines. If you have idle or low-load virtual machines, Dynamic Memory additions in Hyper‑V let you increase consolidation and improve reliability for restart operations. With runtime configuration changes for Dynamic Memory, overall IT productivity is expected to increase with reduced downtime and increased agility to respond to requirement changes. Hyper-V Network Switch The Hyper‑V Extensible Switch in Windows Server 2012 is a layer-2 virtual network switch that provides programmatically managed and extensible capabilities to connect virtual machines to the physical network. The Hyper‑V Extensible Switch is an open platform that lets multiple vendors provide extensions that are written to standard Windows API frameworks. The reliability of extensions is strengthened through the Windows standard framework and reduction of required third-party code for functions and is backed by the Windows Hardware Quality Labs (WHQL) certification program. You can manage the Hyper‑V Extensible Switch and its extensions by using Windows PowerShell, programmatically with WMI or the Hyper‑V Manager user interface.
• #11: Lets take a look at some of the new scale enhancements that Windows Server 2012 Hyper-V now provides. In some cases in 16 times greater then previous versions. Windows Server 2012 Host Up to 320 Logical Processor support Up to 4TB Memory Up to 1024 Virtual Processors per host Virtual Machine Guests Up to 64 vCPU’s Up to 1TB of Virtual Memory 1,024 active virtual machines Up to 64TB storage per Virtual Hard drive Clustering Scale Enhancements Now supporting up to 64 nodes per a single cluster and up to 4000 running VM’s
• #42: What is it that constitutes Cloud Computing? Infrastructure as a Service (IaaS): A model for provisioning hardware (compute, storage, networking, etc...) to the end user where management of the underlying fabric is controlled by the provider, but the end user maintains control of the operating system and applications installed within. This system usually includes a metered-byuse cost model and allows the end user to expand/contract their use of the infrastructure as needed, usually via self-service portals. Examples include: FastHosts, Go Daddy, Rackspace, Amazon EC2, vCloud Express Services (e.g., BlueLock, Hosting.Com, Melboure IT, Terremark), private clouds deployed/managed by IT as service to business units (internalIT’s end-customers), Azure Service with the VM Role Platform as a Service (PaaS): A model for delivering complete development platforms as a cloud service. PaaS offerings facilitate development, testing, deployment and on-going maintenance of applications without the cost of buying the underlying infrastructure and software environments. Examples include: Microsoft Azure Platform, Google App Engine, VMforce.com Software as a Service (SaaS): A model where an application is delivered over the Internet and customers pay on a per-use basis. It is the most common form of cloud computing delivered today. Examples include: BPOS, Salesforce.com, Hosted Exchange, Salesforce.com Core attributes of Cloud Computing: Scalable Elastic Multi-tenancy Metered by Use Self Service
• #43: These are but illustrative examples (there’s a lot more) of how we’ve implemented our core learnings from operating large scale cloud services as capabilities/features within Windows Server so you can also benefit from the resulting agility and cost benefits. We’re committed to this approach as we continue the journey towards realizing our Cloud OS vision. 1. High performance storage on industry-standard hardware File & storage services – Microsoft’s large scale cloud services implement hot-hot nodes in many cases so the application services can seamlessly failover in case of node failures. Analogously, Active-On technologies in Windows Server 2012 are those capabilities than can survive the failure of a key component without disrupting the service provided - for example, the File and Storage Services server role can be deployed in an active-active cluster which can endure the failure of a server while maintaining connectivity for clients. Even open file handles are maintained. Offloaded data transfer – This feature came about by applying a key design principle that Windows Azure uses to Windows Server 2012, which is to push work to hardware as much as possible so CPU cycles can be optimally leveraged for the hosted application. Storage spaces - We routinely provision commodity hardware in our pods thereby making hardware failure a non-event; we wanted to provide enterprise customers with the ability to easily and flexibly provision inexpensive storage. Windows Server 2012 enables you to use file servers for even critical workloads like Hyper-V server and SQL with arguably similar reliability and performance characteristics as more expensive storage (e.g. SAN). 2. Multi-tenant environments with isolation Server core - Small footprint with server core so you do not have to worry about updating or fixing unnecessary code, meaning less churning or required reboots for the host. This is similar to how we designed the Windows Azure hypervisor where we’ve only retained (and optimized) critical code paths that are applicable to large scale cloud scenarios. Hyper-V Network Virtualization - Windows Azure was designed from the ground up to be multi-tenant environment with isolation given that it hosts apps and infrastructure that belongs to different customers. Over time, it became increasingly evident that this same requirement would be pertinent to enterprises implementing cloud infrastructure (albeit at smaller scale) given the need to distinguish between infrastructure allocated to different departments, integrating infrastructure from acquisitions, assigning costs etc. We then built the ability to isolate networks logically in multi-tenant enterprise environments with Hyper-V using similar architectural considerations. Web-sites & virtual machines - These are great examples of how we’ve bought consistent experiences and cloud-optimized services from Windows Azure to Windows Server and System Center environments. 3. Software-defined Networking Hyper-V Network Virtualization - The network virtualization features within Windows Azure and Windows Server enable customers to easily extend their on-premises datacenters into third party datacenters (e.g. Windows Azure or service providers) by allowing for flexible placement of VMs along with the ability to continue using the same IP address. Cross-premises connectivity- The cross-site connectivity capabilities offered by Windows Server 2012 and Windows Azure (Windows Azure Virtual Network) complement each other well for VPN-like scenarios. Network QoS – Windows Server 2012 provides the ability to programmatically adhere to a given SLA by guaranteeing minimum bandwidth available to a VM or a port. We’re assessing how to implement this in Windows Azure (albeit at a much larger scale) to shift from “best effort” to quality of service guarantees, thus providing an example of this bi-directional virtuous learning between our platforms. 4. Policy based Automation Cluster aware updates (CAU) – Windows Server 2012 lowers downtime by coordinating software updates on all servers in a failover cluster while maintaining continuous availability as appropriate. For example, in case of applications with continuous availability features such as Hyper-V with live migration, or an SMB 3.0 file server with SMB Transparent Failover, CAU can coordinate automated cluster updating with no impact on service availability. This is very similar to Windows Azure which offers the ability to maintain high availability for its cloud services (e.g. web role/worker role) and virtual machines through Upgrade Domains (UD) and Availability sets. Dynamic optimization – System Center 2012 - Virtual Machine Manager provides the ability to load balances the resources within a host cluster automatically using a policy based approach (e.g. move workloads VMs between hosts based on defined CPU, memory, or I/O thresholds). Such policy based automation is implemented by the Windows Azure Fabric Controller too using various load balancing approaches. 5. Application elasticity Service templates – Similar to how an application owner (or developer) can specify application instance counts and scale thresholds in Windows Azure, service templates in System Center enable application owners to specify similar parameters that enable scaling the application once relevant thresholds are reached.
• #44: We’re going to take a look at how enterprises and service providers can offer a consistent experience in this section—but I want to start with how Windows Azure works. Windows Azure subscribers--let’s call them customers—access the public cloud through a website, known as the management or customer portal. Basically, this portal is the gateway to a wide range of IT services that are delivered on top of the compute, storage, and network resources found in Microsoft datacenters around the world. Now, at each of these datacenters, there are Microsoft IT administrators that manage resources, allocate those resources to the various services being provided, and manage customer subscriptions. In addition, they bill customers for the services consumed. For the customer, everything is taken care of, so they get the services they need almost instantly. For instance, a developer could provision a test environment in minutes—a far shorter time than many face in their enterprise environments.
• #45: Now let’s take a look at how this translates to an enterprise or service provider dataceter. You can see it looks exactly the same. The only difference is that its on-premise, rather than in the cloud. Who are the customers? Well, if you’re a service provider, they’re the customers who pay you to provide IT services. If you’re an enterprise, they’re the employees who consume IT services. Within your datacenter, your administrator performs the exact same functions as in the Microsoft datacenters that Windows Azure uses. He or she configures and defines the resources that support your customers and manages access to services. Admins can also monitor services consumed, so that service providers can price and bill, and enterprises can charge users, departments, or divisions.
• #46: Now let’s take a look at how this translates to an enterprise or service provider dataceter. You can see it looks exactly the same. The only difference is that its on-premise, rather than in the cloud. Who are the customers? Well, if you’re a service provider, they’re the customers who pay you to provide IT services. If you’re an enterprise, they’re the employees who consume IT services. Within your datacenter, your administrator performs the exact same functions as in the Microsoft datacenters that Windows Azure uses. He or she configures and defines the resources that support your customers and manages access to services. Admins can also monitor services consumed, so that service providers can price and bill, and enterprises can charge users, departments, or divisions.
• #47: Let’s take a look at IT administrator tools—the same tools that Microsoft administrators have for Windows Azure. Just as customers get a consistent experience through the Customer Portal, IT administrators get their own dedicated portal too. The admin portal provides a range of management capabilities: Administrators can build resource clouds, which define the compute, storage, network, and other resource clouds that provide the underpinning of each of the customer services. Administrators can automate services by creating powershell ‘runbooks’ that define a range of processes that should execute under specific circumstances. This automation reduces the need for human intervention, which can help drive down the cost of managing the cloud. Administrators can define pricing structures associated with customer services for billing. Administrators get Tenant Management, which are the administrative tasks associated with managing customer user accounts and report against activity. Administrators can establish hosting plans, which are various combinations of services, resource clouds and quotas, which can then be made availabile to specific groups of customers who can subscribe to them. So how do you get from where you are now to this model? After all, it’s a huge shift from traditional provisioning.
• #48: Here’s how: The Windows Azure Pack. Windows Azure Pack for Windows Server is a collection of Windows Azure technologies, available to Microsoft customers at no additional cost for installation into your data center. It runs on top of Windows Server 2012 R2 and System Center 2012 R2 and, through the use of the Windows Azure technologies, enables you to offer a rich, self-service, multi-tenant cloud, consistent with the public Windows Azure experience. That’s the long version. Here’s the short one. WAP is a free download that puts Azure in your datacenter.
• #49: Let’s take a look at each element of the Windows Azure Pack in more detail. <click> The Management portal replicates the Windows Azure Developer portal experience in Windows Azure albeit with a subset of the services available in full Windows Azure. All the capabilities available in the Management portal can be accessed programmatically through the Service Management API, an Odata/Rest API. This allows the portal to be completely replaced, for example if a service provider has their own portal which they want to integrate the Azure services into. By bringing the Windows Azure Service Management Portal experience and API to Windows Server, Microsoft is enabling service providers to administer Web sites and Virtual Machine services on Windows Server while also offering their customers the rich, self-service user experience to provision and manage their Web sites and Virtual Machines. Because this management portal is built on a REST-based Service Management API, these portal experiences are customizable and extensible including possibilities such as partner branding, billing integration and incorporation of incremental solutions and scenarios as well as integrating into existing portals.
• #50: High-density, scalable website hosting services. Simple to deploy and administer. Easily move customers to higher revenue offerings Web Sites is a highly scalable web hosting service for public and private clouds that is optimized for cloud hosting economics and integrated with the world’s most popular OSS Web apps, frameworks and tools. Create high-density, scalable website hosting services that are simple to deploy and administer, operating tens of thousands of sites in a single web farm. Out-of-the-box automation lowers customer onboarding costs while resource metering and throttling can help tailor customer offerings. Supports many frameworks including ASP.NET, Classic ASP, PHP, and Node.js with full Git integration for source code control. Integration of the Web App Gallery allows customers access to popular web applications. Web Sites makes it easy to deploy and administer a scalable website hosting services at high density, similar to the experience on Windows Azure. Out of the box automation lowers customer onboarding costs and streamlines upselling from shared to reserved infrastructure while metering and throttling of resources can help tailor customer offerings. This solution is fully integrated into the Web App Gallery to allow customers to deploy the most popular web applications in seconds without incremental coding. Out of the box integration with many popular tools and frameworks streamlines customer adoption. It can easily operate tens of thousands of sites in a single web farm and scale an individual customer’s footprint with the click of a mouse. Web Sites is an easy to deploy high-density web farm solution that provides a self-service experience to allow service providers’ customers to be in control. This interface reduces onboarding cost and also drives down the potential need for direct interaction which increases operating margin. Taking advantage of new metering and throttling technologies, service providers deploying Web Sites can confidently offer more customized services based on customer need. The scenario also completely automates the transition of a customer from a low margin shared infrastructure to a high margin reserved instance. A single Web Site farm can easily operate tens of thousands of websites. Web Sites offers end customers broad choice in how they build and manage their websites. The solution supports many frameworks including ASP.NET, Classic ASP, PHP and Node.js and allows customers to easily deploy many of the most popular web apps with the click of a mouse. It also allows you to build on Windows, Mac or Linux and publish using the tools of your choice – Visual Studio, Git, FTP or WebMatrix. Web Sites also allows customers to easily scale their number of instances as well as moving from shared to reserved infrastructure with a single mouse click. The Technology The primary design point of Web Sites was that of a cloud service meant to operate at large scale. Now deployable on Windows Server and working in conjunction with IIS, this high density solution is primarily enabled by the Dynamic Windows Process Activation Service which centralizes web farm configuration into a SQL Server database and allows for dynamic site binding and configuration. The solution also incorporates resource metering to allow for incorporation in billing services. Web Sites can also take advantage of enhancements in Windows Server 2012 such as resource throttling which can allow for more fine grain customer offers guaranteeing capacity availability.
• #51: Out-of-box the Windows Azure Pack provides a ready-built Web PaaS and IaaS solution for enterprises to offer internal services of service providers to build customer offerings on. Using the advanced features in Windows Server and System Center you can build the solution on inexpensive, industry standard hardware. Shared mode When a website is first created it runs in Shared mode, meaning that it shares available compute resources with other subscribers that are also running websites in Shared mode. A single instance of a website configured to run in Shared mode will provide somewhat limited performance when compared to other configurations but should still provide sufficient performance to complete development tasks or proof of concept work. If a website that is configured to run in a single instance using Shared mode is put into production, the resources available to the website may prove to be inadequate as the average number of client requests increases over time. Before putting a website into production, estimate the load that the website will be expected to handle and consider scaling up or scaling out the website by changing configuration options available on the website's Scale management page.
• #52: Use the power of System Center and Windows Server to create an Infrastructure as a Service solution for your customers to provision and manage virtual machines. This slide is a good overview of the Virtual Machines service we’re enabling on top of Windows Server and System Center. It really does allow you to deploy an IaaS solution for VM provisioning and management that is completely self-service in nature. In the following slides, we’ll walk through the architecture in more detail and you’ll better understand, from the administrator side, how easy it is to deploy the service and how much flexibility you have in defining and controlling the customer experience. From the customer side, I’ll also show you how easy it is to create and manage these VMs without ever needing to physically contact the service provider. Virtual Machines uses the power of System Center and Windows Server to easily create an Infrastructure as a Service solution for customers to provision and manage Windows Server and Linux VMs, similar to the experience already provided in Windows Azure. This flexible multi-tenant solution puts your customer in control of how they grow their datacenter resources while lowering your support costs. This solution also leverages the new IaaS Gallery to provision server workloads on virtual machines such as SQL Server and SharePoint. It also allows customers to provision self-service networks with their own IP addresses. Virtual Machines is an easy to deploy Infrastructure as a Service solution that puts customers in control of how they extend their datacenter. This self service solution provides a fully automated customer experience to provision and manage virtual machines and server workloads (such as SQL and SharePoint) that can drive down costs associated with on-boarding and servicing of customer accounts. This solution can easily integrate with your current infrastructure such as billing, portals and other services. Virtual Machines offers tenants a very simple self-service experience for requesting, receiving and managing virtual machines and server workloads from their service provider – truly Infrastructure as a Service. This flexible solution provides customers the ability to specify their needs for Windows and Linux virtual machines and networks and truly take control of how they extend their datacenter. The Technology Virtual Machines takes advantage of the new REST oDATA API in System Center 2012 called ‘Service Provider Foundation’. This exposes IaaS in terms of the Virtual Machine Manager component, VM networks, multi-tenancy and quotas, monitoring data from the Operations Manager component and runbooks from the Orchestrator component. This solution also takes advantage of the new VM network technology in Windows Server 2012 which allows customers to create their own private virtual network. Virtual Machines also leverages the service template functionality in System Center 2012 to define and publish fully configured server workloads (such as SQL Server and SharePoint) to the customer portal for purchase and deployment. The solution also utilizes other enhancements in Windows Server 2012 and System Center 2012 such as multi-tenancy and quotas to create custom offerings at higher margins and support billing scenarios. Virtual networks enable the creation of tenant specific IP-address schemes independent of the actual IP infrastructure on which it depends. This means that tenant VMs can be migrated to the cloud without the need to reconfigure IP address schemes or the risk of colliding with other tenants’ IP addresses and security concerns therein. (self service IaaS) (virtual networking), (SC and WS)
• #53: The Windows Azure Pack Virtual Machines Service offers two types of Virtual Machine services to Tenants: Standalone VMs The traditional view of Infrastructure as a Service where a single VM is provisioned against a template selected from a gallery of available images VM Roles Here a template is selected from a gallery of VM role templates configured by the administrator. However, unlike standalone VMs, VM role templates are designed and configured to be able to scale dynamically. Both types of VMs can be attached to pre-configured virtual networks.
• #54: The Windows Azure Pack also enables you to offer your customers multi-tenant access to your existing SQL and MySQL databases for use with their applications.
• #55: Windows Azure Pack delivers IaaS powered by System Center 2012 R2 and Windows Server 2012 R2. The Service Management API integrates with the System Center 2012 R2 Service Provider Foundation API. Service Provider Foundation is provided with Microsoft System Center 2012 Orchestrator, a component of Microsoft System Center 2012 Service Pack 1 (SP1). Service Provider Foundation exposes an extensible Open Data Protocol (OData) API over a Representational State Transfer (REST) web service that interacts with components of System Center. This enables service providers and large enterprise organizations to design and implement multi-tenant self-service portals that integrate IaaS capabilities available through Microsoft System Center 2012 SP1. Feature Description:- VM Management Create, update, and operate VM’s Virtual Machine Roles Deploy, update, and operate VMM services. Scale-out within VM tiers Self-service VM networks Create, update, and use self-service VM networks. Self-service Tenant Administration Tenant admin manages access to hosted IaaS cloud. Tenant admin configures and manages self-service user roles. Enterprise identity for SPF Tenants can authenticate and authorize using corporate identity Extensibility for hosted cloud API Extensible REST OData API surface. Use runbooks to customize processes driven through SPF API.
• #56: This is the Management Portal for Tenants and has a strong consistency with the Windows Azure Developer portal. Tenant users can list items, view their status and provision new items. (Compare to Azure with image)
• #57: Once the VM has been provisioned, the Tenant has a rich set of information to help them manage the environment. They can track usage data as well as do some configuration of the VM based on the permissions of the plan to which they subscribed. And, of course, this dashboard easily provides the ability to stop, start and pause the VM as well as connect remotely to it. (Compare to Azure with image)