Cisco sees cloud adoption as a journeyThe journey starts with increasingly capable pubic cloud services and private cloud platforms for the enterpriseFrom there, new APIs and protocols, virtual private clouds services, hybrid cloud capabilities and open standards for cloud will appear and evolve over timeThe eventual goal, from Cisco’s perspective, is an open, public cloud networking infrastructure, which we call the Inter-Cloud
If we’ve described the problem in the current economics of the data center, what does the future of IT look like? For many, it is best represented by the world of many clouds. This is what will allow IT to deliver services back to the business as opposed to being considered a cost center. If IT can make the shift towards more of these innovative projects, they stand a much better change of transitioning to IT as a Service and cloud. By harnessing the power of public, private, hybrid clouds or traditional data centers, they can more rapidly deliver value to the business. Although many customers will start with private clouds, which deliver pivotal functionality to they will also want to choose from offerings available to them via the public cloud. This has the benefit of eliminating the need for shadow IT that so many groups end up creating. It may not surprise you to realize that service providers around the world are launching their cloud services based on a Cisco Unified Data Center.
From the standpoint of internal IT, your business users demand and expect faster, better, more agile and dynamic IT services. But many legacy data centers are built on old technology that just wasn’t designed for the demands and pace of today’s business. The infrastructure is complex, it’s messy, and it’s ill-suited for the future of IT.And from a management perspective, most IT organizations have dozens of different systems that don’t necessarily work all that well together. If you’ve ever played Jenga, you know what happens when you try to remove the wrong block too quickly, it falls apart. That complexity and inherent fragility applies to most legacy existing IT systems – they are very expensive to manage and maintain, and many IT organizations are ill-equipped to meet the demands of the business for faster service delivery.For example, in most IT departments, the process for data center infrastructure service requests is slow, complex and expensive. Each request, like a request to host an application whether it’s for dev or testing or production, is treated as essentially a new project with new requirements, architecture review, sizing, etc., even if that same or similar request has been made dozens of times. The resulting process is manual, cumbersome, inconsistent, and often takes several weeks for the end-to-end delivery of the requested service.By comparison, the operating model for cloud computing and infrastructure-as-a-service (or IT-as-a-Service in general) is one that requires on-demand provisioning with standardized options and processes, and rapid elasticity with a shared pool of resources. The graphic here on the right simple highlights those key attributes for cloud computing from the standard NIST definition (the National Institute of Standards & Technology).The cloud operating model means faster, simpler, more flexible and more cost-effective IT. This is what’s so appealing about public cloud computing options like Amazon Web Services. And what we’ve found is that IT organizations can deploy this operating model in a private cloud (and pave the way for a hybrid cloud model), but it requires a new approach to management and data center infrastructure.
This diagram depicts the key elements from a management standpoint that are required for IT-as-a-Service, regardless of the business applications and IT services delivered – whether it’s for a dev and test environment for your SAP applications team or a production environment for Oracle database hosting. This framework is similar to the architecture for private cloud recommended by analysts firms like Gartner and Forrester. We’re not trying to replicate all of the existing IT management systems (like your existing service desk / ticketing systems and CMDB) that you use to run your legacy data center environments. Instead, this diagram represents the new capabilities necessary for IT-as-a-Service; the mandatory requirements for this new approach include a self-service portal and orchestration, together with policy-based infrastructure resource management.At the top level of the diagram you have the self-service portal, with on-demand provisioning from a catalog of standardized IT options, governance and approvals, as well as tracking the lifecycle of service usage to prevent sprawl and to enable chargeback or showback. This portal can provide users with a unified online “menu” of options for requesting IT services, whether the infrastructure resources are hosted in your own data centers or potentially sourced externally in a hybrid cloud model.From an automation and integration standpoint, you need to combine the portal with an orchestration engine that can provision the requested service and the underlying infrastructure – with policy-based infrastructure resource management and controls across a shared pool of compute, storage, and network resources, whether physical or virtual.And finally, although IT-as-a-Service demands a new approach, it must complement your legacy systems and management tools. So this new management approach needs to integrate with the existing IT environment for operational processes including monitoring and service assurance, configuration management and a CMDB, as well as business processes like user management in your directory systems and financial management - whether you start with a showback model or evolve to pay-per-use billing and chargeback.
The impact for our customers is taking what was very long and complex legacy management process – that can take weeks for end-to-end, often manual delivery of a request, from the upfront architecture and design of the system to the deployment – to minutes in an automated self-service provisioning environment.And this is exactly what we did at Cisco IT, which deployed Cisco UCS and Nexus – together with Cisco Intelligent Automation for their private cloud.
Accelerated Cloud DeploymentDeployment readyPre-integrated portal, orchestration, and OS/ESXi provisioning softwareOut-of-the-box content: pre-built, fully supported portal and orchestration contentPartner-enabled services; install and configure within 2 weeksCompute as a Service on UCSReduce provisioning time from weeks to minutesEnd-to-end request to fully automated provision/deprovision/lifecycle managementGet more applications into cloud through business-friendly IT storefrontEnd users have full lifecycle management of their serversPhysical and Virtual provisioning and automation for CloudStrong integration with Cisco UCS Manager API and service profiles enables repurposing blades as needed very quicklyIntegration with vCenterEnd-user and administrator workspacesResource pools and policy management.Strong IT control over resources and consumptionCustomizable lease policiesStrong consumption limitsRole-based access controlGet most out of your infrastructure by using superior VM density of Cisco UCSAbility to grow your automation and cloud use cases and coverageExtensible platform for adding new services (hybrid cloud, composite applications, etc.)Integration into existing IT operational management toolsMultivendor virtualization and infrastructure capable through partner and Cisco servicesClear upgrade path to Enterprise Edition and VMDC support when needed
Bringing the conversation back to the crawl walk run sprint model, this is our recommended pathway for a customer deploying clouds along a safe and sane roadmap. Starting with a basic cloud and moving forward will ensure a successful journey.