When administrators try to build clouds using legacy infrastructure, they run into operational complexities created by disconnected technology silos never built to work together seamlessly. Physical infrastructure was traditionally static and difficult to manage. It was built assuming frequent manual intervention. Virtualization, on the other hand, is inherently programmable. The problem comes in layering flexible virtualization software on static infrastructure. This boundary is frought with complexity and expense in setup, configuration, and ongoing management. <click - animation>I call this an accidential architecture. No one wanted things to be this way, but infrastructure was designed largely before virtualization and has retained essentially the same overall design. Until the UCS, of course, which was designed explicitly to bridge the gap between physical and virtual by providing pre-integrated, programmable infrastructure.In fact, UCS was designed by one of the original VMware founders, Ed Bugnion. Ed wrote the original ESX hypervisor and understood as well as anyone, the difficulties in managing infrastructure dependencies in a virtualized environment. As good as his software was, it relied on extensive and costly manual configuration of compute, network, and storage to run applications efficiently in high quality virtual machines. By integrating the physical compute and network layers with hypervisors and their virtual machines, Ed was able to overcome many of the traditional bottlenecks in data center and cloud systems. Beyond performance, which could be considered a capex issue – do I have enough hardware? – the UCS improves operational processes by enforcing and managing the physical/virtual boundaries.
The market has spoken clearly: Cisco UCS is the #2 x86 blade manufacturer in North America and #3 globally. With a run-rate over $1Billion, UCS has sustained market traction and deployment across the world’s leading companies and institutions. Customers including Qualcomm, Freeport McMoran, and Levi Strauss run business-critical workloads on UCS including SAP and Oracle business applications. Travelport, Moses Cone Health System, and Boston University are examples of customers who have utilized UCS rapid deployment and configuration capabilities to more rapidly respond to their application requirements with low operational costs.Qualcomm:http://www.slideshare.net/Ciscodatacenter/why-cisco-ucs-for-oracle-solutions, http://www.cisco.com/en/US/prod/collateral/switches/ps9441/ps9670/at_a_glance_c45-617345.pdfFreeport:http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns944/dcs_uc_for_erp_freeport.pdfLevi:http://www.sap.com/about-sap/newsroom/press.epx?pressid=13294Travelport:http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns944/casestudy_c36_674549.pdfMoses Cone:http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns944/case_study_c36-588298.pdfBoston University: http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/bumc_cs.pdf
If we can say one thing about cloud services – it is that they are growing, resulting in ever increasing workload on system administrators. UCS is a stateless system that leverages the power of dynamic configuration to rapidly deploy compute nodes in whatever configuration is best suited to the applications at hand.UCS understands the concept of infrastructure templates and enables administrators to automatically provision dozens of physical servers at a time with policy-enforced configurations at the compute and network layers. Whenever the administrator wants to expand or create a hypervisor or application cluster, it is only necessary to specify the number of nodes in UCS Manager. UCS will automatically deploy the requested server and network configuration and then boot those nodes with the desired system image. The result is a dramatic savings in time and reduced potential for manual errors.Key points for cloud (beyond the normal UCS service profile value proposition)On-demand physical provisioning like virtualDeliver physical resources programmatically like a VMAutomate physical provisioning like software
Key points for cloud (beyond the normal form-factor independent value proposition)Any physical resource can be delivered on-demand: local storage, PCIe cards, more memory, etc.Treat all types of x86 hardware (blades, racks, PCIe, storage) as pools of resources
Nothing here about virtual machines, linux or Windows, object stores or SANsThis looks like how IT resources are delivered as compared to what those resources actually areSounds a lot like traditional IT– you need resources? let’s consider what you needThere is a big difference between the infrastructure that runs your ERP system versus your public-facing web site
If you have a purely stateless application with minimal performance requirements, you can run almost anywhere; not so if your analytics farm requires low latency network and storage
Using the Cisco Unified Computing System as the basis for your virtual desktop infrastructure unites computing, networking, virtualization, and storage access using technology innovation to bring ideal platform for desktop virtualizationSpecifically, the Cisco UCS costs at least 20% less than competitors due to its radically simplified architecture requiring fewer components .Supports 60% more virtual desktops without impacting user performanceIs easy to deploy and scaleMassive Scalability – Full system scales to 320 servers x 120 virtual desktops/server = up to 38,400 desktopsLargest and most cost-effective memory footprint for a 2-socket server combined with screaming fast, secured I/O avoiding bottlenecks
Another unique benefit that we can offer is the ability to manage existing/new “enterprise applications” and emerging “big data applications” from the same management domain.This is enabled by the unique capability of Cisco UCS to transparently integrate and manage infrastructure for running traditional enterprise applications like Oracle RAC on Cisco UCS B-Series Blade Servers (typically deployed with enterprise SAN storage arrays) with infrastructure for deploying Big Data on Cisco UCS C-Series Rack Servers (with internal storage capacity and bandwidth). Integrated Server ManagementIntegrated Network ManagementIntegrated Data Management
UCS is Cisco’s solution for virtualized data centers. Basically it is an aggregation of Network, Physical and Virtual System and Storage technologies all in one appliance – so everything runs on one appliance.CLICK: You migrate your tier 1 apps to UCS and deliver and manage them more cost-effectively through virtual desktop services. Centralization & virtualization can prove workforce productivity and IT costs – if you have the right management solution!Just as the UCS is a converged or unified infrastructure, you need a unified approach to managing UCS to assure ensure exceptional Quality of Experience for the applications UCS delivers and Quality of Service of the underlying UCS infrastructure. Cisco is an expert at managing service layers 1-3, but we are experts at managing layer 1-7 and also provide automation and more.This requires what we call a CA Service Assurance solution approach to management: End-to-End and Top-to-Bottom visibility and control to End-to-End:from the user all the way through the entire infrastructure and backTop-to-Bottom: from the app all way the to packet on the networkCLICK: First we discover the environment. We connect to the Cisco UCS Manager’s XML API – and we discover the UCS’ components and their interconnectivities (relationships and dependencies)The benefit: detect fault quickly and suppress symptomatic alarms.CLICK:Next we do performance and capacity monitoring of all the components within the USC, traffic between them, and mananeg applications hosted there and the business transactions they drive.The benefit: proactively address infrastructure issues in the USC and the rest of the infrastructure.CLICK: Finally rather than looking at things only from an infrastructure fault and performance perspectives, we have a business service dashboard that shows you how the UCS (and infrastructure around it) impacts that applications that the UCS is hosting to deliver business services.The benefit: manage service quality , risk and availability and SLAs
Mention the Monitor portions and the Service Desk portion too – unified Management in an ITIL focussed ITSM packageDiscovery – important in terms of context – what VMs are running on what blade for instance?