Spider & F5 Round Table - The Flexible Data Center

Editor's Notes

1. Today we are going to talking BIG-IP LTM 11.3 Updated, the Solar Release. We are going to be primarily talking about our Secure Application Delivery solution, which deals with our SSL solutions. As you probably know we are releasing new hardware with new capacity and new performance on the BIG-IP hardware and these Secure Application Delivery solutions will drive addition workloads on to that new platform to help drive additional sales and larger sales with in your hardware solutions. So let’s go ahead and get started.
2. But unfortunately the infrastructure will very quickly outgrow the standard firewall and ADC solutions. As new applications come under IT management, such as SaaS and cloud computing – technologies that were designed to offer services directly to users and not be managed on-premise – the challenges of user management, application performance and availability, and security don’t go away. In fact, SaaS and cloud computing bring new challenges to IT: How do you apply application access and performance policies to applications that are outside of the data center? Firewalls and simple ADC’s don’t solve the problem of managing more endpoints, more delivery models, and more applications in more locations.
3. In order to address changes impacting customers, a new, more sophisticated model needs to replace the standard firewalls and simple ADC’s: an Intelligent Services Platform is the key to managing application delivery in the new world of on- and off-premise applications accessed by users with multiple devices. An Intelligent Services Platform manages users from any location on any device, applies application delivery policies to all application requests in both directions, and connects them to applications regardless of where the application lives. There is no way to manage mobile device and application growth simultaneously without an Intelligent Services Platform acting as a full proxy between the users and applications.
4. With co-existing IPv4 and IPv6 networks, add to this the complexity added to DNS to handle it. Then bring in remote networks with overlapping address ranges. And we’ve only talked about IP so far. Migrating from native HTTP to SPDY will require them to co-exist. Then there’s the fact that delivering the same application via a mobile provider, 3G, 4G, LTE, requires a different delivery strategy than for desktops/laptops because of the nature of these protocols.
5. One size fits all application delivery architectures don’t scale. The network needs intelligence in both understanding the different types of behaviours that are required by various devices and networks and when to apply them for the best performance and service resilience.
6. Consider a transport company contracted to takes boxes from one location, lets say New York, to another location, this could be Singapore. The checks and processes in place might ensure almost guaranteed success in delivery of these boxes but what if the boxes contained the wrong goods. Now apply this same logic to Application Delivery. If network has no concept of services and their consumers, how they differentiate in behaviour and requirements, then how will it know whether its delivering errors or the intended data? How will it know if its delivering a 404 Page Not Found or a 503 Gateway Timeout? It wouldn’t. it would deliver these errors as fast as efficiently as it could.F5’s Intelligent Services Platform does differentiate between right and wrong. It adhere’s to the business values of brand protection and takes its instruction from the business owner. F5’ s Intelligent Services Platform:Monitors and analyses application and network healthAchieves real time application availability and fail-overDelivers application visibility and scale for capacity and troubleshootingStats that are availableServer LatencyPage Load TimeThroughputResponse CodesMethodsURLsClient IPsClient GeographicUser AgentUser SessionsCPUMemoryDNS
7. Route users to the best location based on application availability and performance Route user to most available locationMonitor application services globallyDeploy automatic disaster fail-over with minimal downtimeStats that are availableServer LatencyPage Load TimeThroughputResponse CodesMethodsURLsClient IPsClient GeographicUser AgentUser SessionsCPUMemoryDNS
8. Message data is what tells us the ‘who and where’ a transaction is coming from. It also tells us where it thinks it should go. F5’s Intelligent Services Platform taps into these messages bringing flexibility and control. A single namespace like xyz.com can be presented to the consumer where the F5 Big-IP understands that xyz.com/b2b might be a completely different resource, potentially in a different physical location. What is important here is that the complexity of everything ‘behind the scenes’ is abstracted from the consumer. F5 optimizes traffic management decisions based on contextual message data to: Translate between multiple application protocolsAvoid rewriting applicationsSupport for 3600 application protocols natively
9. In 2011, public and private cloud drive 15% of IT spending. Worldwide Private Cloud deployments are projected to increase to over 1.1 billion by 2014 So cloud adoption is very real. But with it comes some problems. And many of these problems come from trying to achieve all of the dynamic greatness that Cloud should deliver. According to NIST (National Institute of Standards and Technology) these are “essential characteristics “of Cloud:On-demand self-serviceBroad network accessResource poolingRapid elasticityMeasured service So, how do we deliver on them?
10. Agility comes from Programmability and Extensibility in the network. And from this we get a scalable infrastructure, application mobility and Software-enabled services.
11. The ability to scale infrastructure on-demand allows for the freeing up of resources for use by other services. In short, you can dynamically spec the required resources based on real time demand. Versus spec’ing resources based on worst-case requirements. The latter, a model that brings with it significant waste. The key to success is the intelligence in measuring when to scale safely. Many organizations still use server CPU & RAM utilization as a gauge of application performance. F5’s integrated Intelligent Services Platform measures the End user experience. This information is passed to the cloud infrastructure platforms resulting in end-user experience driven data center workflows. Integrating F5 with leading cloud infrastructure platforms enables you to: Employ automatic provisioningScale on demandIntegrate with management platforms
12. Scale out to the cloud as demand for application resources spike.  Moving to a Cloud model should deliver simplification and ease of use over what we’ve experienced in the past. Implementing Cloud successfully is more than Pay-as-you-Go hypervisor rental w/ some management tools. It should provide an extension to your existing data centre’s. An increase in commodity resources when needed. And this model is very attractive to organisation that want to increase datacentre utilisation. The adoption of virtualisation took many from averages of sub-10% utilisation to more cost effective values ranging from anything like 30 – 50% at peak time. Sweating the infrastructure further requires workload agility like Cloud Bursting. The ability for the infrastructure to see when resource consumption has reached a level whereby the end user experience is beginning to see impact. Data centre agility, by means of an integrated Intelligent Services Platform, allows for the private datacentre to dynamically extend its workloads into a cloud provider to consume extra resources. And, equally, to retract workloads from cloud providers when workloads decrease.
13. F5’s Intelligent Services Platform delivers the ability to route user requests to applications or workloads that reside in multiple clouds. This with the workload visibility and agility of a private datacenter deployment delivering: workloads to the nearest cloud provider for better performancegranular user control and management to remote resources no matter their location, or xaaS provideradopt cloud effectively without adding risk
14. Many have invested in server virtualisation without realising the returns they had expected. As often with new technologies come new problems:Virtual machine creep - VM’s created, powered on and forgotten about.Shared architecture adding complexity to security – virtual and physical devices co-existing.Lack of network agility/integration stalling what should be faster deployment times
15. To reach the full potential of server and desktop virtualisation, the network must deliver consistency, programmability and extensibility. Furthermore, offloading services like encryption, compression and caching to improve Virtual Machine density.
16. SDN will greatly improve the management of Layer2/3 solutions – switching/routing/network firewalls. But it falls short on the when & why to implement a workflow. The DC infrastructure is built to host and delivers services to their respective consumers. Consequently, its only with knowledge of and integration into these services that SDN can deliver value. The service visibility derived from F5’s Integrates Services Platform is key to: Bring application awareness to an SDN Infrastructureflexibility to add and extend new SDN servicesIntegrate with existing SDN solutions from VMware and Microsoft. 
17. There’s a better way – the Application Delivery Controller. It’s a platform designed specifically to solve these issues, by making decisions in your network to ensure that your applications are always fast, secure, and available. BIG-IP Local Traffic Manager is the market-leading application delivery controller. It load balances, secures, and optimizes application traffic, giving you the control to add servers easily, eliminate downtime, improve application performance, and meet your security requirements.
18. There’s a better way – the Application Delivery Controller. It’s a platform designed specifically to solve these issues, by making decisions in your network to ensure that your applications are always fast, secure, and available. BIG-IP Local Traffic Manager is the market-leading application delivery controller. It load balances, secures, and optimizes application traffic, giving you the control to add servers easily, eliminate downtime, improve application performance, and meet your security requirements.
19. F5 Virtual Edition 11.3 now has unmatched flexibility and performance across all public/private cloud virtualization platformsNew HyperV, KVM and AWS supportHigher performance for Xen and Vmware3G performance for VMware, HyperV and KVM Supported Hypervisor environments:VMware vSphereESXiHypervisor 4.0, 4.1, 5.0 (Update 1), 5.1 and vCloud Director 1.5Citrix XenServer 5.6 and 6.0Microsoft Hyper-V for Windows Server 2008 R2 and 2012KVM – Linux Kernel 2.6.32 (RHEL 6.2/6.3, CentOS 6.2/6.3)AWS (EC2) - CentOS 5.6 userland NOTE: Following updates added to VE 11.4:ESXi 5.0 Update 2