Breaking the Kubernetes Kill Chain: Host Path Mount
OpenStack Opportunity - Citrix
1. Delivering on the OpenStack Opportunity Gordon MangioneVP Cloud | Citrix Systems
2. “We have proven… polymorphism will … defeat any model based detection method. Identification of bounded length Metamorphic viruses is NP-complete” Song et al., 2010 “If the workload CAGR remains steady, server volumes will decline in 2011, and we won’t see 2010’s volumes again in this decade.” Tom Bittman, Gartner Fellow, Feb 2011 “Flash-based DIMMs open another front in the flash war on older storage technologies A 48 GB SO-DIMM should soon be less than $200.” ZDNet, 10/2010 “The implicit change to networking is that the last-hop switch is a feature of the hypervisor or server hardware and not a switch in the physical network” ACM Queue, 2010
16. An Open Cloud Alternative Enterprise Customers: “Let me easily adopt, change and move” Cloud Providers: “Let me differentiate by adding my own IP” Everyone: “No single-vendor lock-in” “Freedom to choose best products to deliver Cloud Economics”
17. Open Large Scale Restful The Open Way to Build Clouds OpenStack Compute OpenStackObject Store OpenStackNetworking openstack
PunchLine: OpenStack has made good progress on compute and storage. There is significant opportunity to advance the science of openstack networking, and to differentiate and monetize our offerings on the basis of it.OpenStack Compute: A distributed compute platform, written from scratch for OpenStack.OpenStack Object Storage: A distributed, replicated, self-healing storage platform. This is the production-quality code used for Rackspace Cloud Files (4PB under management today). -- unspoken notes --OpenStack Compute consists of many components that are distributed around the datacentre: nova-api, nova-compute, nova-scheduler, nova-network, nova-volume. Connections to the hypervisor are managed through nova-compute.OpenStack Object Storage also consists of distributed components: swift-proxy, swift-auth, swift-account, swift-container, and swift-object.
And that’s exactly what our OpenCloud products do for you… (talk through each)… then go to demo
PunchLine:Lots of interesting work going on inside OpenStack – some new ground being covered on both scale and openness. This is the kind of thing that architects and engineers love!This is a zoomed-out view of very high-level OS architecture. The things to note:The system is designed to scale horizontally – you increase the controller capacity, simply by adding more machines to the control farm.When you talk about managing 100K machines, traditional designs, like some of what VMW or a number of startups are doing, start running into limitations. A key difference with OS is that we use asynchronous messaging with a distributed message queue.Design for openness and pluggable-everything is another key consideration – in the design of APIs and coupling.
Transition from previous slide …So to that end, we believe that there three “E”s to cloud.Economics, Elasticity and Pay-as-you-goWe are all aware of the two fundamental “E”s of cloud :Economics .. and ElasticityNo architectures are “true cloud” without natively aligning to these core value proposition.However, to truly create a “sustainable” plan for Cloud, we need to chart a pragmatic path forward that aligns with the specific the applications and services that move to the Cloud