IBM Cloud Manager with OpenStack Overview

Editor's Notes

• #3: First a bit of history lesson, most IT infrastructure has been built in support of legacy applications, databases, emails systems, financial systems, customer relationship management, and the dominant approach has been referred to as IT consolidation where many workloads have been combined together on as few large scale-out computers and disk arrays as possible. To make efficient use of these resources, compute hypervisors like VMware, Hyper-V and PowerVM have been introduced, big iron is involved and virtualization is the management focal point. But today new applications are being built, these new applications, mobile, social, analytic, big data, etc, really the cloud applications are being built to leverage a different type of infrastructure. Pioneered first by the mega public cloud providers these infrastructures tend towards clusters of small standard building blocks and as a result can be much less expensive. The virtualization management focal point has given way to a cloud management focal point as the infrastructure is delivered as a service. The dilemma for your customers is that while this new cloud world offers great business flexibility and financial benefits, it also has a completely new look and feel to IT. There is no dominant cloud infrastructure vendor; IT staffs aren’t skilled yet on deploying, managing, and operating these clouds. Innovation is happening at really a break neck speed and there is just a lot of change going on. IT managers are looking for help and this is where our solution is targeted.
• #4: See the trend for both private and public cloud IT infrastructure growth from now through 2017. Clients and service providers have the money and motivation to move rapidly with cloud adoption. Both the worldwide private cloud IT infrastructure forecast at the top and the public cloud IT infrastructure forecast at the bottom demonstrate a tremendous growth opportunity. Two things to take away from this slide are: first that both the opportunity in private and public cloud are significant and IBM needs to be able to execute and obtain more than its fair share in both market segments and for the first time the public cloud IT forecast will slightly outpace the growth in private cloud. So it just underscores the fact that IBM needs to be able to address both our traditional enterprise clients as well as service providers to really take advantage of the opportunity in this market segment.
• #5: 4
• #7: A quick word about OpenStack, IBM’s strong belief in and support of Linux is well documented; we’ve helped progress Linux to the point of today. Almost all of the top 500 fastest computers in the world are running on Linux and over 8,000 developers are actively contributing to the operating system. Our goal with OpenStack as it relates to cloud is similar. To put it simply, IBM’s objective for this layer of cloud technology, this infrastructure-as-a-service layer, is that we want a ubiquitous open source infrastructure as a service layer, period. OpenStack was already heading in that direction when we helped create the OpenStack foundation in 2012. Today innovation on OpenStack is being pressed forward by over 1,000 developers from around the world. Initial OpenStack deployments were being lead by cloud technology companies, ISVs in this space, but the OpenStack community is maturing. IDC predicted that 2013 will be the year that private enterprises will begin in earnest to deploy OpenStack.
• #9: The aim of our Enterprise OpenStack IaaS solution is to accelerate that adoption and to capture more than our fairs hare of the footprints. Now when you are in a conversation with a client it is important to understand what OpenStack is and importantly what it is not. OpenStack is community developed and it is open source software, it is an infrastructure-as-a-service platform, and importantly it is a set of open cloud APIs that when applications are coded to those APIs they become portable across both public and private OpenStack based cloud platforms. And it is really a series of interrelated projects that cover the whole gambit of function required to provide an infrastructure-as-a-service, form compute, storage and networking to a dashboard for operations to image service, telemetry, and orchestration. But it is not a complete infrastructure-as-a-service by itself. Installing OpenStack as the operating system doesn’t give you a complete cloud you have to have some surroundings technologies. OpenStack from the community does not include installers, online update tools, documentation, or support from a trusted vendor. It doesn’t include a virtual infrastructure, something that is quite important to the implementation of a full infrastructure-as-a-service, the virtual networking, storage and compute. As much as we like OpenStack, it’s not perfect, yet, there may be 1,000 developers working with it but it is also designed to be extended by vendors like IBM who offer OpenStack solutions.
• #11: IBM Cloud Manager with OpenStack… Is a complete OpenStack distribution delivered in the same quarter as the community release. 100% of the OpenStack core projects and APIs. Extends OpenStack in areas important to enterprise implementation Heterogeneous cloud management across any x86 environment, IBM Power and IBM System z, and across multiple OpenStack domains including legacy VMware estates; Accelerates time to value with Chef-based installation and configuration; Improves application performance and reduces infrastructure costs by extending the OpenStack Nova Scheduler with runtime policies for ongoing optimization of virtual machine placement; Raises application availability with runtime monitoring and policies for moving virtual machines from at-risk physical servers to other available resources. Simplifies end user self-service with a portal and standardized service catalog built on OpenStack for automated delivery of requested services without IT intervention Integrated Management: Approvals, metering, billing, users and projects through a single ‘pane of glass’ Is supported by IBM. Five full years of support from the date of release with the option to extend for at least an additional three years following end-of-service. Product upgrades—new releases and new versions—at your convenience IBM led the formation of the OpenStack Foundation, is a Platinum Sponsor and Board Member, and is a leading contributor to OpenStack is currently 2rd overall in code contributions and reviews behind Red Hat. Has over 250 employees working on OpenStack and over 500 employees work on open source initiatives relative to Cloud
• #12: So let’s take a quick look at our IBM enterprise IaaS solution, were starting with 100% of the core APIs and projects, that’s important because it helps us to leverage the rapid community innovation and pass that on to our customers as they progress their cloud journey. It’s also important because we are leveraging all of those cloud APIs that I mentioned from OpenStack, which means that our clients as they build applications to those APIs, those applications will be portable across public and private OpenStack based clouds. We are also extending the community OpenStack in some important ways. One of those ways is in a resource aware scheduler that we will talk about in more detail in a moment. This is a significant technology that IBM has that provides a much improved flexibility as well as a much improved cost benefits to our clients. We also include more enterprise cloud database for storing all the infrastructure the OpenStack infrastructure components we include DB2 in our solutions. Multi-cloud federation becomes important when it comes to hybrid configurations and we also include some drivers for enterprise class infrastructure that may not have made all the way to the community distributions of the code yet, as I mentioned OpenStack by itself is not a cloud, it requires a virtual infrastructure. Many of our clients have already made decisions on virtual compute, on the hypervisor they are using, but interestingly most of those decisions have been made for legacy infrastructures as we talked about earlier. What we are finding is as clients move towards a more cloud infrastructure they are increasingly choosing KVM as their hypervisor environment, KVM has great integration with OpenStack. We are also augmenting the environment with important virtual storage and virtual networking capabilities to complete the virtual infrastructure the operating system needs to operate as a full on infrastructure-as-a-service. And finally we are surrounding the software components that I mentioned so far with automation and tooling that to help IT administration provision and manage the whole solution. That is our enterprise OpenStack IaaS solution.
• #13: Let’s take quick a look at some of the specific value differentiations in our solution that you will need to point out to our clients. The first thing that an IT manager is going to need to have done for him is to get this infrastructure-as-a-service software environment build up on his cluster of scale-out common building block hardware. Since this is all new to IT, most clients have a big concern over the training requirements to get this done reliably. That’s why we are including IBM differentiated automation for provisioning and management in our solution. The automation takes care of provisioning the IaaS environment onto bare metal and then monitors performance and utilization taking automatic actions to ensure service levels are met.
• #14: Another important IBM differentiated extension is in the OpenStack resource scheduler. The OpenStack resource scheduler is responsible for making sure virtual machines are efficiently placed in the infrastructure. The community scheduler is a fairly static implementation, sort of a once placed always placed implementation. We dramatically extended the capabilities of the scheduler giving our clients much more flexibility and higher quality of service and improved efficiency. Our scheduler extension performs initial placement AND ongoing monitoring and optimizations
• #15: Our resource scheduler extensions offer a wide variety of policies, for example, maybe a packing policy might be appropriate for the development phase of a application when cost efficiency is the primary goal, and a stripping or load balancing policy might be appropriate when that application is moved into production for more performance optimizations. Additionally our resource scheduler extensions are dynamic; they cover both initial placement and ongoing monitoring and optimization as the needs of the environment change. More detail, if needed Our scheduler extension performs initial placement AND ongoing monitoring and optimizations. Takes advantage of the underlying hypervisor and live migrates VMs. The community scheduler only does initial VM placement. Once the VM is placed there is no more interaction with the nova-scheduler: Place and forget it. Our scheduler enables all metrics to be available to all policies. The community scheduler filters only handle the specific metrics that have been coded into the filter logic. Our scheduler enables any metric to be used in any policy. To add a new metric with the community scheduler, filter logic needs to be written or updated (i.e., coded) to handle the metric. Our scheduler enables policies to be global or specific to an individual request. With the community scheduler, filters are global, they apply to all requests. Our scheduler allows new policies and metrics to be added without the need to restart the services. When new policies and metrics are added to the community scheduler all the services need to be restarted.
• #16: Another IBM differentiation is in the virtual storage layer, as physical compute is flexed up and down, as the VMs are migrated per policy, it is important that data access is maintained. As the big data used in analytics applications grows it’s important that there is enough high performance storage capacity, and for many of these clouds that are using internal storage for common building blocks it is important to construct enterprise class storage services from these common disk capacity building blocks, this is what we have added to our solution and you don’t find that in most any other competitive alternative. Additional GPFS/Elastic Storage details if needed The IBM GPFS Cloud Vision GPFS as a single scale-out data plane for the entire data center Unifies VM images, block devices, objects, and files Single name space no matter where data resides De-clustered parity - GPFS Native RAID (GNR) Data in best location, on the best tier (performance & cost), at the right time All in software
• #17: So that’s an overview, a quick overview of what we are trying to accomplish with this enterprise OpenStack IaaS solution, our goal is the help clients deploy this type of IaaS environment automatically, on standard hardware building blocks and with a minimum of skilled IT staff resources. It’s also to help them manage their cloud infrastructure efficiently and cost effectively with the differentiated scheduler that’s continuously reevaluating the virtual infrastructure and making adjustments so usage is maximized. And it’s to help them build new cloud services on a platform that is open and portable. The unaltered and fully supported cloud APIs make this possible.
• #18: Cloud Manager with OpenStack was formally announced in May 2014. It delivers rapid time to value for clients wanting to move beyond basic virtualization. It has been in the marketplace for over 3 years and prior to this point it was offered as IBM Smart Cloud Entry. IBM Cloud Manager with Open Stack is an easy to deploy, simple to use cloud management software offerings based on OpenStack, with open APIs. Importantly we include IBM enhancements that feature a scheduler extension providing dynamic resource management, a self-service portal for workload provisioning, virtual image management and monitoring. It's an innovative, cost-effective approach for clients that want to get started with cloud and/or clients that want to get started with an OpenStack-based solution. It also includes automation, metering, and basic security. New for the 4.2 release are advanced hybrid capabilities where two CMO-based clouds can be connected to form a single multi-region cloud so that they appear to be a seamless extension of each other. So you can feel confident engaging with your clients because this is a market tested and validated solution that isn't brand-new to the marketplace today. It's continuing to build on its foundation, and it supports heterogeneous compute, storage and network, that means it works across IBM's entire family of compute infrastructure from the basic System x up to Pure Systems, Power, and now System z. It supports all the major hypervisor's and it provides additional features that clients are looking for as they look to implement their cloud strategy. It also provides a hybrid enablement and multi-cloud federation to use multiple instances of OpenStack and hybrid clouds, hybrid clouds on and off premise and on SoftLayer. As we continue to refine and evolve this offering, we will continue to try and keep it focused on its core target market for clients that are looking to implement a basic cloud solution by keeping it open, simple and innovative.
• #19: IBM Cloud Orchestrator enables Infrastructure, Platform & advanced Orchestration Services: Eases coordination of complex tasks and worklflows, necessary to deploy applications Deploy application topologies or patterns Take advantage of the huge pattern library in the IBM PureSystems Center IBM Cloud Manager with OpenStack enables basic Infrastructure Cloud Services: Cloud provisioning and automation based on OpenStack Simplified implementation, lifecycle management, resource management, self-service portal, monitoring & metering Full access to OpenStack APIs – All IBM server architectures and major hypervisors now available to choose from Integrated platform management, backed by IBM enterprise-grade lab services and support {TRANSCRIPT} So as far as positioning cloud management solutions from IBM, we like to talk about modular capabilities with a common cloud management services layer. So IBM Cloud Manager with Open Stack performs the basic infrastructure cloud services that we’ve just reviewed and it would be a great place for many of our clients to get started with cloud. But some clients may be looking for some more advanced cloud features right out of the gate and if that’s the case then IBM Cloud Orchestrator may be a better fit. Orchestrator provides the capability of facilitating more complex workloads and other tests necessary to implement pattern solutions and take advantage of the huge library in the IBM Pure System Center associated with deploying more advanced cloud capabilities.