First, a quick side not on “What is cloud computing?” There is so much hype in the industry it is worth just defining what I mean by this. First, it is about compute (virtual machines), storage, and networks. These are the resources that make up dev, integration, test, staging, and production environments. Second, they are delivered as a service. You may have heard of Infrastructure as a Service. This is essentially the heart of cloud computing. The service is available from both web interfaces and APIs in a self service manner. Cloud computing is characterized by the ability to provision apparently infinite resources in minutes vs. hours. The reverse is true, resources can be returned to the pool in minutes so that they can be repurposed. Finally, the business model is pay for use, operating expense vs. a purchase/capital expense where you pay for the resources regardless whether they are being used or not.
Slide Purpose/Objective: Define what is a dynamic workload, reinforce leading Skytap Cloud use cases, and tee up the Skytap differentiators slide Talking Points: Now that I’ve introduced you to the Skytap Cloud, which was designed to make the process of supporting dynamic workloads fast, easy and secure, let’s spend a for moments talking about why the cloud is ideal for your dynamic workloads . First, what are the common characteristics of a dynamic workload? Dynamic workloads often present businesses a range of challenges making them difficult to manage and maintain in an efficient, cost-effective fashion. By their very nature, dynamic workloads are difficult to resource since they have unpredictable capacity needs due to fluctuating demand which may require increased scale at a moments notice. For example, a product development process enters into the latter stages of performance testing or an IT training class for your new product offering is exceeding registration numbers. In both these examples, IT would either need to have spare capacity handy and idle or would need to rapidly procure, prepare and provision new equipment to meet these demands. In today’s volatile economy, few businesses are willing to plan and purchase IT resources based on a forecasted “high-water mark” of demand, only to have these equipment sit underutilized. Requires frequent changes to applications and underlying infrastructure , as experience with the growing adoption of Agile development processes, is a second characteristic of a dynamic workloads. Users, such as a developer, can be frustrated having to wait for operations to “load up” the next build of their project. This not only results in stalled progress, but could increase the risk to quality if adequate testing cycles are stymied by slow, manual processes. Adding to this, it is not uncommon—in cases like product demos or virtual training—where entire multi-machine environments need to be repeatedly setup and torn down at a rapid pace. Continuing on the theme, dynamic workloads are ones that experience high demand for sharing and collaboration . Whether it is sharing between in house development teams and contracted test resource or enabling a customer to interact with the proof of concept environment you’ve built out to demonstrate your product’s fit, trying to securely yet easily collaborate with multiple parties is no small task, involving opening up access over the Internet amongst other requirements. Lastly, all of these conditions mean one thing: higher costs . Whether your users need to collaborate with partners across the office or around the world, need complex virtual data centers “spun up” to triage a newly discover product bug, or your new product is driving increased demand for demos, you’ll find yourself spending a lot of time and resources just to keep up. <CLICK TO ADVANCE SLIDE ANIMATION> And, it is not uncommon for businesses to attempt to build out virtual lab infrastructure on-premises in an effort to address these ranging demands dynamic workloads can place on IT. However, there are a number of challenges and risks associated with this approach: It can been tough to keep up with the often times breakneck speed and user demand for fast turn around on infrastructure requests. Being too slow risks application delivery and once you’ve setup an environment, it’s very likely changes will be requested. And, while you may struggle with keeping pace, users will often seek ways to share these virtual environments with collaboration partners within your company and outside. Do you have the mechanisms in place to address these needs in a secure, scalable and timely fashion? Again, costs can add up quickly and these are many of the reasons customers, like yourself, have turned to cloud-based computing solutions to handle these dynamic workloads. <CLICK TO ADVANCE SLIDE ANIMATION> This is why Skytap focuses on dynamic workloads which are ideal candidates for cloud computing. Great examples of these, and why many leading companies like Symantec, TrainSignal, BinaryTree and CommVault have turned to Skytap Cloud, include: Development and testing, especially those who are embracing Agile development Software demonstrations and product evaluations (or proof of concepts) as part of a sales cycle And, after a sale is closed, training your new customers on these technologies (or even your own teams) By now, I suspect you can clearly see why the cloud is perfect for these types of dynamic workloads
Slide Purpose/Objective: Clearly articulate the business value (core value proposition) Skytap Cloud delivers to customers Talking Points: Skytap Cloud provides companies, like yours, with a fast , easy and secure way to use cloud computing to accelerate your business. By mixing the best of what cloud computing can offer with the advanced automation and self-service capabilities of the Skytap Cloud Intelligent Automation Platform , you will experience: Increased business agility to keep pace with the ever changing dynamics of today’s markets, stay ahead of competitors, and support bursts in demand Reduced time to market thanks to the flexibility to quickly meet the often fluctuating needs and frequently changing requirements of your users, especially for those who have embraced methodologies like Agile development The ability to ship better products and do to so more quickly because you have an elastic sandbox to test your new products rapidly, at scale and across multiple platforms and configurations. Find a bug? Simply snapshot and share the environment to give developers an easy way to triage and get to the root cause without slowing down productivity A boost to productivity across teams whether through secure collaboration or training your sales team on the latest functionality, all from the convenience of an intuitive, browser-based interface. And, lowered costs by not having to worry about having enough capacity on hand (which could be underutilized due to fluctuating demand) or the having to spend extra cycles managing it to keep pace with requests
To set the context, describe a traditional software development lifecycle. This is called the “waterfall” method of development as you can see by how it is typically diagramed, as the water runs down the phases. It is characterized by defining requirements up front so you know what you are building, then doing the development, then testing, etc. Typically long development cycles sometimes years. Infrastructure is usually provisioned up front and rarely changes throughout the process. Very hard to react to changing requirements and often projects complete that are way of the mark originally set many month ago.
Agile methods on the other hand are characterized by short release cycles focused on solving specific customers problems and delivering software at each iteration. The short cycles (typically weeks) allow teams to quickly react to changing requirements. The focus on delivering smaller, more incremental changes to customers allows teams to get feedback from customers and then quickly iterate. You may also here this method discussed as “iterative development” We use this method at Skytap with “4 week iterations”. Unlike waterfall methods, there is increased pressure using Agile to deploy/ship features and as such there is more likely hood of infrastructure changes being made more frequently than in water fall method.
So, if you are using Agile development methods. What does that mean for your infrastructure requirements and for operations of that software? The weakest link in the chain from development to test to operations may become the infrastructure as your teams deliver more features faster. Let’s look at the challenges development teams using agile may face.
Development teams face big hurdles when they need new infrastructure. Clearly ordering hardware, having it shipped, unpacked, installed in the racks, and configured can take weeks to months. This is the state of many IT organizations. Projects using waterfall methods can deal with these delays, but not agile development teams. Even companies that are fully virtualized, tend to take 24-48 hours to provision a new virtual machine. On top of all that development/test teams tend to take second priority to keeping the lights on in produtions. Q: How many of you are happy with how quickly your IT organizations deliver new servers? 2) The second challenge is in changing/recycling existing hardware. As an application scales and grows, a development team may decide they need to double the CPU or memory for the servers supporting the application. It is not easy nor quick to do this with existing hardware resources. 3) Sharing environments is not very practical. If a tester finds a bug and the developer is unable to reproduce it in their own environment, the tester may be stalled as the developer takes over the test environment to troubleshoot. Likewise, creating copies and reproducing replicas of existing environments is not feasible. 4) There can be high costs in maintaining dev/test infrastructure. As compute and storage resources grow, maintaining staging environments that closely mirror production can be expensive. A lot of the times these staging environments sit unused. Larger up-front capital expense to start a project. Again for projects that will run for 12+ months this ok, but for agil projects it is very costly to get started.
So, if you are doing agile development, you will want more agile infrastructure. Can cloud computing help address these challenges?
So cloud computing brings rapid provisioning to address issue #1.
Cloud computing also makes it easy to change infrasructure on the fly. Shown here is Amazon web services where you can select specific VM configurations with their own CPU/Memory, you can also create volumes of different sizes. Rackspace has similar pre-configured VMs to chose from.
Thirdly cloud computing can bring better tools for collaborating with teams. A user can build out an set of VMs, define the networking configurations, and get everything running. [BUILD1] These cloud environments can then be “Saved” or snapshotted into a library as a template. [BUILD2] For example, imagine a developer creates a “golden” development environment and saves it as a template. A second developer can then create their own separate environment from that original template. [BUILD3] Another tenant of agile development is sharing software with customers. Rather than stopping development, you can provision a separate enviornment, and then add it to a project or publish it so other team members or even external customers can try out new changes.
Finally cloud computing can lower costs. No up-front capital expenditures. Pay for use. Many people still think they can run things more cheaply in-house.
So in summary, agile development adoption is accelerating. We see cloud computing as offering signfiicant advantages to development and test teams who are making that transition. Agile development requires Agile infrastructure. The next link in the chain to explore is operations and the movement toward dev/ops where development teams are more intimately involved in the production operations of applications.
Transcript of "Choosing Public vs. Private vs. Hybrid Cloud Computing"
Choosing Public vs. Private vs. HybridCloud ComputingCAMP ITBrett GoodwinVP Marketing & Business DevelopmentSkytap, Inc.
I 2• Cloud Computing Overview• Research & Trends• Cloud Solution Requirements• Where to Get Started?• Integrating With Existing EnvironmentsAgenda
I 3Predicting The FutureWhen he saw ademonstration of thetelephone in 1880, aU.S Mayor declared:“One day every town inAmerica will have atelephone!”
I 4Predicting The Future133 years later…Over 100 million totalsmartphone users inAmerica
I 5Innovation Is Accelerating…2005 2006 2007 2008 2009 2010 2011YouTube is bornApple launchesmovies on iTunes.The beginning ofthe end for moviestores2012Google buysYouTubeFacebookopens upThe worldstarts tweetingGoogle Books startsscanning the world’sliteratureThe iPhoneis bornObama’s campaigngrows $21M - $150Musing social media1 of 8 couples inthe US gettingmarried met onlineCloud computinggoes mainstreamOver 2.1 billionglobal internetusersFacebook has over800M active usersWikipedia grows to25M+ pages2M soldin24 hoursiPhone 5
I 6The Future Is NowThe internet revolutionchanged everything… Shop Date Conduct business Consume media Pay our bills Conduct research Learn CollaborateCloud computing will too… “Unlimited” capacity On-demand Self-service Increased collaboration Pay-as-you grow Faster innovation Increased agility Next generation ofproducts & services
I 7• Compute, Storage, and Networking Resources• Delivered as a service• Self service access via web interfaces and APIs• Provision / release resources in minutes• Pay for usage modelWhat Is Cloud Computing?7
I 8Cloud Computing ArchitecturesSecure connection:VPN (IPSec)“Hybrid Cloud”InternalInfrastructure“PrivateCloud”“Public Cloud”
I 9How Is Cloud Computing Different?Cloud Computing
I 10Cloud Evolution – 10 Year ForecastTodaySource: Forrester Research
I 11Corporate Use Of Public Clouds Is Accelerating
I 18Top SaaS Applications Used“What are your firm’s plans to use software-as-a-service (SaaS) tocomplement or replace the following applications?”Top categories:>30% adoptionNext group ofSaaS solutions:20% to 30%Base: software decision-makers at firms with 20+ employees who currently use or are planning to use eachapplication; Source: Forrsights Software Survey, Q4 2012
I 19What’s Driving The Move toCloud?Agilityandspeed
I 22Cloud Segments Are ConvergingSaaS› Fixed application› Defined structures› Singular purpose› Look and feelcustomizationsIaaS› Virtual infrastructure› OS/MW configuration› Network and storagedefinition› SaaS extensions› Customization tools› Integrations› Developer services› APIsPaaS› Application container(s)› App life-cycle mgmt.› Development tools
I 23Private, Hybrid, or Public Clouds?Gartner: Private clouds are a last resortThorough analysis required to identify cloud computing benefitsBy Neal Weinberg, Network WorldOctober 19, 2011 10:00 AM ETORLANDO, Fla. -- Enterprises should consider public cloud services firstand turn to private clouds only if the public cloud fails to meet their needs.That was the advice delivered by analyst Daryl Plummer during GartnersIT Symposium Tuesday. Plummer says that there are many potentialbenefits to deploying cloud services, including agility, reduced cost,reduced complexity, increased focus, increased innovation and being ableto leverage the knowledge and skills of people outside the company.
I 24Hybrid Cloud Architecture24Your Infrastructure Public CloudMulti-VPN via IPsecOn-PremiseVirtualized ResourcesCloud-Based VirtualEnvironments
I 25• Some enterprise workloads are ideally suited forin-house; some are ideal for cloud delivery• The hybrid cloud becomes a secure, IT managedextension to your existing internal IT infrastructure• Hybrid clouds provide the best of both worlds– Elasticity, on-demand, self-service nature of public cloud– Security, control, and visibility of on-premise resourcesWhy Hybrid Cloud?
I 26IT Workload Management• Contextual for IT operations• Limited IT resources• User managed changes• 30% of the environment 70% of change requests• Core to IT operations• Dedicated IT resources• IT managed changes• 70% of the environment 30% of change requestsPredictable Dynamic• Enterprises are moving dynamic workloads to the cloud• Cloud model provides for dynamic capacity management• Automation solutions decrease IT support burdenDynamicPredictableSource: Customer interviews
I 27Cloud Computing – Where to Start?27Anatomy of aAnatomy of aDynamic WorkloadDynamic Workload• Unpredictable capacity needs• Frequent change requests• Requires collaboration• High costsLeading DynamicLeading DynamicWorkloads forWorkloads for the Cloudthe Cloud• Development and Testing• Software Demos and Evaluations• Virtual TrainingThe cloud is ideal for your dynamic workloadsCommon challengesCommon challenges buildingbuildinginternal infrastructureinternal infrastructure• Slow to provision• Difficult to change• Challenging to share• Costly to maintain
I 28Business Benefits of Cloud Computing28IncreaseIncrease business agilitybusiness agilityReduceReduce time to markettime to marketShipShip better products fasterbetter products fasterBoostBoost productivity across teamsproductivity across teamsLowerLower your costsyour costs
I 29Example:Application Testing & DevelopmentWhere to Get Started?
I 30• Define requirements up front• Difficult to change the outcome if requirements change• IT infrastructure rarely changes during long dev/test cyclesTypical Software Development Life-CycleDesignDesignDeploy & SupportDeploy & SupportDevelopDevelopTestTest6 to 12 month development and test cycle6 to 12 month development and test cycleRequirementsShippable softwareTestable software
I 31• Short release cycles which deliver customer value• Welcome changing business requirements• Collaboration with business and customer stakeholdersAgile Development Model31Design DevelopTestShip ShipCustomerProblem 1CustomerProblem 2CustomerProblem 3Design DevelopTestDesign DevelopTest3 to 6 week release cycle3 to 6 week release cycle 3 to 6 week release cycle3 to 6 week release cycle 3 to 6 week release cycle3 to 6 week release cycle
I 33• Ordering and provisioning requests take weeks or longer• Lower priority than production operationsInfrastructure Challenges with Agile Development33Slow toProvisionSlow toProvisionDifficult toChangeDifficult toChange• Difficult to change and recycle hardware once ordered• Not easy to re-configured Memory, CPU, DiskHigh CostHigh Cost• Costly to maintain state of the art dev/test hardware• Costly to duplicate production environments• Capital expense vs. Operating expenseDifficult to ShareDifficult to Share• Sharing environments or “copies” of environments with remoteteams or customers is not practical• Reproducing complex bugs requires snapshots of full environments
I 34Cloud Computing34Can Cloud Computing Help Address these Challenges?
I 35Rapid Provisioning1 day2 to 3 weeksIn-house data based on customer inputDays• Quicker Release Cycles• Faster Time to Market• Positive Business Impact• Quicker Release Cycles• Faster Time to Market• Positive Business ImpactTime liberated to focusfor strategic priorities35
I 37Improved Collaboration37Save as TemplateNew EnvironmentAdd to Project /PublishEnvironmentTemplatesMy Environment
I 38• No up-front capital expenditure• Pay for use model– Allows for simulating production environmentsLower Cost38
I 39Putting the Agile into your Infrastructure39Agile development adoption is acceleratingCloud computing offers significantadvantages for development andtest workloads
I 40• Can I import my existing VMs/configurations into my cloudservice provider “as is”?• Do I have to write/rewrite my application to a new/specific setof APIs? If so, will I be “locked in”?• Can my cloud service provider handle complex networking andmuti-tiered environments?• Can end users access the cloud resources from any browser?Also RDP; SSH clients?• Am I comfortable that my cloud environment is secure? My DCto cloud; users to cloud; not exposed to the Internet unless Iwant it to be…Integrating With Existing Infrastructure
I 41• Self-service and ease of use• Visibility and control• Fast on-demand provisioning• Complete complex environments vs. VMs• Designed for teams vs. individuals• Hybrid/public cloud architectureKey Things To Look For
• Cloud transformation is underway• Take a holistic view– Private infrastructure– Private cloud– Public/hybrid cloud computing• Cloud computing enables best of both worlds– In-house private clouds/data centers for business critical– External hybrid/public clouds for dynamic workloadsSummary
Thank You!CAMP ITBrett GoodwinVP Marketing & Business DevelopmentSkytap, Inc.firstname.lastname@example.org
Choosing Public vs. Private vs. HybridCloud ComputingCAMP ITBrett GoodwinVP Marketing & Business DevelopmentSkytap, Inc.