Cloud Architecture For Dummies - Bitpipe

5of systems with few changes, but cloud computing promotes multiple large systems (scale) with constantchange (velocity) to support the cloud. Expect the magnitude and dynamic nature of the cloud to be fundamentally different than traditional IT systems.Benefits of a Cloud ArchitectureMoving to a cloud computing architecture represents asignificant potential shift in your IT mentality. To makesuch a fundamental change requires strong motivatingbusiness factors to encourage the change. Cloud computing comes with many such motivators, but they cangenerally be grouped into three key areas: lower cost,business agility, and a focus on innovation and corecompetencies.Lower costCloud computing drives down IT cost. Rather than purchasing and supporting dedicated resources to supportthe peak demand for each application, cloud resourcesare shared among multiple applications via resourcepooling. Pooling enables higher resource utilization andless excess capacity, resulting in lower hardware, software, administration, and infrastructure costs.A measured, metered service model ensures that usersonly pay for the resources they consume, not forunneeded capacity. Whether it’s on-demand use of apublic cloud, or a private cloud, lower cost for theorganization is realized.Business agilityRapid time to market and simplified deployment fornew applications are key motivators. On-demandThese materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

6provisioning and rapid elasticity allow applications tobe deployed and scaled very rapidly. Faster deployments increase the organization’s agility and allow ITto be more responsive to the business. Excess capacitymay be reduced and systems can be scaled as neededfor increasing demand.Focus on innovation and corecompetenciesShifting to a cloud architecture liberates your organization’s best and brightest staff to focus on innovationand core competencies rather than supporting traditional IT systems. Your staff can focus on the business’s core competencies, differentiators, and sourcesof competitive advantage.Challenges to CloudArchitectureAs with any new idea, cloud architecture does facesome challenges. Although some challenges are organizational and philosophical, others are more tangible.Here is a list of common challenges faced by clouds: Abstraction. An effect of cloud architecture is thatdetails of the IT system are abstracted and simplified. Determining the most useful level of abstraction and granularity is a challenge. Delegation of control/lack of control. Undercloud architecture, the underlying technologiesare no longer under the organization’s physicalcontrol. Coming to terms with not having exclusive, direct physical control over server hardwareThese materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

7and infrastructure is a significant organizationalchange in IT culture. Application portability. Applications with uniquehardware platforms or nonstandard infrastructurerequirements are difficult to support using capacityon-demand, shared-services environments.Standards are essential. Security. Placing critical, sensitive information ina cloud will generate an emotional response inmany people. But, with the right technical measures, processes, and policies, high-level securitycan be achieved. Proliferation and version control. Organizationsmust ensure that they are disciplined not to growand sprawl out of control. Unnecessary resourcesand out-of-date components need to be decommissioned and released back into the cloud forthe benefit of all cloud consumers. Disciplineand metered service contracts and fees enforceresponsible cloud use. Regulatory compliance. Cloud computing emphasizes distributed IT. But different industries andregions of the world have unique regulatoryrequirements about the location of data and howdata must be managed and accessed. Transparency. The degree of information the consumer requires regarding the technical hardwarearchitecture of the cloud varies based on thecloud model. Theoretically, the infrastructure ofthe cloud shouldn’t matter to the consumer(unless there is a regulatory issue), but it doesmatter to the service provider.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

8The challenges of cloud computing are manageable andare outweighed in comparison to the benefits to theorganization. A key way of dealing with these challenges, and reaping the desired benefits, is to makesure that you adopt a thoughtful architecture (seeChapter 2).These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

Chapter 2Cloud ArchitectureIn This Chapter Identifying three cloud service models Explaining four cloud deployment models Describing when hybrid clouds make sense Defining architecture principles and guidelinesThis chapter examines service models, deploymentmodels, hybrid cloud scenarios, and the architecture principles and guidelines for cloud architecture.Service ModelsThe type of computing resource that is offered in acloud defines a cloud’s service model. NIST and theindustry have identified three common service modelsthat are based on what cloud services are provided:applications, platform, and/or infrastructure.Software as a Service (SaaS)With Software as a Service (SaaS), service consumersget their software applications from the service provider. The consumer uses the software as anThese materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

10 pplication while the provider manages the underlyingaplatform software and infrastructure hardware.Platform as a Service (PaaS)In Platform as a Service (PaaS), the consumer (a developer) uses programming languages, tools, and the runtime platform from the provider for applicationdevelopment and deployment. The platform mayinclude databases and middleware in addition to application development tools.Infrastructure as a Service (IaaS)With Infrastructure as a Service (IaaS), the providermanages the underlying physical cloud infrastructure(servers, storage, network, and the associated virtualization and operating systems software) while the consumer deploys and runs his or her own application andplatform software. Virtualization software is often a keyenabler for IaaS architectures.Figure 2-1 shows how service models are related.In Figure 2-1, you see SaaS providing cloud applications, PaaS providing cloud platforms, and IaaS providing cloud infrastructure with usage examples of eachenvironment. It is important to note that each model isnot dependent on the others. A SaaS applicationabstracts the underlying hardware and software, soSaaS need not run on an exposed PaaS or IaaS.Similarly, PaaS abstracts the underlying hardware andneed not run on an exposed IaaS.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

11DeveloperBusiness licationPlatformSaaS CloudPaaS CloudConsumerIT oviderConsumerDifferentUsersIaaS CloudFigure 2-1: Cloud service model hierarchy.Deployment ModelsThis section defines four cloud deployment models.These models are based on who has access to thecloud.Private cloudsPrivate clouds are for the exclusive use of a single organization. The cloud is often hosted by the same organization in a private data center, but it may also behosted by a third-party service provider. Private cloudscan offer a high level of control and visibility by themanaging organization and are perceived as potentiallyproviding a higher degree of privacy and security.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

12Public cloudsPublic clouds are used by multiple organizations on ashared basis. The cloud is hosted by an outside partyproviding the cloud as a service. Public clouds do notrequire an upfront capital investment and are thereforefast and inexpensive to get started — but consumerstypically have less control and visibility over the service, compared to private clouds, since public cloudsare shared and service providers are motivated to provide a standardized offering to reduce their costs.Community cloudsCommunity clouds are used by related organizations orconsumers with a common purpose. These are notavailable to the general public, but are available tomembers of a specific community. Examples includebranches of the military, government or educationalorganizations, or industry partners and suppliers.These are especially useful where the consumers sharespecial needs (for example, high security), or there is arequirement for common services.Hybrid cloudsHybrid clouds occur when an organization’s IT requirements are supported by a mix of private, public, andcommunity clouds in order to take advantage of thebenefits of multiple cloud types. Today, these are mostlikely special use cases, but in the long term this is thelikely future of all IT. Some specific types of hybridclouds are covered in the next section.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

13Hybrid Cloud ScenariosReal-world implementations may not fit exclusively in apublic, private, or community cloud. In many cases,using both public and either private or communityclouds makes sense. This section talks about threecommon scenarios: life cycle distribution, functionaldistribution, and cloud bursting.Life cycle distributionThis scenario involves separating development andtest from production. An enterprise may want to leverage public clouds for rapid, low-cost development andthen move to the corporate data center for “missioncritical” production.Figure 2-2 shows these separate environments with differing requirements deployed to different clouds.Private CloudPublic gure 2-2: Life cycle distribution model.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

14Here, you see a lifecycle in which development and testing is located in a public cloud. Then the applicationmoves to production in a private cloud. On the otherhand, a startup might develop and test on a private“cloud” (for example, a developer’s laptop) and thenmove the resulting apps to production in a public cloud.Functional distributionOrganizations might split functionally different applications or different parts of an extended business process into different clouds depending on characteristicsof the applications. These applications run in differentclouds but are integrated with each other across theclouds. A possible scenario here is using generic business functionality as SaaS from a public cloud, butkeeping custom or proprietary components in the private cloud. This is an example of using cloud to focuson innovation and core competence.Figure 2-3 shows different business applicationsdeployed across public and private clouds.Private CloudPublic CloudHRMClaims ProcessingCRMEnterpriseProviderFigure 2-3: Functional distribution model.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

15In this scenario, you see a sensitive claims-processingapplication deployed to an internal private cloud. Theclaims application takes advantage of the privatecloud’s benefits but still has access to the generic HRM(human resource management) and CRM (customerrelationship management) applications deployed in apublic cloud.Cloud burstingSome applications have temporary workload spikesthat exceed the processing capacity in the existingcloud. In these situations, the additional workload isoff-loaded to another cloud to handle the processing.Figure 2-4 shows cloud bursting due to capacitylimitations.Private CloudPublic CloudSynchronizationClaims ProcessingEnterpriseClaims ProcessingProviderFigure 2-4: Cloud bursting model.In this example, normal application processing occurs inthe private cloud. Once workload exceeds the capacityof the private cloud, that excess workload is moved to apublic cloud for processing.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

16Cloud bursting can also support Disaster Recovery (DR)where the primary site fails and the application failsover to another cloud that may be a different model.One complication of this scenario is the needto simultaneously synchronize two distributedenvironments, which may not work for complicated real-time transactions.Architecture Principlesand GuidelinesTo be successful in using a public cloud or designing aprivate cloud, it’s useful to embrace some architecturalprinciples (rules) and guidelines (recommendations).When designing a cloud or examining a perspectivecloud to use, consider whether it meets the followingprinciples and guidelines: Conformity to standards. Cloud interfaces, formats, and specifications must conform to industrystandards. Interoperability and portability isensured by knowing and following standards. Perceived simplicity. Clouds and computer systems are inherently complex, but to the greatestdegree possible that complexity should bemasked from the consumer. Just automatingmanual provisioning won’t solve the challenge.The cloud should only present the information tothe consumer that is necessary to perform a specific function. Abstractions, models, and policiesare methods of simplifying the environment.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

17 Visibility. If the system cannot be measured, itcannot be effectively managed. The architectureand technology should lend itself to monitoring allresources. Tools to monitor and manage allaspects of the cloud (hardware, software, and network) need to be implemented in accordance withindustry standards, but first decide what yourmeasurable objective is. Transparency. Due to the high level of abstractionand because no single group or organization controls the “cloud,” transparency of key items isessential. Performance metrics, security, reliability, and availability data must be visible to thecloud consumer. Fail in place. Availability should not be limitedby hardware failures, which are inevitable.Redundancy and intelligent management softwarecan leverage standardized resource pools tooperate continuously despite component failures.By following these key architecture designprinciples and guidelines, a cloud has a muchhigher degree of long-term relevance andsustainability.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

18These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

Chapter 3Cloud ModelingIn This Chapter Looking at perspective Defining the cloud maturity model and relevantdomains Exploring logical model views of the cloud Examining cloud management capabilitiesIt may be tempting to just jump right into building acloud environment — but like with all IT, it’s probably better to do some thinking up front. Understandingyour cloud perspective and modeling the componentsof the cloud are critical. From software payloads touser roles, models will guide you to architect, implement, and operate the cloud. Appropriate modelingallows for a more successful cloud experience.It’s really just about making your key cloud decisions.This chapter helps you to identify your cloud perspective. You also find out about the cloud maturity model,cloud architecture logical views, and finally cloudmanagement.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

20Cloud PerspectiveBefore embarking on any multi-step cloud-building process, it is important to first identify your cloud perspective. That is, how do you approach and control theimpact and benefits of a cloud? Everyone should be onthe same page.First, you want to scope out the cloud project. For anenterprise cloud initiative, it’s critical to understandwhat your “perspective” or motivation is for implementing the cloud. Consider where you are on each ofthree dimensions: Cost versus agility. Do you want to reduce ITcosts or do you want to achieve technical agility? IT driven versus business unit driven. Who’s incharge? Is the cloud initiative being driven by thegeneral IT organization (often cost focused) ordirectly by a line-of-business organization (usuallyflexibility and/or innovation focused). IT as support versus IT as a business. If the organization’s business is an IT product or service, or if it otherwise views IT as a differentiatingmission-essential element, then the speed andagility advantages of cloud computing are oftenthe driving motivator.Answering these questions will guide you in using thefollowing tools, designing your cloud, and being able tomeasure your success.These materials are the copyright of John Wiley & Sons, Inc. and anydissemination, distribution, or unauthorized use is strictly prohibited.

21Cloud Maturity ModelA maturity model is used to benchmark your implementation against others in a similar environment to determine your relative position and where you stand interms of achieving your goals. This cloud maturitymodel provides a method to use when charting yoursuccess and identifying areas that need improvement.The cloud maturity model is composed of capabilitieswithin eight

a cloud maturity model and a logical architectural model for cloud, and examines the cloud management infrastructure. Chapter 4: Cloud Implementation. This chapter delves into how engineered systems apply to cloud computing, public cloud options, and technologies that make the most sense for the cloud.