Presentation introduction to cloud computing and technical issues

Download Presentation   introduction to cloud computing and technical issues

Post on 14-Jan-2015

961 views

Category:

Technology

0 download

Embed Size (px)

DESCRIPTION

 

TRANSCRIPT

<ul><li> 1. Introduction to Cloud Computingand Technical Issues Eom, Hyeonsang School of Computer Science &amp; Engineering Seoul National UniversityFuture Internet Summer Camp 2009pCopyrights 2009 Seoul National University All Rights Reserved</li></ul><p> 2. Outline What is Cloud Computing? Clouds vs. Grids Cloud Computing Architecture Cloud Services Technical Issues on Cloud Computing Copyrights 2009 Seoul National University All Rights Reserved 3. Cloud Computing- Buzz and Fuzz Wordcomputation may someday be organized as a public utility John McCarthyMcCarthy, 1960cloud computing can take on different shapes depending onthe viewer, and often seems a little fuzzy at the edges James Obriena cloud is a pool of virtualized resources that can host a varietyof different workloads, allow workloads to be deployed andscaled-out quickly, allocate resources when needed, and supportq y, , ppredundancyGreg Boss et al., IBM Chaisiris presentation Copyrights 2009 Seoul National University All Rights Reserved 4. What Is Cloud Computing? Internet computing Computation done through the Internet No concern about any maintenance or managementof actual resources Shared computing resources As opposed to local servers and devices Comparable to Grid Infrastructure Web applications Specialized raw computing servicesCopyrights 2009 Seoul National University All Rights Reserved 5. Cloud Computing Resources Large pool of easily usable and accessiblevirtualized resources Dynamic reconfiguration for adjusting to different loads(scales) Optimization of resource utilization Pay-per-use model Guarantees offered by the Infrastructure Provider bymeans of customized SLAs(Service Level Agreements) Set of features Scalability, p y py pay-per-use utility model and virtualizationy Copyrights 2009 Seoul National University All Rights Reserved 6. Technical Key Points User interaction interface: how users of cloudinterface with the cloud Services catalog: services a user can request g q System management: management of availableresources Provisioning tool: carving out the systems from thecloud to deliver on the requested service Monitoring and metering: tracking the usage of thecloudl d Servers: virtual or physical servers managed bysystem administratorsCopyrights 2009 Seoul National University All Rights Reserved 7. Key Characteristics (1/2) Cost savings for resources Cost is greatly reduced as initial expense and gyprecurring expenses are much lower thantraditional computingg Maintenance cost is reduced as a third partymaintains everything from running the cloudto storing data Platform Location and Device independencyPlatform, Adoptable for all sizes of businesses, inparticular small and mid sized onesmid-sized Copyrights 2009 Seoul National University All Rights Reserved 8. Key Characteristics (2/2) Scalable services and applications Achieved through server virtualization gtechnology Redundancy and disaster recovery Cloud computing means getting the best performing system with the best monetary value Copyrights 2009 Seoul National University All Rights Reserved 9. TimelineGrid ComputingUtility Computing Software-as-a-Cloud ComputingService (SaaS)VolunteerHPs Utility y Google AppsgppGoogle AppgppComputingC tiData CenterEngine(e.g.SETI@home) Amazon EC2 MicrosoftGlobusSun Grid AzureToolkitT lkitComputingC ti(from GT2-UtilityGT4) Copyrights 2009 Seoul National University All Rights Reserved Wikipedia.com 10. Clouds vs Grids vs. Distinctions: not clear maybe because Cloudsyand Grids share similar visions Reducing computing costsgpg Increasing flexibility and reliability by using third-partyoperated hardware Grid System that coordinates resources which are notsubject to centralized control, using standard, open,general-purpose protocols and interfaces to delivernontrivial qualities of service Ability to combine resources from differentorganizations f a common goal i tiforl Copyrights 2009 Seoul National University All Rights Reserved 11. Feature Comparison (1/3) Resource Sharing Grid: collaboration (among Virtual Organizations,for ) fair share) Cloud: assigned resources not shared Virtualization Grid: virtualization of data and computing resources Cloud: virtualization of hardware and software platforms Security Grid: security through credential delegations Cloud: security through isolation Copyrights 2009 Seoul National University All Rights Reserved 12. Feature Comparison (2/3) High Level Service Grid: Plenty of high level services Cl d N hi h lCloud: No high level services d fi d yet. l idefined t Architecture Grid: Service oriented Cloud: User chosen architecture Platform Awareness Grid: need for Grid-enabled client software Cloud: SP (Service Provider) software working in acustomized environmentCopyrights 2009 Seoul National University All Rights Reserved 13. Feature Comparison (3/3) Scalability Grid: node and site scalability Cloud: node, site and hardware scalability Self Management g Grid: reconfigurability Cloud: reconfigurability and self-healing g y g Payment Model Grid: rigid Cloud: flexibleCopyrights 2009 Seoul National University All Rights Reserved 14. Cloud Computing Architecture Front End End user, client or any application (i.e., web browser,etc.) Back End (Cloud services) Network of servers with any computer program anddata storage system It is usually assumed that clouds have infinite storage capacity for any software available in market Copyrights 2009 Seoul National University All Rights Reserved 15. Copyrights 2009 Seoul National University All Rights Reserved 16. Cloud Service Taxonomy LLayer Software-as-a-Service (SaaS) Platform-as-a-Service (PaaS) Infrastructure-as-a-Service (IaaS) Data Storage-as-a-Service (DaaS) g ( ) Communication-as-a-Service (CaaS) Hardware-as-a-Service(HaaS) TType Public cloud Private cloud Inter-cloud Copyrights 2009 Seoul National University All Rights Reserved 17. Cloud Computing Service Layer Copyrights 2009 Seoul National University All Rights Reserved 18. Software-as-a-Service Software as a Service (SaaS) Definition Software deployed as a hosted service and accessed over the Internet Features Open, Flexible Easy to Use Easy to Upgrade Easy to Deploy Copyrights 2009 Seoul National University All Rights Reserved 19. Overall Technology Stack SaaS layer Most of SaaS Applications are implemented based on available modules il bl d l in the way that new business logics are realized; e.g., salesforce.com g, Copyrights 2009 Seoul National University All Rights Reserved 20. SaaS Business Example 1 Copyrights 2009 Seoul National University All Rights Reserved 21. SaaS Business Example 2 Copyrights 2009 Seoul National University All Rights Reserved 22. Platform-as-a-ServicePlatform as a Service (PaaS) Definition Platform providing all the facilities necessary to support the complete process of building and delivering web applications and services, all available over the Internet Entirely virtualized platform that includes one or more servers, servers operating systems and specific applications Copyrights 2009 Seoul National University All Rights Reserved 23. What Is PaaS ?Copyrights 2009 Seoul National University All Rights Reserved 24. PssS Example: Google App Engine S i th t allows user t d l users W bService that ll to deploy Webapplications on Googles very scalablearchitecture Providing user with a sandbox for users Java andPython application that can be referenced over theInternet Providing Java and Python APIs for persistentlystoring and managing data (using the Google QueryLanguage or GQL)Copyrights 2009 Seoul National University All Rights Reserved 25. Google App EngineGoogle App Engine lets you run your web applications on Googles infrastructureGoogle App Engine supports apps written in several p g gpp gpppp programming languages including Python and gg gg y Java Google clusterData Store Cluster G o o g le s R P CHTTPnodenodenode G o o g l ee s L o a d nodeRequestnode node nodenodebrowseHTTP er mm e c h a n i s mm Google clusterData Store BalacnerResponsenodenode Clusternode node node noder nodenode Python (or Java) Web Server Persistent LayerCopyrights 2009 Seoul National University All Rights Reserved 26. Google App Engine (Python Example)WebAppFrameworkF k Django FDjFrameworkk WebOb Framework W bOb F k (Googles (Third Party) (Third Party)CGI Inframework) nterface Python RuntimeMemCData Mail UsersacheStore APIs APIs APIs APIsgoogle.appengine.ext.dbli t db QuotaMemCache Expand EnginemodeloQuery GQLTransactionQueryPropert Manager Engine i y BigTableCopyrights 2009 Seoul National University All Rights Reserved 27. Infrastructure-as-a-Service Infrastructure as a Service (IaaS) D fi itiDefinition Provision model in which an organization outsources the equipment used to support operations, including operations storage, hardware, servers and networking components Infrastructure as a Service is sometimes referred to asHardware as a Service (HaaS). The service provider owns the equipment and is responsiblefor housing, running and maintaining it The client typically pays on a per-use basis Copyrights 2009 Seoul National University All Rights Reserved 28. IaaS PlatformWorkload / Applications(Self-) Service Tools(S lf ) S i T l y Utility Fabric Delivering IaaS gDynamic Resource Management Virtualization ProvisioningAutomation Server StorageNetworkInfrastructureCopyrights 2009 Seoul National University All Rights Reserved 29. What Is Infrastructure-as-a-Service (IaaS) Characteristics UtilityUtilit computing and billi model tid billing d l Automation of administrative tasks Dynamic scalingD ili Desktop virtualization Policy-based services Internet connectivityCopyrights 2009 Seoul National University All Rights Reserved 30. Use Scenario for IaaS keithpij.com.dnnmax.com Copyrights 2009 Seoul National University All Rights Reserved 31. Business Example: Amazon EC2 Automation CommunicationDaaSClient InterfacesAmazon S3 FilesystemCopyrights 2009 Seoul National University All Rights Reserved 32. Amazon EC2 CLI(Client Level Interface)Copyrights 2009 Seoul National University All Rights Reserved 33. Amazon EC2 Automated ManagementWeb Traffic Elastic Load High Available AmazonManagement Scalable CloudWatch Balancing Fault tolerant Real-Time Visibility No deployment Better resource utilization No maintenance Cost effective Auto Scaling Potential cost savings Ensure capacity availability Tools APIs Capacity on demand Amazon EC2 Boot in minutes Pay-as-you-goCopyrights 2009 Seoul National University All Rights Reserved 34. Data Storage as a Service (DaaS) Definition Delivery of data storage as a service, including database-like services, often billed on a utility computing basis Database (Amazon SimpleDB &amp; Google App EnginesBigTable datastore) Network attached storage (MobileMe iDisk &amp; NirvanixCloudNAS) Synchronization (Live Mesh Live Desktop component &amp;MobileMe push functions) Web service (Amazon Simple Storage Service &amp; NirvanixSDN) Copyrights 2009 Seoul National University All Rights Reserved 35. Amazon S3PUT http://S3hostname/bucket/objectnameGET http://S3hostname/bucket/objectnamewith shared key S3 REST Interfacebucketbucket objectCopyrights 2009 Seoul National University All Rights Reserved 36. Putting It All Together Case Study of Alexa Web AlAlexa W b SWeb Search hweb service T build customizedTo b ild t i dsearch enginesagainst the massive gdata that Alexa crawlsevery night T query the AlexaToth Alsearch index and getMillion Search Results(MSR) back as outputCopyrights 2009 Seoul National University All Rights Reserved 37. GrepTheWeb Architecture Amazon S3For retrieving input datasets and forstoring the output data setAmazon SQSFor durably buffering requests acting asa glue between controllers Amazon SimpleDBFor storing intermediate status, log, andfor user data about tasksAmazon EC2For running a large distributedprocessing Hadoop cluster on-demand HadoopFor distributed processing, automaticparallelization, and job scheduling Copyrights 2009 Seoul National University All Rights Reserved 38. Data and Control Flow Copyrights 2009 Seoul National University All Rights Reserved 39. MapReduceCopyrights 2009 Seoul National University All Rights Reserved 40. Public Cloud Services D fi itiDefinition The standard cloud computing model Th SP makes resources, such as applications and storage,The khli ti d tavailable to the general public over the Internet Free or offered on a pay-per-use modelp yp Examples of public clouds Amazon Elastic Compute Cloud (p (EC2), IBMs Blue ), Cloud, Sun Cloud, Google AppEngine and Windows Azure Services Platform. Copyrights 2009 Seoul National University All Rights Reserved 41. Private Cloud Services Internal cloud or corporate cloud Definition Proprietary computing architecture that provideshosted services to a limited number of people behinda firewall. Designed to appeal to an organization that needs or wants more control over their data than they can get by using a third-party hosted service Copyrights 2009 Seoul National University All Rights Reserved 42. Copyrights 2009 Seoul National University All Rights Reserved 43. Inter-Cloud Inter Cloud Definition Federation of clouds based on open standards Concept primarily promoted by Cisco Similar Concepts Cloud of Clouds, Cloud Interoperability, etc., py, Copyrights 2009 Seoul National University All Rights Reserved 44. Why Inter Cloud? Inter-Cloud ? InterGrid Acquiring resources from peering Grids to serveoccasional peak demands Solving the problem of resource provisioning inenvironments with multiple Grids Evolvable and sustainable systemCopyrights 2009 Seoul National University All Rights Reserved 45. Cisco s Ciscos VisionCopyrights 2009 Seoul National University All Rights Reserved 46. Architecture of Inter-CloudStandards (by Cisco) Copyrights 2009 Seoul National University All Rights Reserved 47. Technical Issues of CloudComputing Virtualization SecurityVi t li ti Sit Reliability Monitoring ManageabilityCopyrights 2009 Seoul National University All Rights Reserved 48. Virtualization Security (1/2) Virtualization Abstracting the underlying resources so that multiple operating systems can be run on a single physical simultaneouslyy Improving resource utilization by sharing available resources to multiple on demand needs Copyrights 2009 Seoul National University All Rights Reserved 49. Virtualization Security (2/2) Virtualization security Including the standard enterprise security policies on access control, activity monitoring and patch management M t enterprises just starting to grasp but not fully Most t ij t t ti t b t t f ll understanding Many IT people still believe that the hypervisor and virtual machinesare safe Becomeing one of the factors when virtualization technologies move into the cloud Access control and monitoring of the virtual infrastructure will be ontop of providers mind providersCopyrights 2009 Seoul National Uni...</p>

Recommended

View more >