Cloud: Session 3 - Defining the True Cloud
Download as PPTX, PDF0 likes341 views
This document discusses the definition and evolution of cloud computing. It begins by exploring early concepts of distributed computing and how they evolved into today's definition of the cloud. It then examines definitions of cloud computing from Wikipedia and NIST. The document also discusses barriers to cloud adoption, including security, integration, and scalability concerns, and provides arguments for why organizations should adopt cloud computing infrastructure.
Cloud: Session 3 - Defining the True Cloud
- 1. Cloud defining the true cloud
- 2. esteban kolsky thinkJar “cloud purist” @ekolsky
- 3. first mention ever of cloud computing model? 1906 Tweet: #SCON12
- 4. evolution of the cloud distributed network is network state of computing the computer computing sirocco 1931 1970s 1984 1990s 1999 2004 2007 grid CORBA cloud computing DCOM computing Tweet: #SCON12
- 5. windstorm that lifts up clouds of dust (similar to a hadoop) Tweet: #SCON12
- 6. defining the cloud by “the book” cloud computing is the delivery of computing as a service rather than a product, whereby shared resources, software, and information are provided to computers and other devices as a utility (like the electricity grid) over a network (typically the internet) wikipedia cloud computing is a model for enabling ubiquitous, convenient, on- demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction NIST Tweet: #SCON12
- 7. defining the cloud, the “purist” way leverage, mashup parts of history, current definitions distributed computing single instance with replication uses systems management computing as a service separation into three layers open network connected and integrated measured by the usage, not by the user public, public, and public Tweet: #SCON12
- 8. a picture worth a thousand words or more S S P P presentation I validation I interaction personalization SaaS S security business rules P compliance I directory PaaS S P connectivity database I legacy internet IaaS Tweet: #SCON12
- 9. barriers to cloud – irrational fear of the unknown Tweet: #SCON12
- 10. barriers to cloud – change management Tweet: #SCON12
- 11. scalability integration security barriers to cloud –three laughable IT excuses Tweet: #SCON12
- 12. debunking security in the cloud: problem system 1 sys 2 sys 3 system x partner vendor Tweet: #SCON12
- 13. debunking security in the cloud: solution security token, created and managed by PaaS S S S S P P P P I I I I security service S S P P I I Tweet: #SCON12
- 14. debunking integration: problem system 1 system 2 system 3 system x partner vendor Tweet: #SCON12
- 15. debunking integration: solution a S S S P P I I S P I SaaS-SaaS integration, within the application, leveraging the cloud P similar to existing, uses service I calls, middleware Tweet: #SCON12
- 16. debunking integration: solution b S S S P I P I S P PaaS-PaaS integration, leveraging I the cloud, no need for middleware, easiest for elasticity (scale out), establish once, reuse many times as necessary P I Tweet: #SCON12
- 17. scalability: three sustaining pillars cloud scalability define scalability scalability by design for must occur in scalability in two ways – Capacity the cloud up and out Availability Performance the entire concept of cloud computing was conceived for scalability sake Tweet: #SCON12
- 18. making the case for cloud computing E1 conomics easiest part of the equation: someone else hosts hardware, software, maintenance anyway to look at it, remote, distributed computing is cheaper move from CAPEX to OPEX is very attractive to managing stakeholders cloud ROI traditionally done by comparing IT budgets to cloud expenses #mmmkay scalability, integration are cheaper and easier, security is still under (hot) debate Tweet: #SCON12
- 19. making the case for cloud computing E2 lasticity inherent to cloud architectures makes case for scalability, high-availability, failover planning done wrong – remember the failure of amazon EC2 and the problems it caused is a scale-out model in single-tenancy, multi-instance, that cannot be tested requires systems management for proper instance maintenance Tweet: #SCON12
- 20. making the case for cloud computing E3 nterprise application stores mobile, social, user empowerment – all changed the game users aren’t putting up with big, complex applications – they want “apps” on iPad best bet for support for all business is adoption of EAS, cloud is a must for that model scalability, flexible interfaces, myriad device access via platform – all part of the cloud two-to-three year trend, if you haven’t started – better get going Tweet: #SCON12
- 21. making the case for cloud computing E4 xpansion plans for all business units in next few years include “cloud” and SaaS plans for IT over next three-to-four years include cloud infrastructure integration, plans from partners, providers, suppliers, associates include cloud in next five years expansion of any organization must include cloud going forward, no questions if all other fails, cite the fact that competitors are expanding into cloud #Winning Tweet: #SCON12
- 22. making the case for cloud computing E5 volution evolution of the world comes down to cloud, organizations have to adapt vendors are evolving their products from hosted, on-demand to cloud ventures with partners, other organizations will demand cloud; three-to-four years customers are demanding cloud –stirred up by microsoft “to the cloud” TV ads with differentiation of what is cloud and how to deploy in it, definitely a must-do Tweet: #SCON12
- 23. making the case for cloud computing E6 infrastructur reports say from 10% to 30% of IT budgets put on cloud infrastructure reports highlight more money spent by BU than IT on SaaS applications; continuing organizations cannot leverage the cloud as an external service only, must integrate to make cloud work across organization, to retain security, to enhance integration – IT must be responsible for centralized cloud infrastructure; CEOs are seeing this Tweet: #SCON12
- 24. Issues: hosted != on-demand != cloud hosted: a vendor puts a web-interface to their apps MT (multitenancy) is a big deal here licensing is by the user, as usual reduced functionality or poor performance by comparison on-demand: vendors leverages SaaS,PaaS model web interface MT benefits vendor; not so much clients usually SaaS and PaaS “all in one” failing to leverage cloud cloud three layer model as described earlier MT, single-instance hard to scale; ST, multi-instance better licensing by usage, starting, is winner model for biz Tweet: #SCON12
- 25. issues: private cloud short, simple, sweet any well done definition of the cloud specifically calls for open networks to support the cloud; private clouds violate core tenet of cloud today’s private cloud is used as a bridge between client-server, web-based, on-demand, and SaaS worlds to get to the real cloud infrastructure use if you must, you are better off planning around it Tweet: #SCON12
- 26. your next steps define the cloud (for you) take charge of your existing “cloud” initiatives across stakeholders make case for central IT control setup 3-5 years strategy and investment get going, do your infrastructure while you leverage SaaS and on demand solutions learn, implement, learn, implement – you get the idea Tweet: #SCON12
Editor's Notes
- #4: American society of mechanical engineersNew concepts technologies, such as distributed computingUsing computers? No, but the concept was developed with views to the future
- #5: 1931 –Air Force invest $3MM towards understanding how distributed computing can change the world1946 – RAND defines distributed computing models1970 – IBM starts playing with models for grid computing (not like today, think shared timeshare, plus the “owned” most of the “iron)1976 – NCC the future of distributed computing systems, “future trends in distributed computing systems” mentions grid computing1984 – SUN (stanford university network) computer1991,1993 –- CORBA, COM/DCOM - Larry Elison, networking computing is nothing new – why make such a fuss?1999 – salesforce launches using SaaS model (from distributed computing fame, does not mention cloud yet – no software, hosted was their model – MT begins in earnest)2004 – eric schmidt make first reference to cloud computing since internet came about, starts craziness - 2009 – sirocco (wikipedia -
- #10: FearChange of infrastructureUpgrade cycleThree core problemsScalabilityIntegrationSecurity
- #19: EconomicsElasticityEnterprise Application StoresExpansionEvolutionInfrastructure