A Cloudy Weather Forecast R. Wolski, UCSB 1 Trends in Web Search - PDF document

Introduction to Cloud Computing Electrical and Computer Engineering Department Rutgers The State University of New Jersey Rutgers, The State University of New Jersey A Cloudy Weather Forecast R. Wolski, UCSB 1

Trends in Web Search (ack: Google) Legend: Cluster computing , Grid computing , Cloud computing Nov 15 2007 – IBM Introduces 'Blue Cloud' Computing, CIO Today � Apr 14 2008 – Google and Salesforce.com in cloud computing deal, Siliconrepublic.com � Jun 27 2008 – Yahoo realigns to support cloud computing, 'core strategies', San Antonio Business � Journal Jul 8 2008 – Merrill Lynch Estimates "Cloud Computing" To Be $100 Billion Market, SYS-CON � Media Jul 23 2008 – Cloud Computing Firm Closes $1.5m Series A, SYS-CON Media � What is Cloud Computing � The Definition Wikipedia - Cloud computing is Internet ('Cloud') based development and use of computer � technology ('Computing'). The cloud is a metaphor for the Internet (based on how it is depicted in computer network diagrams) and is an abstraction for the complex infrastructure it conceals[1]. It is a style of computing where IT-related capabilities are provided “as a service”[2], allowing users to access technology-enabled services from the Internet ("in the cloud")[3] without knowledge of, expertise with, or control over the technology infrastructure that supports them[4]. According to the ti ith t l th t h l i f t t th t t th [4] A di t th IEEE Computer Society "It is a paradigm in which information is permanently stored in servers on the Internet and cached temporarily on clients that include desktops, entertainment centers, table computers, notebooks, wall computers, handhelds, etc."[5]. “ � No Consensus on a good definition of “Cloud computing” - Today anything and everything Internet will come with a cloud computing logo � The Bottom-line � The Bottom line � Changes the economics of Computing from being a Capital investment to Utilities (You buy electricity, you don’t buy generators ) � Changes the way software is developed – Hardware provisioning , Deployment and Scaling now part of developer lifecycle as a program / script as compared to a Purchase order � Automates a whole bunch of infrastructure related tasks and activities leading efficiencies and cost savings Niraj Juneja, WebScale Solutions 2

What is a Cloud? SLAs Web Services Virtualization R. Wolski, UCSB Clouds … . � IT resources provisioned outside of corporate data center � Resources accessed over the internet � Variable cost of services � SaaS, PaaS, IaaS, HaaS � A virtual computing environment � Build and deliver always-on, pay-per-use IT services � Near infinite-scale computing, storage, database, related Web services, AND users � Scaling resources and services up and down � Abstraction of the hardware from the service � No need for on-premises software and servers 3

Motivations for Clouds Why Now? � Consolidation and virtualization of cluster-based enterprise Grids and datacenters has become essential essential � Increasing size, cost and energy consumption make efficiency and utilization critical � Experience with very large datacenters, Grids � Unprecedented economies of scale � Other factors � Pay by use instead of provisioning for peak � Risk of over-provisioning: underutilization � Heavy penalty for under-provisioning: Lost revenue, users 4

Economics of Cloud Providers � 5-7x economies of scale [Hamilton 2008] Cost in Cost in Resource Ratio Medium DC Medium DC Very Large DC Very Large DC Network $95 / Mbps / month $13 / Mbps / month 7.1x Storage $2.20 / GB / month $0.40 / GB / month 5.7x Administration ≈ 140 servers/admin >1000 servers/admin 7.1x � Extra benefits � Extra benefits � Amazon: utilize off-peak capacity � Microsoft: sell .NET tools � Google: reuse existing infrastructure UC Berkeley RAD Lab Economics of Cloud Users • Pay by use instead of provisioning for peak Capacity Resources Resources Capacity Demand Demand Demand Time Time Static data center Data center in the cloud Unused resources UC Berkeley RAD Lab 5

Economics of Cloud Users • Risk of over-provisioning: underutilization Capacity Unused resources Resources Demand Time Static data center UC Berkeley RAD Lab Economics of Cloud Users • Heavy penalty for under-provisioning es Resource Capacity Demand Resources 2 3 1 Capacity Time (days) Lost revenue Demand 2 3 1 es Time (days) Time (days) Resource Capacity Demand 2 3 1 Time (days) Lost users UC Berkeley RAD Lab 6

The Cloud of Cloud Companies • Amazon • Akamai • Google • Areti Internet • Salesforce • Enki • Microsoft • Fortress ITX • Joyent • Sun • IBM • Layered Technologies • Rackspace • Oracle • Terremark • EMC • Cloudera • Xcalibre • • Cloudsoft • Wolfgang Gentzsch, DEISA Spectrum of Clouds � Instruction Set VM (Amazon EC2, 3Tera) � Bytecode VM (Microsoft Azure) � Framework VM � Google AppEngine, Force.com Lower-level, Higher-level, Less management Less management More management More management EC2 Azure AppEngine Force.com 7

The Cloud of Clouds – A Market Analysis Niraj Juneja, WebScale Solutions Peter Laird’s Taxonomy of Cloud Computing Source: http://peterlaird.blogspot.com/2008/09/visual-map-of-cloud-computingsaaspaas.html 8

17 Cloud Computing Challenges � Privacy: companies need to be ready to hand over their data to a third party => big leap of faith � Need of massive server farms and high-speed networks (power, cooling, management, … ) (p , g, g , ) � Scalability, any � Security of data, applications, and user-related data � Reliability: outages damage customers’ business (S3 in July) � Vendor lock-in and industry standards � Vaporization (cloud-enabling) of applications � Vaporization (cloud enabling) of applications � Service level agreements, regulation, … � Self-healing architecture, redundancy handled through software � Rapidly growing volumes of data � ….. 9

When is your app ready for current Clouds ? � If there are no issues with licenses, IP, secrecy, sensitive data, privacy, legal or regulatory issues, . . . � If your app is (almost) architecture independent, not optimized for specific architecture (i e single process optimized for specific architecture (i.e. single process, loosely-coupled low-level parallel, I/O-robust) � If it’s just one app and zillions of parameters � If latency and bandwidth are not an issue � If time (wait, wall, run) doesn’t really matter � If your job is low-priority, simple SLAs, can re-run, . . . � Current cloud “killer” apps � Mobile and web applications � Extensions of desktop software � Matlab, Mathematica � Batch processing / MapReduce � Hadoop at NY Times Characteristics of Clusters, Grids, Clouds (1/ 2) 10

Characteristics of Clusters, Grids, Clouds (2/ 2) Top 10 Obstacles and Oportunities for Adoption and Growth of Cloud Computing UC Berkeley RAD Lab 11

Cloud Computing: A possible Definition “A computing cloud is a set of network enabled on demand IT services, scalable and QoS guaranteed, which could be accessed in a simple and pervasive way .” | Marcel Kunze | OpenCirrus, 23 NeSC Edinburgh | March 2009 24 | 12