COMS E 6998 6 Cloud Computing and Data

COMS E 6998 -6 Cloud Computing and Data Center Networking Sambit Sahu http: //www. cs. columbia. edu/~sambits/

Course Scope § Graduate level seminar course on Cloud Computing and Data Center Networking: This is a seminar course on Cloud Computing and Data Center Networking intended to provide students exposure to recent advances in these areas. The target audience is graduate level students with the knowledge of operating systems, computer networks and web services. It is expected that students will read and present research papers and work on a course project. The course project is intended to motivate students to identify and work on engineering as well as research issues in these areas. The first part of the course will focus on broad range of topics in Cloud Computing - starting with overview of technologies that enable Iaa. S Cloud such as Amazon EC 2, provisioning and monitoring, storage cloud, elasticity and resource provisioning, and cloud applications. The second part of the course will focus on the recent advancements on data center networking and Open. Flow based network architecture design. § First part: Cloud Computing – – – – Provisioning and automation Enterprise cloud: moving to cloud Resource scaling Storage cloud Applications and programming models Security and privacy Energy issues § Second part: Data Center Networking – New data center network architecture and design – Open Flow based network architecture and research 2

Requirements from Students § Students will present and critic papers -- 50% grade – Presenter presents the paper – Others need to critic the paper (email me the critic) – There will be standard template for presentation and critic § Course project -- 50% grade – Students may team up – Two choices • Project: come up with a project idea and carry out the project • Compile a report: choose 10 papers in a particular topic and turn in a detailed report § Project Ideas – Discuss with instructor to decide on a course project – Initial list of choices will be provided – Team up to work in groups 3

Course Structure Each week § Background material (if any) covered by instructor § 2 -3 papers will be presented by students § Discussion on possible research issues § Guest speakers on engineering and research issues § Any hands-on experience/demo 4

What is Cloud? § Allows users to request computing resources through web interfaces § Extreme automation through virtualization – Hide away physical resource details – Hide away tedious and error prone configuration issues from users – Provides an utility view to computing resources § Pay as you go – Resources on-demand § Read this paper: http: //www. eecs. berkeley. edu/Pubs/Tech. Rpts/2009/EECS-200928. pdf § Some links: – http: //aws. amazon. com/ – http: //www. eucalyptus. com/ – http: //www. opennebula. org/start – http: //cloudcomputing. sys-con. com/node/770174 5

Different versions of Cloud § Iaa. S – Amazon EC 2 – Open Source: Eucalyptus, Open Nebula § Paa. S – Google. Aps, Azure § Saa. S – …. 6

Why Cloud? § User does not need to own and configure machines – Management of infrastructure left to cloud providers – User needs to only worry about what to do with the machine/resource – not how to prepare it for that purpose § Request resources when needed § Simple web based interface for – request resource – monitor and manage resource § Extreme scaling – Can scale the footprint from 1 server to 1000+ servers in a matter of few minutes or less § Economic model – Rent vs lease – Different models makes sense for different usage pattern – Management cost is often higher than resource cost 7

Life Cycle in a typical Cloud Operation § User of a cloud – APIs to request a machine – Build and load virtual machine images with software stack – Request and attach storage – Attach IP address – Request monitoring capabilities § Cloud administrator – Manage physical resources – Manage virtual resources 8

Some Research Topics in Cloud Computing § Provisioning and automation using virtualization § Enterprise cloud: moving to cloud § Resource scaling § Storage cloud § Applications and programming models § Security and privacy § Energy issues 9

Conference venues § Usenix Hot. Cloud Workshop § ACM SOCC 2010 § Usenix Annual Technical Conference § Sigcomm § NSDI 10

Next Week § Need two volunteers to present the following papers: – Above the Clouds: A Berkeley View of Cloud Computing – http: //www. eecs. berkeley. edu/Pubs/Tech. Rpts/2009/EECS-2009 -28. pdf – Map. Reduce: Simplified Data Processing on Large Clusters http: //labs. google. com/papers/mapreduce. html 11