Grid, Mesh,

1
Grid, Mesh, Cloud Computing
2
Agenda

• Grid computing
• Evolution of networks
• Key ideas
• Key emerging applications

3
An Industry Transformed
Delgo www.delgo.com
http//www.boxofficemojo.com/
4
Amazons Grid Foray

• Amazon Elastic Compute Cloud
• aws.amazon.com/ec2

5
The Big Picture

• The Web is a service for sharing information over
  the Internet
• The Grid is a service for sharing computing power
  and data storage capacity over the Internet
• goes well beyond simple communication between
  computers
• global network of computers ? one vast
  computational resource

6
Grids Why Now?

• 4 Drivers
• Demand to do more with less
• But growing infrastructure needs costs
• Tremendous processing needs for complex problems
• Supercomputers cant deal with it
• Much IT infrastructure underutilized (15-20)
• Examples?
• Huge pressure on IT Re applications backlog
• Sudden bursts of demand
• Invest using the estimated peak load (network
  design)
• Not average demand
• /. _at_ iastate.edu
• Design Goal
• Appears to an end user as one large virtual
  computing system

7
Overview Solution

• Model for solving massive computational problems
  by
• using unused resources CPU cycles/disk storage
• of large numbers of disparate computers/administra
  tive domains
• treated as a virtual cluster
• embedded in a distributed telecommunications
  infrastructure
• Focus
• solving problems too big for any single
  supercomputer
• retaining the flexibility to work on multiple
  smaller problems
• caters to intermittent demands of large
  computational exercises
• Examples?

8
The P2P Architecture Inspiration

• Much like a utility
• computing power on demand
• Pay-as-you-go (usage driven)
• Public grids emerging
• Rent your spare CPU cycles

9
Emerging Apps

• Generic
• SETI
• forecasting weather
• modeling nuclear explosions
• analyzing seismic data
• protein folding
• financial modeling
• Commerical
• Pratt Whitneys aerospace airflow modeling (jet
  engines)
• Bristol-Myers Squibb potential drug compounds
• Charles Schwab RD stock transaction speeding

10
Where it Works

• Large applications
• with small data transfer that dont require
  shared memory
• Apps will be delayered into business logic,
  data, and application components
• Transforming basic network application
  architectures
• Different network app design objectives
• Limewires Supernode model

11
Other advantages

• Avoids hardwiring apps
• Avoids large, unchunked IT investments ?ROA
  connection
• Increases org ambidexterity

12
Shrek, Delgo, and Others

• Why did Dreamworks use this?
• Upsides?
• Downsides?

13
Mesh Computing in Action
14
Hype or Reality?

• 40 of the largest firms active use it
• 30 more considering it
• Emerging players
• Microsoft
• IBM e-Business on Demand
• HP
• Sun (network appliance model, Java)
• Globus Toolkit as their core middleware
• Commercial Offerings
• DataSynapse (http//www.datasynapse.com)
• IBM Grid Computing website (http//www.ibm.com/gri
  d)
• Sun Microsystems Grid Computing
  website (http//www.sun.com/grid)
• Oracle Corp. "Oracle Grid"

15
Virtualizing Computing Resources

• Involves sharing heterogeneous resources using
  open stds
• different platforms
• different hardware/software architectures
• different languages
• belonging to different administrative domains
• Cluster computing/ grid computing
• homogenous versus heterogeneous
• confined to data centers vs. encompass user
  desktops

16
What Will Change in Telecom

• Grids operate through
• Network standards
• App layer collaboration
• Will become more important
• Robustness of network key to reliability
• App architectures will shift away from
  client-server
• To what model?

17
Summary