1. A Survey of Cloud Computing
Demystifying
the
CLOUD
Typically vague, abstract, nebulous -> clarify
Stories: Most of your computing happens in your browser; NYT converts 15 million stories
A Survey of Cloud Computing
Jesse Dunietz
SASS Talk

2. Cloud service = unified-looking, network-accessible resource pool
Large, homogeneous resource pool
On-demand self-service
Ubiquitous access over network
Room for on-demand increase of resources
Cloud service looks like single, locale-independent entity
Self-monitoring for provider and consumer

3. The central ideas of CC are not new.
The grand vision: J.C.R. Licklider’s “intergalactic computer network” (1963)
Utility computing
“If computers of the kind I have advocated become the computers of the future, then computing may someday be organized as a public utility just as the telephone system is a public utility…The computer utility could become the basis of a new and important industry.” --John McCarthy (1961)
Mainframes  desktop computing  utility/cloud
This was even before the Internet!
Mainframes differ a bit from cloud: not public,

4. Grid computing: an intermediate step
Grid is “a system that:
coordinates resources that are not subject to centralized control
using standard, open, general-purpose protocols and interfaces
to deliver nontrivial qualities of service” (--Ian Foster)
Distributed, parallelised computation
Loosely coupled
Extension of clustering
Computing as accessible as the electric grid
E.g. (BOINC)
When was it created?

5. Grid computing: an intermediate step

6. Recent years have seen an explosion of cloud-like services.
1999: salesforce.com
2002: Amazon Web Services
2005: Zoho Office
2006: Amazon EC2
2007: Gmail/Google Apps
2009: Chrome OS announced

7. Recent years have seen an explosion of cloud-like services.

8. Cloud computing depends on many recent software technologies.
Late 1990’s: widespread broadband Internet
Late 1990’s: OS virtualisation
Xen hypervisor developed at Cambridge CL!
: Intel/AMD offer hardware-assisted virtualization
Server farm management (scaling)

9. “As a service” (XaaS) is the new computing paradigm.
Google Docs,
Photoshop Express
J2EE, Google App Engine
Amazon EC2, GoGrid,
Microsoft Azure
Infrastructure
Software
Platform
IaaS is usually virtual hardware. Real hardware is at the bottom of the stack, but can’t cloudify it.

10. Cloud deployments come in all shapes and sizes.
Private cloud
Public cloud
e.g. Amazon
Community cloud
e.g.
Hybrid
Attribution: Cloud Computing Use Case Discussion Group

11. CC has obstacles, but offers many benefits.
Disadvantages
Privacy
Reliability
Migration
Legal issues
Account security
Lock-in (loss of control)
Advantages
Cost savings
HW/OS independence ([S|I]aaS)
Server security
“Thin” clients for wimpy devices
Update speed

12. New uses for cloud technology are constantly popping up.
Rich web applications
Computing as a utility (IaaS)
New “back office” functionality applications
E.g. Apple Push Notification Service
Anywhere, anytime computing
Talk about GPush

13. Not everyone’s head is in the clouds.
A few treasures from Richard Stallman:
“[The] term [‘cloud computing’] is too nebulous to refer to anything in particular. If it has any meaning, it can only be, ‘Don't pay attention to who controls your data or your computing. Just trust every company.’”
“The issue I've raised is about a more specific kind of scenario: Software as a Service. We can avoid it, and avoiding it is the only way to maintain our freedom.”
Another from Oracle CEO Larry Ellison:
Cloud computing has been defined as “everything that we currently do” and will have no effect except to “change the wording on some of our ads”
Recent outages reduced trust in reliability
Greenpeace has complained about energy usage
Advantage of being at MIT: Richard Stallman’s s
Don’t need to trust everyone if can switch on-demand via open standards
Freedom issue debatable
No change: Seems different in significant ways – XaaS is new as a common paradigm
Reliability: Still an issue; QoS guarantees difficult
Energy: Wikipedia claims

14. CC is still full of unsolved problems.
Interoperability, generic design blueprint
Making sporadic self-healing converge on consistency
Formal models for “loose coupling” vs. brokenness

15. “Everything we think of as a computer today is really just a device that connects to the big computer that we are all collectively building.” -- Tim O’Reilly, CEO, O’Reilly Media