This document discusses cloud computing, virtualization, and the future of these technologies. It provides definitions and explanations of key concepts such as cloud computing, virtualization strategies including OS virtualization, hardware emulation, and paravirtualization. Experts are quoted discussing how cloud computing allows computation to migrate into the network, how virtualization frees up the relationship between software and hardware, and how cloud computing provides on-demand access to computing resources and infrastructure and allows users to pay for only what they use. The document examines issues like server virtualization addressing lack of space in data centers and the benefits of virtualization in improving hardware utilization and reducing costs and energy usage.

1. Cloud compu0ng,
Virtualisa0on
& the Future
Aake Edlund, PhD
KTH, Sweden
Bal1cGrid (Project Director)
Bal1cGrid Innova1on Lab and Bal1cCloud (Ini1ator, Manager)
Pawn Promo1on (CEO), Numeri (Co‐founder)

2. Cloud compu0ng,
Virtualisa0on
& the Future
Paradigm change
“Cloud compu1ng is the ability to migrate the
computa1on that used to happen
at the edges into the network..”
Randy H Katz, Berkeley RAD Labs
See berkeleyclouds.blogspot.com
July 17, 2009 ISSGC09 ‐ Sophia An1polis 2

3. Cloud compu0ng,
Virtualisa0on
& the Future
Paradigm change
“Free up the relationship between
software and hardware and
a whole new industry is born”
Alan Williamson
www.aw20.co.uk
July 17, 2009 ISSGC09 ‐ Sophia An1polis 3

4. Cloud compu0ng,
Virtualisa0on
& the Future
Cloud vs Grid
July 17, 2009 ISSGC09 ‐ Sophia An1polis Google Trends, 1st of July 2009
4

5. Cloud compu0ng,
Virtualisa0on
& the Future
Cloud Compu0ng – short version
Cloud = Virtualization + Automation
1. The cloud is IT infrastructure as a service.
2. The IT infrastructure is delivered as virtual machines.
3. Automation moves those VMs around, thereby
providing the delivery mechanism for the service.
Rachel Chalmers, 451 group, October 20th, 2008
July 17, 2009 ISSGC09 ‐ Sophia An1polis 5

6. Cloud compu0ng,
Virtualisa0on
& the Future
Virtualisa0on
‐ abstrac0on of computer resources
“Not physically exis1ng as such but made by
so^ware to appear to do so”
July 17, 2009 ISSGC09 ‐ Sophia An1polis 6

7. Cloud compu0ng,
Virtualisa0on
& the Future
Virtualisa0on
‐ abstrac0on of computer resources
PlaBorm virtualiza0on – Separates an opera1ng system
from the underlying plaaorm resources
Encapsula0on ‐ The hiding of resource complexity by the
crea1on of a simplified interface
July 17, 2009 ISSGC09 ‐ Sophia An1polis 7

8. Cloud compu0ng,
Virtualisa0on
& the Future
Virtualisa0on
‐ abstrac0on of computer resources
Storage virtualiza0on ‐ The process of completely
abstrac1ng logical storage from physical storage
Network virtualiza0on ‐ Crea1on of a virtualised network
addressing space within or across network subnets
July 17, 2009 ISSGC09 ‐ Sophia An1polis 8

9. Cloud compu0ng,
Virtualisa0on
& the Future
Virtualisa0on
‐ abstrac0on of computer resources
Computer clusters, grid compu1ng, and cloud compu1ng ‐
the combina1on of mul1ple discrete computers into larger
metacomputers
Applica0on virtualiza0on ‐ The hos1ng of individual
applica1ons on alien hardware/so^ware
July 17, 2009 ISSGC09 ‐ Sophia An1polis 9

10. Cloud compu0ng,
Virtualisa0on
& the Future
Virtualisa0on – why?
Hardware is underu1lized – Moore’s law
Data centers run out of space
Economy and Environment – Energy cost explosion
System administra1on cost…
July 17, 2009 ISSGC09 ‐ Sophia An1polis 10

11. Cloud compu0ng,
Server virtualisa0on is geHng
Virtualisa0on cri0cal – we’re running out of
& the Future
room in our data centers
July 17, 2009 ISSGC09 ‐ Sophia An1polis 11

12. Cloud compu0ng,
Virtualisa0on
& the Future
Three main strategies
OS virtualisa1on
Hardware emula1on
Paravirtualisa1on
July 17, 2009 ISSGC09 ‐ Sophia An1polis 12

13. OS virtualisa0on
Virtual Virtual Virtual
environment 1 environment 2 environment 3
Installed and runs on top of an
Applica1on
so^ware
Applica1on
so^ware
Applica1on
so^ware exis1ng host OS, and provides a set
System so^ware
System libraries
System so^ware
System libraries
System so^ware
System libraries
of libraries that applica1ons
interact with.
OS | Container Virtualisa1on Layer
Host Opera1ng System
Hardware
Efficient, imposes liile overhead,
Network many containers on one piece of
HW possible.
Good for large number
of homogeneous OS,
Only one OS, same as host OS.
e.g. web hos1ng.
July 17, 2009 ISSGC09 ‐ Sophia An1polis 13

14. Hardware emula0on
OS installed on a virtual machine
Applica1on Applica1on
that emulates the hardware that
OS OS the OS usually interact with. VVM
coordinates access between guest
CPU, CPU,
Memory,
NIC, Disk
Memory,
NIC, Disk
VMs and the actual underlying
hardware – in run1me.
VMware Virtualisa1on Layer
x86 Architecture Unmodified guest OSes. Supports
Device driver issues/ dissimilar OSes.
inflexibility: could limit
the usability. Performance penalty.
July 17, 2009 ISSGC09 ‐ Sophia An1polis 14

15. Paravirtualisa0on
Does not create an entire VM to host
the guest OS, rather enables the
Xen Guest Guest
guest OS to interact directly with the
Device
Tool Drivers OS OS hypervisor.
Stack
Less performance overhead.
Xen Hypercall API
Xen Hypervisor Uses the device drivers contained in
one of the guest operating systems –
Hardware i.e. not limited to the drivers (as in
hardware emulation) contained in the
virtualisation software.
July 17, 2009 ISSGC09 ‐ Sophia An1polis 15

16. Cloud compu0ng,
Virtualisa0on
& the Future
Again:
Cloud Compu0ng – short version
Cloud = Virtualization + Automation
1. The cloud is IT infrastructure as a service.
2. The IT infrastructure is delivered as virtual machines.
3. Automation moves those VMs around, thereby
providing the delivery mechanism for the service.
Rachel Chalmers, 451 group, October 20th, 2008
July 17, 2009 ISSGC09 ‐ Sophia An1polis 16

17. Cloud Compu0ng
– longer version
“Cloud Computing refers to both the applications
delivered as services over the Internet and the
hardware and systems software in the
datacenters that provide those services…
The datacenter hardware and software is what
we will call a Cloud…
Cloud computing has the following characteristics
1. The illusion of infinite computing resources..
2. The elimination of an up-front commitment by Cloud users..
3. The ability to pay for use … as needed…”
July 17, 2009 ISSGC09 ‐ Sophia An1polis 17
UC Berkeley RAD Labs

18. Cloud compu0ng
– once more…
1. The illusion of infinite computing resources..
2. The elimination of an up-front commitment by Cloud users..
3. The ability to pay for use … as needed…”
a) Signup for an AWS account
Use your existing Amazon.com account if you want
b) Register a credit card
Billed on the 1st of every month for usage prior
c) http://aws.amazon.com/ec2
July 17, 2009 ISSGC09 ‐ Sophia An1polis 18

19. Spectrum of Clouds
• Instruction Set VM (Amazon EC2)
• Bytecode VM (Microsoft Azure)
• Framework VM (Google App Engine)
Lower-level, Higher-level,
Less management More management
EC2 Azure AppEngine
July 17, 2009
Automatic scalability and failover
ISSGC09 ‐ Sophia An1polis 19

20. Two views
‐ of the same thing
On the right, Sam Johnston’s 6 layer Cloud Computing Stack
http://samj.net/2008/09/taxonomy-6-layer-cloud-computing-stack.html
July 17, 2009 ISSGC09 ‐ Sophia An1polis 20

21. Layer 1: Cloud Infrastructure
‐ Provisioning of compu0ng resources: CPU, Memory, Processing
‐ Basically an “Opera0ng System” on demand
‐ Usually billed on a per‐hour usage model
Players in this space
Amazon EC2, Flexiscale, GoGrid, Joyent, AppNexeus
Management providers: RightScale, Elas0cServer
Opera0onal Differences
‐ Wildly different CPU/Memory specs
‐ Opera0ng System differences
‐ Difficult/Impossible to move whole images between providers
July 17, 2009 ISSGC09 ‐ Sophia An1polis 21

22. Layer 2: Cloud Storage
- Provisioning of data storage: Either file/object based or Database like
functionality.
- Billed on bandwidth and storage consumed
Players in the space: Amazon S3, Nirvanix, Mosso, Amazon’s
SimpleDB, Google’s BigTable, Azure Storage
Management Providers: Jungle Disk, Elephant Disk, PutPlace.com
Issues
- Different types of data storage models
- Limitations on the size of individual data units
- Different billing models makes it hard to do a straight comparison
- Access to the data generally uses non-standard query syntax
- No common API
- Performance issues
July 17, 2009 ISSGC09 ‐ Sophia An1polis 22

23. Layer 3: Cloud PlaBorm
Provides a complete software stack - An IDE for the cloud
Takes care of: Runtimes, Load balancing, Resource provisioning
Players in the space
- Google App Engine - Python (initially, now also Java)
- Force.com (SalesForce)
- Microsoft’s Azure - .NET
- Heroku.com (RubyOnRails)
Issues
Different languages -- Most platforms are unary
Different operational philosophies
- Google App Engine for example doesn’t permit files
Lots of limitations in terms of deployment
Completely reliant on the provider for complete uptime and operation
Widely different billing models
July 17, 2009 ISSGC09 ‐ Sophia An1polis 23

24. Layer 4: Cloud Applica0ons
- Applications that are completely ‘online’
- Operate on data that is stored in the ‘cloud’ or ‘ether’
- No client software generally required
- Billing: Ad. Revenue, Premium Services
Players in this space
- Google Apps - Gmail / Google Docs
- Apple’s MobileMe
- Microsoft’s Live - Hotmail, Live Spaces
- SalesForce.com
Issues
- Near on impossible to move between providers
- GMail to Hotmail requires major disruption
- End user focused
o Consumer side of cloud computing
- July 17, 2009
Completely reliant on the provider for complete uptime and operation
ISSGC09 ‐ Sophia An1polis 24

25. Layer 5: Cloud Services
- Provides services to which other applications can utilise
- Specific to vertical markets where most “Web2.0” standards live
- Usually free for non-commercial use
Players in the space
- Google/Yahoo Maps
- PayPal / Google Checkout
- Google / Yahoo WebServices
- Amazon Merchant Services
- Amazon Simple Queue Service
Issues/Comments
- Some “Web2.0” services have attempted a standardization path
- Most however are complete vendor lock-in
- Mashup applications utilize Cloud Service
July 17, 2009 ISSGC09 ‐ Sophia An1polis 25

26. Layer 6: Cloud Clients
- Accessing the cloud
- Clients utilize standard access protocols
o XML
o JSON
o REST / SOAP
Browsers
o FireFox / IE / Chrome
Mobile clients
o Google Android / Symbian / iPhone / J2ME
Desktop Apps
o Google Gears / Adobe AIR / Microsoft Azure
July 17, 2009 ISSGC09 ‐ Sophia An1polis 26

27. Complexity and
Vendor lock‐in
Clients
Services
Increasing
Applications level of
complexity
Platform and
vendor lock-in
Storage
Infrastructure
July 17, 2009 ISSGC09 ‐ Sophia An1polis 27

28. Vendor lock‐in, or the
risk/benefit dance
There are alternatives to Amazon, and the list is growing
Many pricing models are confusing and unclear,
and hard to compare
Many pricing models will change (and are already)
Important to get the big picture of what you need,
today and later on
Vendor locking is definitely a problem to address.
Before you jump into it. But this is always true
July 17, 2009 ISSGC09 ‐ Sophia An1polis 28

29. Cloud Storage
Calculator
hip://www.aw20.co.uk/tools/storagecosts.cfm
July 17, 2009 ISSGC09 ‐ Sophia An1polis 29

30. But is that the
full story?
• Like any good model it is only valid for a
broad range of services
• Many services span different layers at once
– For example
• Amazon / Google / Microso^
• Facebook / MySpace
• YouTube / Flickr
• Ebay
...making standardiza1on even harder
July 17, 2009 ISSGC09 ‐ Sophia An1polis 30

31. Why now?
Experience with very large datacenters
Unprecedented economies of scale
Cost in Cost in
Resource Medium DC Very Large DC Ratio
≈ 1000 servers ≈ 50,000 servers
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
Where Possible Reasons Why
Price per KWH
3.6¢ Idaho Hydroelectric power; not sent long distance
10.0¢ California Electricity transmitted long distance over the
grid; limited transmission lines in Bay Area; no
coal fired electricity allowed in California.
18.0¢
July 17, 2009 Hawaii Must ship fuel to generate electricity
ISSGC09 ‐ Sophia An1polis 31

32. Cloud compu0ng,
Virtualisa0on
& the Future
Why now? Technology is ready
– Pervasive broadband Internet
– Fast x86 virtualiza0on
Virtualiza1on makes it all possible
Virtualiza1on is now built‐in to microprocessors
o Intel’s Virtualiza1on (Intel VT)
o Hardware based, very fast
– Pay‐as‐you‐go billing model
– Standard sogware stack
July 17, 2009 ISSGC09 ‐ Sophia An1polis 32

33. Why now?
Business drive – cloud users
• Flexibility
• Eco‐efficiency
• Credit crunch business impera0ves
‐ CapEx to OpEx – pay‐as‐you‐go
‐ Fixed cost to variable cost
• Improved 0me to market
July 17, 2009 ISSGC09 ‐ Sophia An1polis 33

34. Cloud compu0ng,
Virtualisa0on
& the Future
Pay by use instead of provisioning for peak
Capacity
Resources
Resources
Demand Capacity
Demand
Time Time
Static data center Data center in the cloud
Unused resources
July 17, 2009 ISSGC09 ‐ Sophia An1polis 34

35. Cloud compu0ng,
Virtualisa0on
& the Future
Heavy penalty for under-provisioning
Resources
Capacity
Demand
Resources
Capacity 1 2 3
Time (days)
Demand Lost revenue
1 2 3
Time (days)
Resources
Capacity
Demand
1 2 3
Time (days)
July 17, 2009 ISSGC09 ‐ Sophia An1polis Lost users 35

36. Trends: How is cloud compu0ng
being used?
July 17, 2009 ISSGC09 ‐ Sophia An1polis 36

37. Trends: Cloud providers
July 17, 2009 ISSGC09 ‐ Sophia An1polis 37

38. The devil is in the details…
Your data
in the cloud
July 17, 2009 ISSGC09 ‐ Sophia An1polis 38

39. Cloud compu0ng,
Virtualisa0on
& the Future
Adop0on Challenges
Challenge Opportunity
Availability Multiple providers & DCs
Data lock-in Standardization
Data Confidentiality and Encryption, VLANs,
Auditability Firewalls; Geographical Data
Storage
July 17, 2009 ISSGC09 ‐ Sophia An1polis 39

40. Growth Challenges
Challenge Opportunity
Data transfer bottlenecks FedEx-ing disks, Data Backup/
Archival
Performance Improved VM support, flash
unpredictability memory, scheduling VMs
Scalable storage Invent scalable store
Bugs in large distributed Invent Debugger that relies on
systems Distributed VMs
Scaling quickly Invent Auto-Scaler that relies on
ML; Snapshots
July 17, 2009 ISSGC09 ‐ Sophia An1polis 40

41. Cloud compu0ng,
Virtualisa0on
& the Future
Policy and Business Challenges
Challenge Opportunity
Reputation Fate Sharing Offer reputation-guarding
services like those for email
Software Licensing Pay-for-use licenses; Bulk
use sales
July 17, 2009 ISSGC09 ‐ Sophia An1polis 41

42. Cloud compu0ng,
Virtualisa0on
& the Future
Startups opportuni0es
• Startups and prototyping
• One‐off tasks
• Research at scale
• Cloud Killer Apps: Mobile and web applica0ons
…. more about this in the ‘Future’ sec0on
July 17, 2009 ISSGC09 ‐ Sophia An1polis 42

43. Bal0c Cloud
Bringing cloud compu.ng to
the Bal.c States and Belarus
Bal1cGrid
BC
Planned
BGi ‐ Bal1cGrid Innova1on Lab
ac1vi1es
SA1‐3, SME
Bal1cCloud Courses
NA1‐4, JRA connec1vity
July 17, 2009 ISSGC09 ‐ Sophia An1polis 43

44. Open source 44
“cloud in a box”
Our (main) choice:
- Great team! Rich Wolski, UCSB. BC
- Integrated with
- Integrated with
July 17, 2009 ISSGC09 ‐ Sophia An1polis

45. Cloud pilot
45
– spreading
Bal0cCloud + SE + NO +
DK + FIN + Iceland + …
July 17, 2009 ISSGC09 ‐ Sophia An1polis

46. Almost‐there‐first‐summary
Cloud computing is a new way of using existing
technology, driven by business value for the whole
chain, from the providers to their users and their
user’s users.
At this point in time, cloud computing is best used in
startups and prototyping…
July 17, 2009 ISSGC09 ‐ Sophia An1polis 46

47. from http://geekandpoke.typepad.com
July 17, 2009 ISSGC09 ‐ Sophia An1polis 47

48. Cloud compu0ng,
Virtualisa0on
& the Future
Expecta0ons on 2009
• See more cloud computing being used by startups and in-house
quick prototyping
• See more cloud computing being used in academia – esp. HPC
flavours
• See more open source alternatives, and alternative cloud providers
– also in academia (BalticCloud, SweCloud, NordicCloud)
• Improved solutions addressing the dangers in cloud computing.
• Low expectations on standard APIs (takes time, not sure about the
interest from industry, looks like for grids). I.e. industry will not
help here, this is the competition epicenter.
• Hype, problems, business issues: Successful usage will prove the
value. If it is a paradigm shift, it will be clear this year.
• Plenty to opportunities for everyone!
July 17, 2009 ISSGC09 ‐ Sophia An1polis 48

49. Cloud compu0ng,
Virtualisa0on
& the Future
The startup world is changing
“ Startups can be run so cheaply now (with open‐
source so^ware, cloud compu0ng, and virtual
teams spread across the Web) that many more
”
can achieve profitability without any VC cash.
Paul Graham, Ycombinator
July 17, 2009 ISSGC09 ‐ Sophia An1polis 49

50. Cloud compu0ng,
Virtualisa0on
& the Future
The startup world is changing
“ Imagine what it would do to the VC business if
the next hot company didn’t take VC at all. The
less venture capital there is for new startups, the
”
faster the decoupling will begin
July 17, 2009
ISSGC09 ‐ Sophia An1polis
Paul Graham, Ycombinator 50

51. Cloud compu0ng,
Virtualisa0on
& the Future
Rethink VCs
“
IPO market is dead & M&A valua1ons are depressed
Fundamental problem: IT and marke1ng has became so
cheap that VCs may not be needed any more
Angels to take care of earlier stage, VCs move back to
”
later stage
e.g. Allan Mar1nson, MTVP
July 17, 2009 ISSGC09 ‐ Sophia An1polis 51

52. Cloud compu0ng,
Virtualisa0on
& the Future
Cloud compu0ng + Microstartups =
True love!
Bootstrapping – more than ever
Quick start – really quick!
New direc1ons, prototyping, customer projects
Lower your own internal IT costs
Quick to start, agile
“VC = VISA Card”
July 17, 2009 ISSGC09 ‐ Sophia An1polis 52

53. Cloud compu0ng,
Virtualisa0on
& the Future
Future ‐ Fortune 1M
“
Our vision is to enable one person to invent and run the next
revolutionary IT service, operationally expressing a new
business idea as a multi-million-user service over the course of
a long weekend. By doing so we hope to enable an Internet
”
"Fortune 1 million”.
RAD Lab, Berkeley
July 17, 2009 ISSGC09 ‐ Sophia An1polis 53

54. Cloud compu0ng,
Virtualisa0on
& the Future
Enter Microstartups!
“ The zero cost startup has led to the age of the “microstartup.”
It’s no longer two folks in a garage hoping to build a prototype
in order to land a huge VC round, then geAng millions of
dollars to build out an office. Microstartups are sustainable
from prototype to launch and on to a core user base, all for
around $5‐10,000 in costs.
ISSGC09 ‐ Sophia An1polis
July 17, 2009
”
Jason (Calacanis.com) 54

55. STARTUP
@ BGi
Bal0cGrid Innova0on Lab (BGi)
– the star0ng point for early stage startups in the Bal0cs
Scalability – How to build – technology and business aspects
Course ‐ Learning how to leverage on cloud compu1ng
Course – “Startup school”
Networking – alumni, investors, excellence centers, industry
Mentors – finding and follow up
Prototyping resources ‐ the whole Bal1cCloud to “play with”
July 17, 2009 ISSGC09 ‐ Sophia An1polis 55

56. STARTUP
@ BGi
Try it! You don’t have to be in the
Bal0cs
BGi is a unique star0ng point for microstartups
Want to know more?
ake.edlund@gmail.com
July 17, 2009 ISSGC09 ‐ Sophia An1polis 56

57. Acknowledgement and references
Alan Williamson, www.aw20.co.uk
Berkeley RAD lab, berkeleyclouds.blogspot.com
Baltic Cloud team, cloud.balticgrid.eu
Rich Wolski
and his Eucalyptus team, eucalyptus.cs.ucsb.edu
Dan Reed, Microsoft,
www.hpcdan.org/reeds_ruminations/microsoft
451 group, www.451group.com
July 17, 2009 ISSGC09 ‐ Sophia An1polis 57

58. STARTUPS
Links
CLOUDS
Clouds:
BGi and Bal0cCloud: berkeleyclouds.blogspot.com
ake.edlund@gmail.com
ilja.livenson@gmail.com
cloud.bal1cgrid.eu Startups, and Microstartups:
www.infochachkie.com
blog.guykawasaki.com
microstartups.blogspot.com
July 17, 2009 ISSGC09 ‐ Sophia An1polis 58

59. STARTUP
@ UT
SCHOOL
Final slide – yes, you made it!
July 17, 2009 ISSGC09 ‐ Sophia An1polis 59

60. Thanks!
July 17, 2009 ISSGC09 ‐ Sophia An1polis 60