Documente Academic
Documente Profesional
Documente Cultură
Lixia Zhang
UCLA Computer Science Department
February 2011
1
Why this talk
◊ To tell you the basic idea behind NDN
◊ To initiate a discussion
◊ To convince you that NDN represents
the right direction for future Internet
◊ To invite everyone to join the effort
2
Three basic questions
◊ How today’s Internet works
3
How today’s Internet works
◊ TCP/IP architecture
² Defined in RFC791/793 (1981)
◊ Enables any computer on the Internet to
talk to any other computer
² Name boxes
² End-to-end connection
² Datagram delivery
IP
4
30 years down the road
Interconnec4ons
of
computers
A
new
world
of
applica4ons
&
Moore’s
Law
&
compu4ng
devices
silicon
revolu4on
5
1876
2011
◊ Naming boxes got us to where we are today
◊ but no longer fits today’s communication needs
6
What are the problems
◊ Massive scale of data dissemination
◊ Computing devices becoming increasingly
mobile
² Ad hoc networking, disruption-tolerant
networking
◊ Network security
◊ Internet Of Things
◊ Robust data delivery
ISP
ISP
8
Naming location for mobile devices
router
router
router
router
router
router
router
10
Network security: why so difficult?
◊ IP identifies interfaces, networks
◊ Current solutions
² Securing the channel
² Securing the box
² Securing an IP network by firewall
My network
11
Securing perimeters
◊ Firewalls: examining each and all in/out
packets
² But how many entries into your network?
◊ Communication: exchange data across
boundaries
◊ Strong perimeter à barrier to
communication
12
Named-Data Networking
Any
communica4on
media
that
can
provide
best
effort
delivery
Transla4on
PARC.com
Interest(PARC.com/video/foo/s2)
Producer
Consumer
16
Comparison with IP Packet Forwarding
17
NDN Interest Forwarding
get
/parc.com/videos/
WidgetA.mpg/v3/s2
P
/parc.com/videos/WidgetA.mpg/v3/s2 0" 2
data
get
/parc.com/videos/
WidgetA.mpg/v3/s2
18
NDN data retrieval from cache
a/b/c/d
Producer
a/b/c/d
Data
Consumer
a/b/c/d
Producer
Consumer
ISP
ISP
21
ad hoc networking, mobility, DTN
◊ If two or more mobile nodes can
physically reach each other, they can
communicate
22
Securing communications by securing data
◊ NDN makes data the first class
Applications built
directly on top of
named
data
the containers
² Each name associated with a key, data is
signed together with its name at creation
² Data integrity and provenance can be
verified independent from where it comes
23
Design Principle: Hourglass architecture
◊ IP’s “thin waist” has been a key enabler
of the Internet’s explosive growth
◊ NDN keeps the same hourglass-shaped
architecture
² In addition: build security right into the thin
waist
Applications built
directly on top of
NDN data delivery,
named
IP
data
Any communication
media that provide
datagram delivery
24
Design Principle: the end-to-end principle
◊ Internet’s end-to-end principle
² Enables robust applications in face of
network failures
² facilitates support for unforeseeable new
applications
◊ NDN retains and expands the E2E
principle
² End-controlled reliability
² End-to-end security
25
Design Principle:
Routing & forwarding planes separation
◊ Routing-forwarding plane separation:
proven necessary for Internet
development
◊ NDN sticks to the same principle
² Rolling out NDN with the best available
forwarding technology while new routing
system being researched in parallel
26
Design Principle:
No name semantics in the infrastructure
◊ IP’s original design deliberately fixed as
little as possible about how its
'names' (addresses and ports) were
assigned
◊ NDN follows IP's successful strategy
² The only assumption: hierarchically
structured name space
27
New Design Principles
◊ Flow-balanced data delivery is essential to
stable network operation
² IP: open loop datagram delivery
² TCP congestion control was added in later
² NDN designs flow-balance into the thin waist
◊ facilitate user choice and competition
28
Addressing challenges in today’s architecture
h1p://www.named-‐data.net/
32