Upload
june-casey
View
219
Download
0
Tags:
Embed Size (px)
Citation preview
Observation
> Dennis Jennings – original director for NSFnet> The original Internet was neither pre-ordained or a given> Internet was fortunate series of accidental events> We could have easily ended up with X.25 or ATM has the
foundation of the Internet as we know it today> Large part of its success was the telcos ignored Internet
development for the first 20 years– Still ignoring it
> Will a 20 year pure research program result in a better Internet?> Future Internet has to be anchored with practical
implementation issues and business models> Critical role for NRENs & GLIF
NSF/OECD workshop – Jan 31
> In terms of the Future of the Internet there is a current window of opportunity to consider economic, social and regulatory issues in relation to work being undertaken under the umbrella of the Global Environment for Network Innovations (GENI) and the OECD Committee on information and communications policy (ICCP).
> The NSF/OECD Workshop will bring together who’s who of economists, policy-makers, social scientists and technologists, to consider a broad range of factors which have relevance for the future of the Internet.
> In considering a new infrastructure, such as GENI, it is important to highlight the Internet’s growing role as a driver of innovation leading to economic growth and social well-being. Specifically, Workshop participants are asked to draw lessons from the applications and use associated with the evolution of the current Internet and to identify the features that have been critical to the Internet’s success.
> At the same time, the Internet faces many challenges which are not only related to its technical limitations and it is increasingly clear that many of the problems the Internet has encountered concern economic and social issues.
Social & Economic Factors Shaping the Future of the Internet
The research challenge
> The Internet is a tremendous success, but…> Can we meet tomorrow’s needs by incremental improvement of
today’s design? > NSF and its research community have concluded that they
must take a leadership position with respect to revolutionary network research, and must provide suitable infrastructure for this research.
Source: David Clark MIT
Next Generation Internet approaches
> Incremental– Next Generation Internet will be achieved through small
incremental improvements to existing Internet– E.g. Phosphorous, NGI
> Federated– Next Generation Internet will be a federation of different
networks with different architectures and protocols– E.g. CA*net 4, NLR, SURFnet
> Big Bang– Next Generation Internet will have to be a complete fresh start– E.g. NewARCH, GENI
Next Generation Internet Initiatives
> USA – NSF– FIND – Future Internet Directions - $40m PI driven research program– GENI – Global Experimental Network Initiative – proposed $350m
infrastructure to support next generation Internet research > USA DARPA
– NewARCH – New Internet Architecture – CORONET - Dynamic Multi-Terabit Core Optical Networks:
Architecture, Protocols, Control and Management– CBMANET - Control-Based Mobile Ad-Hoc Networking
> EU– INTERNET OF THE FUTURE – – FIRE -Autonomic networks and self organizing networks
> JAPAN– NGI – Next Generation Internet Initiative
FIND: An NSF challenge question
> 1) What are the requirements for the global network of 10 or 15 years from now, and what should that network look like?
> To conceive the future, it helps to let go of the present:
> 2) How would we re-conceive tomorrow’s global network today, if we could design it from scratch?– This is not change for the sake of change, but a chance to free
our minds.
Source: David Clark MIT
Security and reliability
> Define the objective broadly.– “Classic” security, availability, resilience.
> Hard because:– Many problems are in the end-hosts.– Many problems involve a balance of interests.
• Among actors, states and societies.
• We don’t have agreement about the objective.
– Different contexts call for different answers.– We don’t have a coherent approach.
Economic landscape
> In 1975, it was not clear to the early designers that we were designing the landscape of investment and competition.– Now it is.
> Could we do a better job to shape:– Regulation (or lack of)?– Continued investment and innovation?– Options for user choice?– Deployment of new services?– Health of the value chain?
• Consider the role of facilities providers, for example.
– Role of advertising?
Source: David Clark MIT
Technology drivers
> New network technology.– Usual place to start, but I will get to it later.
> New computing technology.– Whatever computing is, that is what the Internet should support.– The Internet grew up in a stable “PC” time.– The cellular industry evolved independently.– Tomorrow: many different views; sensors, cell phones, embedded
processors, $100 laptops, etc.
> Rich space of services and servers.– Design alternatives will have important influence on personal choice,
control, innovation, etc.
Source: David Clark MIT
Summary--integrative visions
> What are candidate designs of a future global network suited for the needs of 2020?
> What are good design principles for large scale future distributed systems and services?– Alternative approaches will lead to different outcomes in the
socio-technical space.
Van Jacobson
> Content Centric Networking> Move from channels to platforms
– The old Internet and telecom was focused on setting up communication channels to distribute data – routers, switches, paths, etc
– But data should not have locality
> Convergence of cyber-infrastructure and next generation Internet
> Integration of Grids, Web 2.0, SOA, P2P, Enterprise 2.0 and NGI
> Possible significant impact on new business models
ISOC – Internet 2020
> Not “Internet 2” too or similar incremental initiatives> Radical transformative network originating from
research community> Pervasive Information systems> Terabit to the desktop
Internet 2020- What will be there
> Optical of course– Terabits to the Desktop
> Hybrid networks– Generic packet based IP-network and– Dynamically established high speed end-to-end optical circuits
• Bandwidth on Demand (BoD)
• Lambda Grids
Solutions?
> Require interdisciplinary research> Long term research, not restricted to 4 years
– More like 10 to 15 years
> Meanwhile we see short term partial fixes
– E.g. NREN’s move towards fully owned optical hybrid networks
• Works within NREN
> Commercial providers work on local solutions
– Network Neutrality debate
– Walled garden approaches or
– Verticalization
GENI: Vision & Implementation
A shared facility that allows: > Concurrent exploration of a broad range of experimental networks
and distributed services> Interconnection among these experimental networks and with the
Internet> Real users utilizing experimental services> Observation, measurement, and recording of the resulting
experimental outcomes
> Funding for new router architecture already underdevelopment> Third R&E network in the USA
Source: Peter Freeman NSF
GENI Facility Conceptual Design
Slicing, Virtualization, Programmability
Mobile Wireless Network
Sensor Network
Edge Site
Source: Peter Freeman NSF
EU Internet of the FutureNetwork & service infrastructures
Rationale– Networks and service infrastructures underpin economic progress and the
development of our societies • 2 billion mobile terminals in commercial operation, 1 billion Internet users, 400
million internet enabled devices• Determining factor in business development in all sectors
– Of a growing and changing demand• For more, and more user control of content/services, for interconnecting «things »,
for convergence in networks, services and devices
– Current technologies can be, and need to be improved significantly• for scaling up, for more flexibility, for more security and dependability
– Europe is in a leading position: Industry, technology and use• Networks equipment and services, business software, security, GRIDs technology
Source: Fabrizio Sestini EU-Canada workshop
Limits of Internet architecture and IP, current solutions
> Scalability – devices (users AND nodes, mostly wireless) will outnumber humans by several orders of magnitude
– Not only numbering: Service-centric networks invalidate the source-destination approach of the Internet (semantic addressing, space-aware models…)
– Need to overcome the intrinsic limitations of a centralised control (self-management, self-organisation), multidisciplinary approaches to cope with complexity (e.g. bio-inspired)
– Multiple points of attachments to the infrastructure, edge diversity
– Routing
> Suitability for ad-hoc/multi-hop/mesh networking– disconnected operation is common (opportunistic, trust and privacy, …)
> Mobility– Mobile IP (or MIPv6) not effective, overhead
> Transparency– QoS???
– Web of trust?
– Usability
> Security– How to embed it, how to cope with heterogeneous networks
> Machine-to-machine communication
patches
patches
patches
patches
patches
patchespatchespatches
patchespatches
patches
patches
Source: Fabrizio Sestini EU-Canada workshop
New Architectures
Situated Services
BIONETSBIONETSAutonomic service evolution
HAGGLEHAGGLEOpportunistic networking (cross-layer)
Common research issues:Security, resilience, self-* (organisation,
evolution,healing, …) interaction of new
paradigms with society
ANAANA Beyond IP
self-org.
Autonomic communication elements
CASCADASCASCADAS
FIRE-baselineFIRE-baseline
A Testbed for Emergence of InnovationUnique NG Router
> Lego Design> Off-the-Shelf Blades> Custom FPGA Blade *> National Scale Network
through CANARIE uclp> Multiple Virtual Networks> Large-Scale Access
Computing
GE/10GE
Infiniband
CANARIE
P
Access through IP
IP
FPGA*
Source: Albert-Leon Garcia University of Toronto
Problem: You can’t make money from broadband
> Broadband networks, whether wireless or FTTx, overbuilders, municipal or customer owned, requires huge capital outlay with a big risk of slow take-up. In addition:
– subject to intense competition from incumbents– High churn– Low takeup – “Tyranny of the takeup”
> Consumers unwilling to spend more than $40/mo for broadband– Generally prefer lower cost, rather than higher bandwidth
• http://news.com.com/DSL+strikes+a+chord+with+frugal+shoppers/2100-1034_3-6084717.html
– Margins are very thin, even for incumbents> Many companies starting to offer free broadband, both wired and
wireless– Inuk, Sky, TalkTalk, Google, Microsoft, Cable and Wireless etc
> The real winners in the broadband game are those providing advanced services
– Google, Inuk, Yahoo, Skype, Joost, etc– CBS is going to deliver prime time TV over Internet this fall – one day ahead of
broadcast TV
Net Neutrality – CALEA- etc
> Universities and Telephone companies face the same problem of small number of heavy users consuming expensive Internet bandwidth
> University solution is to cap bandwidth from dormitories and/or block types of traffic
> Telecom solution is to build a two tiered Internet or doing volume capping– A high speed un-congested channel for the telco traffic particularly
aimed at carrying video> Universities & NRENs can play a leadership role in piloting new
last mile (hundred feet) architectures that address problems of dormitories – May serve as possible model for telcos
> University students are ideal early adopters and were instrumental in diffusion of the Internet throughout larger community
One possible solution
> Following is example of one possible solution– There may be others- this is not intended to be definitive or exclusive
> Work with a few universities & NRENs on a small number of pilots where interested students can lease or control dedicated fiber/copper to university colo point
> They can directly peer with other students in the dormitory across a “white light” switch or user controlled VLAN switch; and/or
> Connect to service providers of their and/or setup point to point user controlled VLANs to other students across NLR, CA*net 4, GLORIAD, GLIF, SURFnet, i2Cat, KREOnet, etc
> Primary application would be collaborative video such as Inuk, Joost, YouTube and/or CineGrid
Advantages for student
> One time very small cost for UNLIMITED bandwidth forever to university colo
> Cross connect to service provider of their choice or research network(s)
> NO or very low monthly Internet service fees for connection to content providers or connection across research networks
> Participate in new global collaborative models such as Inuk, YouTube.com, MySpace, Joost, Inuk or CineGrid
> Direct connection to content and application providers> Student installs transceiver or simple media hub at their
computer– Media hub with CWDM for about $200 which includes laser, Gbe
transceiver etc
Inuk Networks www.inuknetworks.com
> Offering free triple play to university dormitories in the UK
> Free cable TV, telephony> Over the air channels from around the world> Also deliver university content to cable systems and
other institutions around the world> Make money by selling eyeballs> Joost has same strategy
Ottawa backbone fiber
Student Empowered Network
University NeutrolColo switch/RPON
University A
University B
University C
University Neutral Colo switch/RPON
Dark Fiber
ORAN- CA*net 4 Lighpaths
CA*net 4UCLP switch
CA*net 4UCLP switch
UniversityResidence University
Residence
YouTube orCineGridserver
Internet transit provider
Internet transit or content provider
UCLP Virtual RouterUnder control of students
Conclusions
> GLIF essential for international collaboration to support Internet of the Future research projects– Phosphorous a good example– Next Generation PlanetLab– VINI, GENI?
> NRENs can play important leadership role to anchor research with practical real issues– Most important economic models– New wireless Internet models – picoGSM, WiFi/GSM– Wireless Internet peering
Background material
> http://www.multichannel.com/article/CA6332098.html> it only takes about 10 BitTorrent users bartering files
on a node (of around 500) of traditional shared IP network to double the delays experienced by everybody else.
> http://www.news.utoronto.ca/bin6/060222-2074.asp > University students played critical role in diffusion of
the Internet to the global community