SC11: Advanced Research Networks - Rod Wilson

© Ciena Confidential and Proprietary

Advances in High

Performance Research

Networks

Rodney G. Wilson

Sr. Director External Research

SC11, Seattle Washington

2 © Ciena Confidential and Proprietary

We help you unlock your network’s full

potential…

and make collaboration happen


Ciena’s Solutions

Converged

Optical Ethernet

infrastructure

World-leading

coherent

technology

featuring

40G/100G optical

solutions

Cloud / data center

Smart grid

High-frequency

trading

Security/encryption

Cost /

performance-

optimized LTE /

mobile broadband

solutions


Research and Experimental Networks

4

YOU ARE HERE: Research and

Experimental Networks

Partnership and Collaboration Program

Advanced Technology Research

CTO Group

Ciena has a remarkable history of research innovation with the High

Performance Computing and networking community.

The showcase venue for some of our breakthrough continues to be

SuperComputing SCxx

Pragmatists Early Majority

Innovators Pragmatists Late Majority

Laggards

Sunset, M/D Early Adopters

RGWILSON 2005

Source: Chasm Group

Time

Adoptio

n


The R&E Community – Ciena’s Interests

Leading indicator of future network demands, challenges and solutions. History tells us that (often) what R&E does today is what mass markets do tomorrow.

Proving / trial ground and visibility into adoption of next-generation network technologies (100G, Dynamic Circuit Networking, Application aware networking, IPv6, Cloud )

Ciena participation & contribution to R&E community to advanced network technology and architecture research understanding. We do proofs, test beds, experiments and demos together.

Many R&E networks are our customers. This is a vital market segment for Ciena


Examples of past adventures

Some here were long term joint activities

Some were “proof of concept”

Some were demonstration networks

Some were product trials

…some worked well, some less well.




C

BW requirements

# u s e r s

C

A

B

ADSL GigE

A. Lightweight users, browsing, mailing, home use

Need full Internet routing, one to many

B. Business applications, multicast, streaming, VPN’s, mostly LAN

Need VPN services and full Internet routing, several to several + uplink

C. Scientific applications, distributed data processing, all sorts of grids

Need very fat pipes, limited multiple Virtual Organizations, few to few, p2p

ΣA ≈ 20 Gb/s

ΣB ≈ 30 Gb/s

ΣC >> 100 Gb/s

Source: C. deLaat

University of Amsterdam circa 2003


DRAC, and AAA experiments

Pioneered by Franco Travostino,

with many other contributors

Dynamic Resource Allocation Control


Amsterdam

Netherlight

Source Data

Linux Cluster

SARA Source Data

Linux cluster

Electronic Visualization

Lab @ UIC

Tile Display Visualization

University of

Illinois Chicago

Starlight Network

Visualization

Cluster

Source Data

Linux Cluster

Ottawa

Research Lab OME

OME

OME

OME

San Diego

UCSD – CALIT2

12xGigE 4xGigE

4xGigE

4xGigE

Dynamic Resource Allocation Control


Scinet Photonic equipment

100G Infrastructure

100Gb/s long haul WAN

In the “production” Network

http://sc08.supercomp.org/index.php


13

1

VM Traffic

Control

Services Services

VM

data

AAA

DRAC DRAC

AAA AAA

DRAC

VM

data

VM

Search Apps

Virtual Compute Plane

Multi-domain, Network Service Plane

Virtual Compute Plane

Multi-resource Co-ordination Plane User Plane

Application

Rendering

Tampa Amsterdam Seoul VM

Search Apps

VM

Search Apps

Cross-section: Virtualized User, Compute & Network Planes

2

3


The SC06 Demonstrator,

sensor + cpu + data + net virtualization and control

DataCenter

@Tampa

SC|2006

Nortel’s Sensor Services Platform

Korea

KREOnet

Netherlight

DRAC Controlled Lightpaths

Internal/External Sensor Webs

Amsterdam

Computation at the Right Place & Time! Virtual Machines migration, unbeknownst to applications

and clients, for data affinity, load balancing, or power management


2008: 100G @ Supercomputing ( > 8Tb/s per fiber)

Single Carrier 100G @ 50Ghz Spacing

Full C-Band Tunable

Asynchronous Multiplexing 10x10G

Any 10G Client on Any Port

100G 8dB+ Enhanced FEC

Dual Polarization

QPSK Modulation Format

Low Latency

Commercial Form Factor

1 Petabyte of data transfer in 12hours on 1 carrier

+


CVS 2010 Terrestrial Demo Content & Applications – 100G

Chicago

Ottawa MEN LAB 10

100G Connectivity from Ciena

Ottawa to Starlight Gigapop in

Chicago.

Ability to show simultaneous

data-flows sourced from major

collaborators.

MEN LAB 10 Tile Display

Canadian Brain Imaging

Network

CBRAIN portal & Metadata

Montreal

http://www.startap.net/starlight/index.html

http://www.cbrain.mcgill.ca/


HPN enabled research….

http://www.youtube.com/watch?v=HlRTUYUf1Ew&feature=player_embedded

C-Brain Project

http://www.youtube.com/watch?v=HlRTUYUf1Ew&feature=player_embedded

http://www.cbrain.mcgill.ca/


Flight Case

A micro “data centre”

Containing:

Dell Server w/ SSD drive array

Mellanox 40GE nic

48v rectifier

Ciena 6500 transport system

40GE tribs & 40G trunk


Content Source:

Amsterdam

Alt Source: CERN

Termination1:

Copenhagen

Alt Termination:

Stockholm

Topology for

Terena demo May

16 2011

Proved: 40G over

Alien waves CERN

to Stockholm

Enabled: University

of Amsterdam

protocol research

(Ultra long haul

TCP / UDP

comparisons. Torn

down for economic

not technical

reasons)

Future Experiment termination:

Iceland


Latency minimization for new protocols and FTP


..and today


SC11 – Ciena demonstrations & Experiments

High Performance Content Delivery

• 100G network powered data transfer

• Collaborative tool allows rapid,

remote movement of high energy

physics for research velocity and scientific discovery

CANARIE

100Gbps

~ 212 km

U of Victoria

Computing Center

10x10Gbps

Caltech/CACR

Booth 1223

BCNet

SCInet

HPC SCinet Backbone Network

• 100Gbps backbone – 100GE clients

• Utilizes Internet2 and ESnet

• Ciena 5410 10GE switch & 6500 powered 100G

network

SCInet

Internet2

Booth 1327


SC11 – Ciena demonstrations & experiments

Demonstrations at Ciena - booth 635

Optimized L0-1-2 Content Delivery

• Remote photonic and ethernet

level optimized content delivery

• Ciena 6500 provides colourless

ROADM, 40G waveselect optics

and 40G Ethernet transport

40G Ethernet

Servers

40G wavelength

muxed

Ciena Booth 635

Dutch Research Consortium

Booth 642

Servers

Using 100G for Peta-scale Science NASA Booth 615, LAC/iCAIR Booth 2615

NASA GSFC- MD

MAX 100GBE 100Gbps

~700 mi 100GE

CHICAGO

Ciena Booth 635

NASA

Booth 615

LAC / iCAIR

Booth 2615

Scinet WAN

100Gbps

100GBE

10GBE

100GBE 100GBE

10GBE N x 10GE

100GE

• Ciena 5410 10GE OTN & 100GE switch,

6500 powered 100G network

http://www.google.com/imgres?imgurl=http://seabass.gsfc.nasa.gov/images/NASA_Logo.gif&imgrefurl=http://seabass.gsfc.nasa.gov/&usg=__wbAdAE8Qkma-7S_KTbgRQkmygH8=&h=857&w=1005&sz=44&hl=en&start=1&zoom=1&tbnid=m0YhM3vHjtBOYM:&tbnh=127&tbnw=149&prev=/images?q=NASA&um=1&hl=en&sa=N&tbs=isch:1&um=1&itbs=1

http://www.startap.net/starlight/index.html


Using 100G for Particle Physics research


Why ? • Chosen by World’s Top Providers

• Deployed in the world’s largest

mesh network

• Industry’s richest Control Plane;

refined over 11 years

• Scalable today to 800+ nodes

• Industry’s best and most complete

portfolio of OTN transport and

switching platforms

• Seamless portfolio interworking for

SONET/SDH/OTN & Packet switching

• Coherent Optical Processing

• Rich network design tools

• Agile Photonic Networking

• Unmatched scalability



Back – up slides… topics to be developed


GLOBAL LAMBDA INTERCHANGE FACILITY

(GLIF)


International Test Beds

It’s all about the Future Internet

Entering third phase of Internet

• As profound as the World Wide Web

• The next two to three years will define the Next Generation

Internet

Standards and Business Practices are shaping the Net as much as

or more than Law and Regulation

The Internet Revolution is less than 15% Complete

1. Number of Users

2. Total amount of Content

3. Total Bandwidth

4. Number of Devices

5. Number of Applications

To quote Michael Nelson , Visiting Professor Internet Studies, Georgetown University

And http://www.ic.gc.ca/eic/site/ic1.nsf/eng/04879.html

Building block on how to address this are being put in to Test Bed networks,

Usually to reinforce National Science and Technology Policies and nationalistic

leadership


Application Classes that are dependent upon changes happening in the FI design and are

Futuristic for the 2012-2017 timeframe

•Content Aware Applications

•Tiny to Large Data movement

•Examples include Cinematographic file transfers

•Handling heterogeneous sensor networks eg. Healthcare Vertical

•Addressing challenges in future mobility provision

•Context Awareness

•User Centric applications

•Green ICT

•Mobility

•Virtualisation

Future Internet Application Research


Application Classes that are dependent upon changes happening in the FI

design and are Futuristic for the 2012-2017 timeframe



•Sometimes blended with Content Awareness to provide higher level values

•Location and Proximity in Health Care Scenarios

•Communication Enabled Applications that radically change business performance


•Green ICT

•Mobility

•Virtualisation








•That need co-operation by the FI to empower users

•ID as a Service, SaaS, Knowledge as a Service, ETC.

•Cloud Services

•People and devices with sensors and context awareness

•Green ICT

•Mobility

•Virtualisation








•Green ICT

•Radical rethinking of Data Centre paradigm (eg. Nanodatacentres)

•Application efficiency in size of data sets produced

•Network efficiency in transport and connectivity paradigms

•P2P at higher layers

•Cloud computing to follow sun/ moon to acquire clean/ available energy.

•Mobility

•Virtualisation








•Green ICT

•Radical rethinking of Data Centre paradigm (eg. Nanodatacentres)

•Application efficiency in size of data sets produced

•Network efficiency in transport and connectivity paradigms

•P2P at higher layers

•Cloud computing to follow sun/ moon to acquire clean/ available energy.

•Mobility

•Virtualisation








•Green ICT

•Mobility

•Fundamental paradigm shift required because of explosion of mobile appliances

•Implications on basic mobile network design including protocols, air interfaces and

mobile access structures

•Virtualisation








•Green ICT

•Mobility

•Virtualisation

•Improving delivery of services

•Examples in Healthcare;

•International remote technical support (EG. Aerospace);

•Security of Property;

•Use of Avatars; knowledge bases; Artificial intelligence;

Cross Cutting theme: Security, Privacy and Trace-ability

Source: D. Mann; CANARIE



Next Generation Research Network Challenges

1. Higher Speed Networks 400G & Beyond

Introduction of ultra high speed interfaces (NIC’s & ports)

aggregation, of VC’s and IP, or disaggregation, and dynamic movement of large data flows

How to automatically exploit transport efficiency while maintaining service transparency over

long haul spans?

2. Cloud Architectures

Commercial cloud services evolving, future will see creation of unique specialty clouds

Inter-cloud interactions like movement of virtual computing resources and data storage

How to deliver multi-cloud service and dynamic path provisioning in less than 1 min. (or <1 s)

Impact of network energy efficiency requirements… emergence of Green ICT laws

3. Exascale Computing

Creation of chip to chip optical connections and the resulting high data flows

New network architectures for resulting burst flows (can a big enough fast enough network

buffer be created?)

Internal computer bus architectures and total system choke points.

4. Now that the bandwidth problem is solved

How do you control it ? (measure, manage, change it?)

How do you increase spectral efficiency (because you haven’t solved it for long)

OTN switch it, route it, access it. (Open Optical exchange discussion)


Thank you!