Business Modelsfor

Dynamically Provisioned Optical Networks

Tal Lavian

DWDMRAM

DWDMRAM

Data@LIGHTspeed

National Transparent OpticalNetwork Consortium

NTONCNTONCDefense Advanced Research

Projects Agency BUSINESS WITHOUT BOUNDARIES

Concept for “Utility” Bandwidth

• Low latency, high bandwidth services (>1Gb/s) are emerging requirements for business, medical, education, government and industry

• New applications development and business models could be stimulated by affordable and easily accessible high bandwidth in both local and wide area networks

• High bandwidth connections are typically full period today but full period 7x24 bandwidth is not always needed.

• Technologies are now available that suggest plausible new business model options to offer time slots for high bandwidth services– Dynamic provisioning of lambda and sub-lambda time slots– Periodically scheduled (N time slots per day, per week) or ad hoc

Application ProfileA – Lightweight users, browsing,

mailing, home use

B – Current business applications, multicast, streaming, VPNs, mostly LAN

C – Emerging business, government, industry & scientific applications, data grids, virtual-presence

Network ProfileA – Internet routing, one to many

B – VPN services on/and full Internet routing, several to several

C – Very fat pipes (both full and non-full period services), limited multiple Virtual Organizations, few to few

BW requirements

# of u

sers

A

ADSL

C

A

B

ADSL GE

User/Bandwidth Profile

Dynamic Wave Provisioning ServiceBusiness Model Examples

• Business Continuity/Disaster Recovery– Remote file storage/back-up– Recovery after equipment or path failure– Alternate site operations due to natural or man-made disaster

• Storage and data on demand– Rapid expansion of network attached storage capacity– Archival storage and retrievals– Logistical networking – pre-fetch and cache

• Financial community and transaction GRIDs– Distributed computational and storage resources– Shared use of very high bandwidth network resources– Utility computing for pay-as-you-go business models

100s of transactions per second

100s of transactions per second

100s of transactions per second

Remote storage /

processing location

Remote storage /

processing location

Remote storage /

processing location

Dynamically provisioned carrier or large

enterprise network – switched lambdas and

sub-lambdas

Core network is a shared resource• SLAs for graduated performance • Fixed time slots

− Lambdas & sub-lambdas• Dynamically allocated

Transaction GRID Demonstration

• Real-time transactions processed and buffered at collection sites for Businesses “BB” & “SRU”

• Periodic transfer to remote site for batch processing using fixed timeslot dynamic lambda provisioning

• High bandwidth/low holding time connection provides periodically scheduled shared use path between collection and remote sites.

Sheridan

PP8600

Federal

PP8600

1 x GE

Taylor

10GE (λ1)

10GE (λ2)

7.2 km

6.7 km 10.3 km

Lakeshore

10GE (λ3)

24.9 km

24.1 km

PhotonicSwitch

OMNInet

PP8600

McCormick Place

Demo Display Station

LakeShoreHost (192.26.85.147/26)

SheridanHost1 (192.26.85.169/26)

SheridanHost2 (192.26.85.170/26)

FederalHost (192.26.85.130/26) GRIDBRICK

Advantage: By having this setup we have contention for Red λ2 between Lakeshore and Sheridan when App A & B try for the same timeslot.

Media Converter/ Local Sw/hub

Logical Demo Layout

100 Mbps

ControlHost (129.105.25.103/24) NRM

(129.105.220.101/24)

ODIN (129.105.220.46/24)

DemoSender1

DemoSender2

DemoReceiver1

DemoReceiver2

DemoControl nrm odinserver

DemoGUI

Demonstration Parameters

All parameters are dynamic and updated in real time

Parameter UnitsTransaction Collection Rate Records/sec

Record size Kbytes

Queue Load Number of records

Queue size Kbytes

Queue fill rate Kbytes/sec

Next queue delivery Date:hour:min:sec

Time to next queue delivery Hour:min:sec

Last delivery average throughput Kbps

DWDM-RAM Architecture Charts

DWDM-RAM Architecture

DataCenter

λ1

λn

λ1

λn

DataCenter

Data-Intensive Applications

NetworkResource Scheduler

Network Resource Service

DataHandlerService

Informa tion S

e rvice

Application MiddlewareLayer

Network ResourceMiddlewareLayer

Connectivity and Fabric Layers

λ OGSI-ification API

NRS Grid Service API

DTS API

Optical path control

DataTransfer Service

Dynamic Lambda, Optical Burst, etc., Grid services

Basic NetworkResource

Service

DWDM-RAM Architecture

Applications

Data Transfer Scheduling

Network Resource Scheduling

Communication Protocols

ODIN

OMNInetFabric

Connectivity

Resource

Collective

Application

DWDM-RAM ArchitectureODIN – Optical Domain Intelligent Network

• Software suite that controls the OMNInet through lower-level API calls

• Designed for high-performance, long-term flow with flexible and fine grained control

• Stateless server, which includes an API to provide path provisioning and monitoring to the higher layers

Applications

Data Transfer Scheduling

Network Resource Scheduling

Communication Protocols

ODIN

OMNInetFabric

Connectivity

Resource

Collective

Application

DWDM-RAM ArchitectureCommunication Protocols

• Standard off-the-shelf communication protocol suites• Provide communication between application clients and

DWDM-RAM services and between DWDM-RAM components

• Communication consists of mainly SOAP messages in HTTP envelopes transported over TCP/IP connections

Applications

Data Transfer Scheduling

Network Resource Scheduling

Communication Protocols

ODIN

OMNInetFabric

Connectivity

Resource

Collective

Application

DWDM-RAM ArchitectureNetwork Resource Scheduling

• Essentially a resource management service• Maintains schedules and provisions resources in accordance

with the schedule

• Provides an OGSI compliant interface to request the optical network resources

Applications

Data Transfer Scheduling

Network Resource Scheduling

Communication Protocols

ODIN

OMNInetFabric

Connectivity

Resource

Collective

Application

DWDM-RAM ArchitectureData Transfer Scheduling

• Direct extension of NRS service, provides an OGSI interface• Shares same backend scheduling engine & resides on same host• Provides a high-level functionality• Allow applications to schedule data transfers without the need

to directly reserve lightpaths• The service also perform the actual data transfer once the

network is allocated

Applications

Data Transfer Scheduling

Network Resource Scheduling

Communication Protocols

ODIN

OMNInetFabric

Connectivity

Resource

Collective

Application

Data Transfer Scheduling

• Uses standard ftp

• Uses NRS to allocate lambdas

• Uses OGSI calls to request network resources

λData Receiver Data Source

FTP client FTP server

DTS NRS

Client App

DWDM-RAM ArchitectureApplications

• Target is data-intensive applications since their requirements make them the perfect costumer for DWDM networks

Applications

Data Transfer Scheduling

Network Resource Scheduling

Communication Protocols

ODIN

OMNInetFabric

Connectivity

Resource

Collective

Application

DWDM-RAM Modes

• Applications may request a data transfer

Applications

Resource Allocator

Network Resource Scheduling

• Applications may request a set of resources through a resource allocator, which will handle the network reservation

Applications

Data Transfer Scheduling

Network Resource Scheduling

The Network Service

• NRS is the key for providing network as a resource– It is a service with an application-level interface– Used for requesting, releasing, and managing the

underlying network resources– Understands the topology of the network– Maintains schedules and provisions resources in

accordance with the schedule– Keeps one scheduling map for each lambda in each

segment

The Network Service

Scheduling maps: Each with a vector of time intervals for keeping the reservations.

Scheduling maps for each segment

The Network Service• NRS

– Provides an OGSI-based interface to network resources– Request parameters

• Network addresses of the hosts to be connected• Window of time for the allocation• Duration of the allocation• Minimum and maximum acceptable bandwidth (future)

– Provides the network resource• On demand• By advance reservation

– Network is requested within a window• Constrained• Under-constrained

The Network Service

• On Demand– Constrained window: right now!– Under-constrained window: ASAP!

• Advance Reservation– Constrained window

• Tight window, fits the transference time closely

– Under-constrained window• Large window, fits the transference time loosely

• Allows flexibility in the scheduling

The Network ServiceUnder-constrained window

• Request for 1/2 hour between 4:00 and 5:30 on Segment D granted to User W at 4:00

• New request from User X for same segment for 1 hour between 3:30 and 5:00

• Reschedule user W to 4:30; user X to 3:30. Both requests are satisfied.

Route allocated for a time slot; new request comes in; 1st route can be rescheduled for a later slot within window to accommodate new request

4:30 5:00 5:304:003:30

W

4:30 5:00 5:304:003:30

X

4:30 5:00 5:304:003:30

WX

0.5s 174s

9. ODIN Server

Processing

10. File transfer

complete, path released

1. File transfer request

8. Path Deallocation

request

7. Data Transfer (20 GB)

4. Path ID returned

3. ODIN Server Processing

2. Path Allocation request

5. Network reconfiguration

6. Transport setup

End-to-end Transfer Time (Un-optimized)

20GB File TransferSet up: 29.7s

Transfer: 174sTear down: 11.3s

Transfer rate: 920Mb/s Effective rate:744Mb/s

3.6s 0.5s 0.3s 11s25s 0.14s

GigE L2 Switch 10GE switched lambdas

For a 200GB file: Transfer rate:920Mb/s Effective rate: 898Mb/s

20GB File Transfer

8.5 x 11 charts

24x36 poster board

4” banner for title or just leave blankDWDMRAM

DWDMRAM

Data@LIGHTspeed

1.5” margin

1.5” margin

¾” margin