Chapter 4 Lecture Presentation
Download
Report
Transcript Chapter 4 Lecture Presentation
CHEETAH's use of DRAGON
Malathi Veeraraghavan
University of Virginia
August 23, 2006
DRAGON software (current usage)
RSVP-TE for an end-host client
VLSR for a CVLSR to support immediate-request
calls
DRAGON network (proposed work)
Run web caches on DRAGON PoP hosts
Interconnect CHEETAH to DRAGON network via IP
routers
1
4. Immediate-request calls
dynamically setup/released distributed bandwidth management
1. Build a scalable GMPLS network
CHEETAH network
6. Web caching: Breaks up long TCP connection into
two short TCP connections + one wide-area
CHEETAH circuit
Raleigh PoP (MCNC)
SN16000
via NCREN
OC192 Control GbE/
card
card
10GbE
card
UVa
CUNY
NCSU
End hosts
2. Circuit-requestor code
RSVP-TE
NLR/SLR/NCREN)
3. Sycamore GMPLS implementation OC-192 (viaimplements
1Gb/s circuit
ORNL PoP
SN16000
End hosts
Atlanta PoP (SOX/SLR)
SN16000
GbE/ Control
OC192
10GbE card
card
card
OC192 Control GbE/
10GbE
card
card
card
End hosts
OC-192
ORNL
5. Applications: eScience + commercial: web
caching, video & storage
GaTech
2
Circuit/Virtual-Circuit technologies
eScience networks Commercial networks
Raison
d'etre
High-throughput
connectivity
between a few
facilities
Moderate-throughput
connectivity between
millions of users
Key
goal
QoS-guaranteed
connectivity
Bandwidth sharing
(different style from
TCP based sharing)
Call
Long-held
duration
Short-duration
Reach
Partial (HOV lanes)
End-to-end
3
My interest in moderate-BW applications/IR mode:
SCALABILITY; Metcalfe's law re. value of networks
m is the link capacity
expressed in channels
e.g., if 1Gbps circuits
are assigned on a 10Gbps link,
m = 10
Bandwidth sharing modes
Large m
Moderate throughput
Small m
immediate-request
with call blocking + retries
("call queueing")
Short calls
Bank teller
(video, gaming)
immediate-request
with delayed-start times
(file transfers)
High throughput
Long calls
Doctor's office
book-ahead
Mean waiting time is proportional to mean call holding time
Can afford to have a queueing based solution if calls are short
4
Proposed DRAGON use for web
caching
Would like to run squid web caching software on
DRAGON and HOPI PoP hosts
First step:
Next step:
Preconfigure VLANs and test web caching
Run CVLSR or VLSR + CSPF PCE in hop-by-hop mode for
short-lived, moderate-BW VLANs across DRAGON
Call setup delay with SNMP: 149ms; CLI adds 80ms
(open SecureCRT) CUNY report
Demo/test:
Use circuits between web caches through just the
DRAGON network
5
CHEETAH evolution: add IP routers with channelized
OCxx cards to support sub-Gb/s circuits
UVa
GbE
Raleigh PoP
SN16000
CUNY
GbE
GbE
OC192 Control OC192
card
card
card
NCSU
GbE
End hosts
OC-192 (via NLR/SLR/NCREN)
ORNL PoP
SN16000
GbE
End hosts
Atlanta PoP
SN16000
OC192 Control OC192
card
card
card
OC-192
GbE
ORNL
GbE
OC192 Control OC192
card
card
card
End hosts
GbE
GaTech
6
Proposed interconnection of CHEETAH to
DRAGON & HOPI - via IP routers
DRAGON
Web
cache
VLAN Web
cache
Web
cache
CLPK
UMD
DCGW
Web
cache
MCLN
DCNE
ARLG
Web
cache
ISI East
VLAN
McLean, VA
Web
cache
Web
cache
MPLS
NC
10GbE
Web
cache
CHEETAH
SONET
HOPI network: courtesy, Rick Summerhill
DRAGON: courtesy, Jerry Sobieski
Web
cache
TN
Web
cache
7
GA
Opportunity missed if the whole optical
testbed community only experiments with BA
What opportunity?
Enable the creation of large-scale circuit/VC networks
with moderate-rate circuits that can support a brand
new class of applications
economic value for the telcom industry
A "reservations-oriented" mode of networking to
complement today's connectionless Internet
ala airlines that complement roadways
Could prove useful to FIND, GENI, net-neutrality
Alternative pricing models for bandwidth
8
What "brand new class of applications?"
Video, video, video:
Video-telephony, conferencing, elearning,
entertainment, surveillance
Gaming
Remote software access + Sync. storage
Async storage
Multimedia (large) files in web sites
9