CS294-3: Distributed Service
Architectures in Converged
Networks
Randy H. Katz
Computer Science Division
Electrical Engineering and Computer Science Department
University of California
Berkeley, CA 94720-1776
1
Outline
•
•
•
•
•
•
What is this Course About?
Technology Trends
Evolution of the Internet
Business Trends
Implications and Issues
Summary and Conclusions
2
Outline
•
•
•
•
•
•
What is this Course About?
Technology Trends
Evolution of the Internet
Business Trends
Implications and Issues
Summary and Conclusions
3
Traditional “Networking”
Course
• All about protocols and the OSI seven layers
–
–
–
–
–
–
–
–
Protocol details: link-state vs. distance vector, TCP
Protocol layering
Multiaccess technology
Switching and routing
Naming
Error control
Flow control & scheduling
Special topics like multicast and mobility
4
What is New?
• New things you can do inside the network
• Connecting end-points to “services” with processing
embedded in the network fabric
• Not protocols but “agents,” executing in places in
the network
• Location-aware, data format aware
• Controlled violation of layering necessary!
• Distributed architecture aware of network topology
• No single technical architecture likely to dominate:
think overlays, system of systems
5
Definition of Terms
• Converged Networks
–
–
–
–
Public Switched Telephone Network (PSTN)
Internet/Public Switched Data Network (PSDN)
Mobile Internet
Converged Structure?
• Distributed Service Architecture
– Services
» “-Ility” connectivity
» New call “features”
» Infrastructure services
» Enables distributed applications
6
What is this Course About?
• Emerging, yet still developing, view of a new kind
of communications-oriented service architecture
in a highly heterogeneous environment
– Rapid development/deployment of new services & apps
– Delivered to radically different end devices (phone,
computer, info appliance) over diverse access networks
(PSTN, LAN, Wireless, Cellular, DSL, Cable, Satellite)
– Exploiting Internet-based technology core: clients/server,
applications level routers, TCP/IP protocols, Web/XML
formats
– Beyond traditional “call processing” model: client-proxyserver plus application-level partitioning
– Built upon a new business model being driven by the evolution
of the Internet: traditional “managed” networks and services
versus emerging “overlay” networks and services structured
on top of and outside of the above
7
Course Structure
• Seminar! We learn from each other!
– Avoid traditional lecture-oriented course
– More student-led presentations, discussions
– Project will be the group design and evaluation of a
distributed service architecture
– 2 Units/2 Hours per week
– Every student will develop materials and lead one hour
discussion on a selected area of technology;
to be written up by mid-semester as a “term paper”
– Project: depending on class size, we will have collaborating
and competing teams develop a design and evaluation for
future converged network service architecture
– 20% Term Paper
– 30% Class Discussion/Presentation
– 50% Class Project
8
•
•
•
•
•
•
•
•
•
•
•
•
Relevant Technologies
(Partial List!)
PSTN architecture: AIN, SS-7
SLAs
3GPP/GPRS/Edge
Voice over IP with SIP
Internet Multimedia
Architecture (RTSP, SIP, SAP,
RTP/RTCP, IPv6, IP Mobility,
DiffServ, Multicast)
SyncML
Parlay, JTAPI
WAP
Symbian/Embedded OS
Sun ONE, .Net, Corba, TINA
SMS/MMS + Other Messaging
Platforms
MGCP
• SIP Instant Messaging +
Presence Leveraging
• Mobile Location Services
• WAP
• Radius/Diameter/Single Sign On
• BGP
• MPLS
• Core vs. Access Networking
Technologies
• Mobey Formum
• Operator Wireless LAN
• DRM, PKI
• SCPTP/IETF Sigtran
• Architecture of Internet Data
Centers and NAPs
9
Nokia’s Mobile World View
Network
Environment
Internet
Intranet
2G/3G
Messaging
Browsing
Rich Call
Identity/
Addressing
Interaction
URL
Email
Phone #
10
Nokia’s Mobile Service Matrix
Content
Information
Entertainment
News
Banking & Finance
Buy & Sell
Travel
Music
TV
Lifestyle
Fun
Games
Astrology
Dating
Comms
Messaging
Productivity
Business
Organizers
Intranet &
Extranet
Access
E-Mail
Personal Assistants
FAX
Tools
Rich Call
Misc
Info Mgmt
Enterprise
Comms
VPNs
Telematics
11
Nokia’s Mobile Internet
“Business Architecture”
End
Users
Developers
Content
Providers
Service Providers
Consumption
Connection
Consumption
User Interface
Applications
Rich Call Browsing Msg
Application Gateways
Rich Call Browsing Msg
Applications
Rich Call Browsing Msg
Middleware
Middleware
Middleware
OS
Network
Hardware
OS
Network
Hardware
OS
Network
Hardware
12
Application Layer
Other and 3rd Party Applications
Email App
Browser App
Phone App
Instant Messaging Application
Streaming App
Application Programming Interface
Msg Support Rich Call Support
Browser Support
Application Development Interfaces
Mobile Internet Layer
App Framework and UI Support
App Protocols
SDK Interfaces
Network Interface
SDK Libraries
Network Protocols
Platform Applications Interfaces
Platform Layer
I/O Drivers
OS Core
Link Layer Drivers
I/O Hardware
Processors & Memory
Network Hardware
13
Nokia’s mPlatform Architecture
Operational
Support
Provisioning
Self-Care
Operator Care
Billing Data
Collection
Sys Monitoring
Perf Monitoring
Statistics
Common Enabling Application Functions
Presentation
Manager
Operator
Manager
Audit Manager
(Billing)
Navigation
Manager
Notification
Manager
Security
Manager
Session
Manager
Personalization
Manager
Scenario
Manager
Access Functions
Subscriber
Subscriber
Service
Access Control Authentication Access Control
Process
WAP & PDA
Handling
Proxy
Firewall
Network
Connectivity
Load
Balancing
14
One Operator’s Viewpoint
Pricing
Structure
Customer
Service
HLR
Distribution
Packaging
Marketing & Sales
Billing
Apps
Charging
3rd Party
Apps
Services
Content
Support Systems
Service Domain
Data Center
Data Storage
BS
Radio Access
BS
Network
Spectrum
QoS Cap
Capacity
On-Demand
Processing Cap
Local Radio
Access
Transport
Core Network
Capacity
On-Demand
15
ICEBERG Architecture
Access Network
Plane
PSTN
GSM
IAP IAP IAP
ICEBERG
Network
Plane
Pager
CA
PR
PAC
IAP
APC
NMS
A
SF iPOP
NY iPOP
IAP
SF iPOP
B
NY iPOP
Clearing House
ISP Plane
IAP
ISP1
ISP2
• iPOP: Clustered
computing environ.
• Call Agent: handles
signaling, one per
device per call party
ISP3
• Name Mapping Service: Maps ICEBERG unique ID  service end point
• Preference Registry / Personal Activity Coordinator: user profile / user tracking
• Automatic Path Creation service: creates transcoding datapath between endpoints
16
Possible High Level “Layered”
View of Service Architecture
Prototype Applications: Universal In-Box,
Context-Aware UI, Group Collaboration
Context-Awareness Services: Activity Tracking/Coordination,
Preferences Specification/Interpretation
Adaptation Services: Introspection,
Tacit Information Extraction/Organization
Wide-Area Services: Discovery, Mobility, Trust, Availability
Performance Measurement and Monitoring
“Core” Wide-Area Network
Edge/Access Networks
17
Outline
•
•
•
•
•
•
What is this Course About?
Technology Trends
Evolution of the Internet
Business Trends
Implications and Issues
Summary and Conclusions
18
Technology Trends
• Computing
– Convergence, Divergence, Scale
• Networks
– Internet vs. Telephone Network
– Wireless/Mobile Access
• Services
– E-commerce, M-commerce/M-transactions (iMode),
Content
– Messaging as a major application (P2P, P2M, M2M)
• Architecture
– Integrated (“Closed”) vs. Composed (“Open”) Content,
Distribution, Access Architecture
– Managed vs. Overlay Networks and Services
– Competitive vs. Cooperative Service Providers
19
Internet Growth
120
200
Million
Servers
Million
U.S. Surfers
150
80
100
40
50
01
20
00
20
9
19
9
8
19
9
7
19
9
6
19
9
19
9
01
20
9
19
9
7
19
9
5
3
19
9
19
9
Annual Growth Rate > 50%
5
0
0
Annual Growth Rate > 20%
The Industry Standard, 2 July 2001
20
Convergence?
First Color TV
Broadcast, 1953
HBO Launched,
1972
Telephone,
1876
Interactive TV,
1990
Early Wireless
Phones, 1978
Computer
+ Modem
1957
First PC
Altair,
1974
IBM
PC,
1981
Handheld Portable
Phones, 1990
Apple
Apple
IBM
Mac, Powerbook, Thinkpad,
1984
1990
1992
Eniac, 1947
HP
Palmtop,
1991
Pentium
PC, 1993
Apple
Newton,
1993
Red Herring, 10/99
21
Divergence!
Atari Home
Pong, 1972
Pentium
PC, 1993
Game Consoles
Personal Digital Assistants
Digital VCRs (TiVo, ReplayTV)
Communicators
Smart Telephones
E-Toys (Furby, Aibo)
Network
Computer,
1996
Free
PC, 1999
Sega
Dreamcast,
1999
Internet-enabled
Smart Phones,
1999
Pentium II
PC, 1997
Apple
iMac, 1998
Palm VII
PDA, 1999
Proliferation of diverse
end devices and access networks
Red Herring, 10/99
22
“X-Internet” Beyond the PC
Internet Computers
Internet Users
93
Million
Today’s Internet
407 Million
Automobiles
663 Million
Telephones
1.5 Billion
X-Internet
Electronic Chips
30 Billion
Forrester Research, May 2001
23
“X-Internet” Beyond the PC
Millions
15000
10000
5000
X
Internet
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
20
09
20
10
0
PC
Internet
Year
Forrester Research, May 2001
24
The Shape of Things Now
• Siemens SL45
– A cellular phone with voice
command, voice dialing, intelligent
text for short messages
– An MP3 player & headset
– A digital voice recorder
– Supports “Mobile Internet” with a
built-in WAP Browser
– Can store
» 45 minutes of music
» 5 hours of voice notes
» “Unlimited” addresses/phone
numbers
25
The Shape of Things Now
• Kyocera QCP 6035
– Palm OS/CDMA
– Palm PIM Applications
– Supports “Mobile Internet” with a
built-in WAP Browser
– 8 MBytes
26
The Shape of Things to Come
• Toyota Pod Concept Car
– Co-designed with Sony
– Detects driver’s skill level and
adjust suspension
– Detects driver’s mood (pulse
rate, perspiration),
compensates for road rage and
incorporates a mood meter
(happy vs. angry face)
– Inter-pod wireless LAN to
communicate intentions
between vehicles, such as
passing
– Individual entertainment
stations for each passenger
27
The iMode Story:
It is About Services
• 27M Internet-capable cell phone subscribers (10/01); 50K iMode Web Sites
• World’s largest ISP, first to deploy 3G
“Freedom of Multimedia Access” (FOMA)
• Not just about Japanese teenagers
Applications Used
User Ages
Ring Tone
9
11.5
40.5
17
13.5
8.5
2
Games
Info
Transactions
<20
7
20-24
27
Entertain
Database
Unknown
24
25-29
30-34
8
12
20
35-39
>39
Economist Magazine, 13 Oct 2001
28
After the PC …
True “Convergence”
• Not just about gadgets or access technologies
• About services and applications, and how the
network can best support them
• Increasing, not decreasing, diversity
• Bottlenecks moving from core towards edge
• Enabled by computing embedded in
communications fabric: wide-area, topologyaware, distributed computing
29
What is the Internet?
“It’s the TCP/IP Protocol Stack”
Applications
• Applications
Middleware
Services
“Narrow
Waist”
TCP/IP
Transport Services and
Representation Standards
Network
Technology
Substrate
Access
Technologies
Open Data Network
Bearer Service
Where is the next “narrow waist”?
– Web
– Email
– Video/Audio
• TCP/IP
• Access Technologies
– Ethernet (LAN)
– Wireless (LMDS, WLAN,
Cellular)
– Cable
– ADSL
– Satellite
30
Telephony Evolution
• Mobility/Wireless driving end-to-end digitization of
the telephony system
– Shift towards IP-based infrastructure (Nokia “All-IP”
Architecture)
• Converged Services
– AT&T
» Cell Phone, Telephone, ISP, Video on Demand (Cable)
» Universal Billing Systems
– Sprint: $0.05/min local/long distance, wired/wireless
• Computer-Telephony Integration
– Call Centers, Software-based PBXs, PSTN By-Pass
– Consumer-to-Business E-commerce (e.g., Lands End)
– Speech-Enabled Services (e.g., “Concierge”)
31
Internet vs. Telephone Net
• Strengths
– Intelligence at ends
– Decentralized control
– Operates over heterogeneous
access technologies
• Weaknesses
– No differential service
– Variable performance delay
– New functions difficult to
add since end nodes must be
upgraded
– No trusted infrastructure
• Strengths
– No end-point intelligence
– Heterogeneous devices
– Excellent voice performance
• Weaknesses
– Achieves performance by
overallocating resources
– Difficult to add new services
to “Intelligent Network” due
to complex call model
– Expensive approach for
reliability
32
Cellular Services Most Often
Requested
After basic wireless telephony service
•
•
•
•
•
•
Call Forwarding
Paging
Internet/E-Mail
Traffic/Weather
Conference Calling
News
37%
33%
24%
Data
15% Applications
13%
3%
Source: CTIA Web Page
Peter D. Hart Research Associates, March 1997
33
Services and Applications:
E-Commerce
• Consumer Services
–
–
–
–
–
Consumer-driven QoS: improved Web access “experience”
Converged digital video + web content (e.g., HVML)
Unified billing: pay-per-view movie plus ad-induced pizza purchase
Content delivery: file mover/software upgrades/digital audio/video
Infrastructure storage: back-up, photos, mp3s, videos, TV tapings
• Consumer-to-Business Services
– Web-based + (IP-based) Telephone
– New kinds of integrated call centers: e.g., Lands End
• M-Commerce
– Location-sensitive ad insertion
– Unified billing for telecom access + purchases
34
Outline
•
•
•
•
•
•
What is this Course About?
Technology Trends
Evolution of the Internet
Business Trends
Implications and Issues
Summary and Conclusions
35
The ARPANet
SRI
940
UCSB
IBM 360
IMPs
Utah
PDP 10
UCLA
Sigma 7
• Paul Baran
– RAND Corp, early 1960s
– Communications networks
that would survive a major
enemy attack
• ARPANet: Research
vehicle for “Resource
Sharing Computer
Networks”
BBN team that implemented
the interface message processor
– 2 September 1969: UCLA
first node on the ARPANet
– December 1969: 4 nodes
connected by phone lines
36
37
ARPANet Evolves into Internet
ARPANet
SATNet
PRNet
1965
TCP/IP
1975
Web Hosting
Multiple ISPs
Internet2 Backbone
Internet Exchanges
NSFNet
Deregulation & ISP
Commercialization ASP
AIP
WWW
1985
1995
2005
Application Hosting
ASP: Application Service Provider
AIP: Application Infrastructure
Provider (e-commerce tookit, etc.)
38
39
40
Digex
Backbone
Qwest
IP
Backbone
(Late
1999)
GTE
Internetworking
Backbone
Parallel Backbones
41
42
Network “Cloud”
43
Regional Nets + Backbone
Regional
Net
Regional
Net
Regional
Net
Backbone
Regional
Net
Regional
Net
LAN
LAN
Regional
Net
LAN
44
Backbones + NAPs + ISPs
ISP
ISP
ISP
NAP
Backbones
Business
ISP
LAN
LAN
NAP
ISP
Consumer
ISP
LAN
Dial-up
45
Core Networks + Access Networks
DSL
Always on
Cable
Head Ends
@home
Covad
Cingular
Cell
Cell
Cell
LAN
NAP
Core
Networks
NAP
ISP
Satellite
Fixed Wireless
Sprint
LAN
AOL
LAN
Dial-up
46
Computers Inside the Core
DSL
Always on
Cable
Head Ends
@home
Covad
Cingular
Cell
Cell
Cell
LAN
NAP
NAP
ISP
Satellite
Fixed Wireless
Sprint
LAN
AOL
LAN
Dial-up
47
Interconnected World:
Agile or Fragile?
• Baltimore Tunnel Fire 18 July 2001
– “… The fire also damaged fiber optic cables, slowing Internet service
across the country, …”
– “… Keynote Systems … says the July 19 Internet slowdown was not
caused by the spreading of Code Red. Rather, a train wreck in a
Baltimore tunnel that knocked out a major UUNet cable caused it.”
– “PSINet, Verizon, WorldCom and AboveNet were some of the bigger
communications companies reporting service problems related to
‘peering,’ methods used by Internet service providers to hand traffic
off to others in the Web's infrastructure. Traffic slowdowns were
also seen in Seattle, Los Angeles and Atlanta, possibly resulting from
re-routing around the affected backbones.”
– “The fire severed two OC-192 links between Vienna, VA and New York,
NY as well as an OC-48 link from, D.C. to Chicago. … Metromedia
routed traffic around the fiber break, relying heavily on switching
centers in Chicago, Dallas, and D.C.”
48
Service-Level Peering
• Need common architecture for different
vendors to create components and work with
one another while still competing
• Some Observations
– IP originally designed for cooperative administrative
environments
– BGP “recently” retrofitted to architecture to manage
administrative relationships
– How to design in managed peering from first principles?
• Solution Based on Redirection Above IP
– Define the redirection architecture
– New client/infrastructure protocol & API (a la DNS)
– Do so in backward compatible way
49
Application-Specific Overlays
E.g., solve the multicast management and
peering problems by moving up the protocol
stack
Isolated
multicast
clouds
multicast
cloud
multicast
cloud
multicast
cloud
multicast
cloud
multicast
cloud
Traditional
unicast
peering
Steve McCanne
50
Application-Level
Servers/Routers
Solve the multicast management and peering
problems by moving up the protocol stack
Steve McCanne
51
Outline
•
•
•
•
•
•
What is this Course About?
Technology Trends
Evolution of the Internet
Business Trends
Implications and Issues
Summary and Conclusions
52
Layerized Internet Service
Business Model
Applications
(Portals, E-Commerce,
E-Tainment, Media)
Appl Infrastructure Services
(Distribution, Caching,
Searching, Hosting)
AIP
ISV
Application-specific Servers
(Streaming Media, Transformation)
ASP
Internet
Data Centers
ISP
CLEC
Application-specific
Overlay Networks
(Multicast Tunnels, Mgmt Svrcs)
Global Packet Network
Internetworking
(Connectivity)
53
A New Kind of Internet
Web Site Caching
Comparison Shopping
Interactive TV Guide
Local Ad Insertion
Streaming Media
Application
Services
Customer J
Terminal Equipment &
Access Network
Applications
Regional Communications
PC, Set-top Box.
Smart Phone, Game
Console, E-toys
Web, E-mail, Chat,
E-commerce,
E-tainment
ISP
Web Hosting
Server “Platform”
ISP Caching
Search Engine
Server Computing
Wide-Area Communications
Infrastructure
Services
High Performance
Backbone
54
Open vs. Closed Access to Services
Covad
DSL
Time/Warner
Roadrunner
AOL Dial-up
AT&T Cable
Access
Cable, DSL, MMDS,
LMDS, Satellite
CNCX
AOL
@Home
Local Network
Management
ISP
Williams
AOL
@Home
Routing &
Distribution
Backbone
Provider
Web
AOL/Netscape
Time/Warner
Excite
Content
Portal
Web Sites
• Closed end-to-end pipe: optimized performance
• But companies developing compelling infrastructure
technology that any content provider or ISP can adopt
• Closed system can’t benefit from these
55
From Network Management to
Service Management
Service Level Control
Server
Load Balancing
• Server
and site
availability
• Balanced
server and
site load
Advanced Traffic
Management
• Rapid change
• Network and
application
flexibility
• Scalability
• Complex site
administration
• Rapid problem
diagnosis/
isolation
• Service level
measurement
• Multi-tier
resource
monitoring
• Preferential
Services
• Resource
Provisioning
• Self-tuning
• Problem
prevention
Chris Morino, Resonate
56
Service Reliability is Critical
• ISP connection down
• LAN segment
overloaded
Network
Failure 18.2%
• CPU overloaded
• NIC failure
• Process hung
• Slowed database
performance
Source: IDC
Applications
Systems
Failure 28.5%
Server Failure 20%
OS Failure 24.6%
Administration 8.7%
Chris Morino, Resonate
57
Competition vs. Cooperation
• Internet Service Providers: Competition
– Peering for packet transport: BGP protocol
– Charging based on traffic volumes
ISP A
Peering
Point
Hot Potato
Routing
Peering
Point
ISP B
58
Mobile Internet Might Be
Different Than Wired Internet
• Wireless is a smarter pipe
– Location-awareness
– UI dictates need for personalization, mediation
• Clear billing authority: it’s the access provider
– People actually do pay for transport
– Reverse billing allows content provider to charge for service
• Peering as a necessity
–
–
–
–
Operators provide local service
Roaming agreements provide basis for service peering
Well understood arrangements for settlements
New economies driving towards shared network deployment
• Person-to-Person communications is a killer app
• Microsoft’s non-monopoly
59
Cooperation and Peering
• 3G Spectrum Auctions: 150 billion ECU;
Capital outlays may match spectrum expenses,
all before first revenue
• New business models in Mobile Networks
– Compelling services make the difference
– Collaborate on deployment of physical network
– Compete on provisioning of services
• Peering For More Than Connectivity
– Horizontal architecture of services on top of networks
– Virtual Home Environments
– Relationships between operators, billing agents,
service providers
60
Any Way to Build
a Network?
• Partitioning of frequencies independent of actual
subscriber density
– Successful operator oversubscribe resources, while less popular
providers retain excess capacity
– Different flavor of roaming: among collocated/competing service
providing
• Duplicate antenna sites
– Serious problem given community resistance
• Redundant backhaul networks
– Limited economies of scale
61
The Case for Horizontal
Architectures
“The new rules for success will be to provide one
part of the puzzle and to cooperate with
other suppliers to create the complete
solutions that customers require. ... [V]ertical
integration breaks down when innovation
speeds up. The big telecoms firms that will
win back investor confidence soonest will be
those with the courage to rip apart their
monolithic structure along functional layers,
to swap size for speed and to embrace rather
than fear disruptive technologies.”
The Economist Magazine, 16 December 2000
62
Feasible Alternative: Horizontal
Competition vs. Vertical Integration
• Service Operators “own” the customer, provide
“brand”, issue/collect the bills
• Independent Backhaul Operators
• Independent Antenna Site Operators
• Independent Owners of the Spectrum
• Microscale auctions/leases of network
resources
• Emerging concept of Virtual Operators
63
Business as Usual:
Vertical Integration
PBMS
Sprint
Access
Network
Backhaul
Network
Access
Network
Backhaul
Network
PSTN Network
(Multiservice Provider today)
Internet
(Multiservice Provider today)
• Each operator owns own frequencies, cell sites,
backhaul network
64
Business Unusual:
Horizontal Competition
Sprint “leases”
frequencies from
PBMS, on-demand,
based on the density
of its subscribers
“Mom&Pop”
Cell Site
Operators
Access Network
Access Network
Backhaul
Network
PSTN
Network
Backhaul
Network
Internet
65
Virtual
Operator
• MVNO: Virgin Mobile and One2One in UK
– Distinguish based on marketing and billing plan innovations
– VM competes for subscribers but uses One2One’s network
• “Operators without subscribers”: local premises
deploy own access infrastructure
– Better coverage/more rapid build out of network
– Deployments in airports, hotels, conference centers, office
buildings, campuses, …
• Overlay service provider (e.g., PBMS) vs.
organizational service provider (e.g., UCB IS&T)
– Single bill/settle with service participants
– Operator Wireless LAN
• Support for ensemble devices
– Cell Phone + Wall Camera & Display
66
Outline
•
•
•
•
•
•
What is this Course About?
Technology Trends
Evolution of the Internet
Business Trends
Implications and Issues
Summary and Conclusions
67
What will be the Next Generation
of Driving Applications?
• Location-aware/context-aware information delivery
and presentation
– Extends UniIn-Box: loc-based, exploits calendar info
– Mediation to translate formats
• IP Telephony, Packet VoD, Teleconferencing
– Streaming media, multicast-based
– Bandwidth, latency, jitter, lose rate constraints
– Clearinghouse provisioning
• Event Delivery for Distributed Applications
– Performance/reliability constrained messaging
– Management of Content Delivery Networks, Distributed Service
architecture?
• Interactive Games? Distributed Storage
(OceanStore)? Telemetry?
68
What Will Be the Next Generation
Operational Environment?
• Virtual Operators/Service Provider (VOSP)
– Provide service to end users with no server/network
infrastructure of own
– Independent “Path” providers (e.g., ISPs) and Server providers
(e.g., Internet Data Centers)
– Many-to-many relationship between VOSP and Path/Server
Providers
• Confederated Service Provider
– Service-level peering: sharing of paths and servers to deploy
end-to-end service with performance and reliability constraints
• Note: Akamai runs “the world’s largest service
network” without owning a network!
69
Alternative Operational
Environments
• Confederation Model
– Providers share (limited) information about topology, server
location, path performance
– Cooperatively collect internal information and share
• Overlay Model
– Reverse-engineer topology and intra-cloud performance
– Collection done by brokers outside of the cloud
• SLAs, Verification, Maintenance of Trust
Relationships different in the two models
• Is there an operational/performance advantage to
the Confederation Model?
70
Open Issues/Questions
• Traditional Overlay Networks
– Server (“Application Level Router”) Placement
» For scaling, reliability, load balancing, latency
» Where? Network topology discovery: WAN Core,
Metro/Regional, Access Networks
– Choice of Inter-Server “Paths”
» For server-to-server latency/bandwidth/loss rate
» Predictable/verifiable network performance (intra-ISP
SLA)
– Redirection Mechanisms
» Random, round-robin, load-informed redirection
» Net vs. server as bottleneck
71
Open Issues/Questions
• Performance-constrained Service Placement
– Separation of Service, Server, Service Path
» Assume “Server Centers” known, can be “discovered” (how
does OceanStore deal with this?), or register with a Service
Placement Service (SPS)
» How is Service named, described, performance constraints
expressed, and registered?
» How is app/service-specific performance measured and made
known to Service Placement Service?
– Brokering between Server Centers and Service Creator, Path
Provider and Service Creator
• If core network bandwidth becomes infinite and
“free”, does it matter where services are placed?
– Latency reduction vs. economies of centralized management
72
Emerging Reference Architecture
Constraint
Specification
Distributed Application
Marshal Resources
Based on Economic Constraints
Service Registration
Service Placement Service
Path Broker
Server Broker
Perf Measurement Service
Verify
SLAs
Path
Provider
(ISP
Cloud)
Path
Provider
(ISP
Path Provider (ISPCloud)
Cloud)
Adapt
Service
Redirection
Pricing
Service
Server Registration
Advertisement
Registration
Path
Provider
(ISP
Cloud)
Path
Provider
(ISP
Cloud)
Server Center Provider
73
Methodological Framework
• Problem: performing scaled, wide-area networking
studies in the current Internet environment
• Possible Solution: Wide-area Network Emulation
– Virtual WAN (VWAN) on Large-scale Multicomputer Testbeds
– Build operational model on top of VWAN
» Traffic generation and measurement infrastructure
» Build Confederation and Overlay operational models
» What part of mechanisms for measurement, negotiation,
registration, redirection, etc. the same and which are
different?
74
Outline
•
•
•
•
•
•
What is this Course About?
Technology Trends
Evolution of the Internet
Business Trends
Implications and Issues
Summary and Conclusions
75
Cable
Modem
Connectivity and Processing
Premisesbased
Access
Networks
Core Networks
WLAN
Transit Net
WLAN
Operatorbased
Cell
Cell
Cell
Regional
LAN
Transit Net
Premisesbased
WLAN
LAN
Internet
Datacenter
NAP
Public
Peering
Data
Voice
Analog
Transit Net
H.323
RAS
H.323
PSTN
LAN
Private
Peering
DSLAM
Data
Voice
Wireline
Regional
76
Challenges for Converged
Networks
• Services spanning access networks, to achieve high
performance and manage diversity of end devices
• Not about specific Information Appliances
• Builds on the New Internet: multiple applicationspecific “overlay” networks, with new kinds of
service-level peering
• Pervasive support for services within “intelligent”
networks
–
–
–
–
Automatic replication
Document routing to caches
Compression & mirroring
Data transformation
77
Managing Edge Versus
Core Services
• Wide-area bandwidth efficiency
• Increasing b/w over access networks, but impedance
mismatch between core and access nets
• Fast response time (and more predictable)
• Opportunity to untegrate localized content
• Associated with client (actually ISP), not server
• Examples:
–
–
–
–
–
–
Caching: exploits response time, b/w efficiency, high local b/w
Filtering: form of local content transformation
Internet TV: b/w efficiency, high local b/w, predictable response
Transformation: adapt content for end user/diverse access devices
Software Rental: sxploits high local b/w
Games, chat rooms, ….
78
Yielding a New
Research Agenda
• New Definition of “Quality of Service”
– Perceived quality depends on services in the network
– Manage caches, redirection, NOT bandwidth
– Enable incorporation of localized content
• Bandwidth Issues
–
–
–
–
Tier 1 ISP backbones rapidly moving towards OC 192 (9.6 gbs!)
Better interconnection: hops across ASs decreasing over time
Emerging broadband access networks: cable, DSL, ...
End-to-end latency/server load dominate performance
–
–
–
–
IP Multicast, DNS, …
Rethinking the End-to-End Principle
Service/content-level peering, just like routing-level peering
Secure end-to-end connection compatible with service model?
• Supporting Old Services in the New Internet
79