apnoms 2006 1 prosperous de facto communication technologies outside academic societies sep. 27,...
TRANSCRIPT
APNOMS 2006 1
Prosperous De Facto Communication Technologies outside Academic Societies
Sep. 27, 2006
Tohru AsamiGraduate School of Information Science and Technology
The University of Tokyo
APNOMS 2006 2
Contents
1. Operational Department-initiative Developments
2. A case study in Japan
(1) Ether over Ether
(2) 3GPP2 Mobile IP
3. Roles of Academic Societies for Commercialization
Contents
1. Operational Department-initiative Developments
2. A case study in Japan
(1) Ether over Ether
(2) 3GPP2 Mobile IP
3. Roles of Academic Societies for Commercialization
APNOMS 2006 3
1. Operational Department-initiative Developments
Competitive services Shorter Development Cycle
1. Operational Department-initiative Developments
Competitive services Shorter Development Cycle
APNOMS 2006 4
Telecom Market after the Collapse of the IT Bubble
Economy
US Telcos: Return to monopolization Development power of new services weakened No leadership for new L1&L2 services since 2000
Telcos in Japan: US Teleco model in ’80s Competitive service developments among service
providers New services deployed by Japan & Korea especially
in 3G Markets in 21 Century So in fixed communication markets Japan&Korea=Test bed for new services &
technologies in the world
APNOMS 2006 5
In Reality, Larger Gap between R&D and Operational Department
Looks like a boon to R&D Labs. as well as academic societies Business divisions cannot win the market just
by following given services by others Better environments: Business as well as
infrastructures(3G, FTTH,etc.)
However, business divisions developed such services by themselves Less contributions to academic societies Larger risks of developments
APNOMS 2006 6
2. A case study in Japan (1) Ether over Ether
(2) 3GPP2 Mobile IP
2. A case study in Japan (1) Ether over Ether
(2) 3GPP2 Mobile IP
APNOMS 2006 7
Business Needs vs. Researches
Banters on Mobile IP and Multicast Technologies Everybody knows it but nobody uses. Research for research. Failures in the Internet Technologies What is the killer application? When is it put into real markets?
But they are important in NGN as well as FMC Are they such dumb technologies?
APNOMS 2006 8
2. A Case Study in Japan
Example 1: Ether over EtherA Nation-wide Ethernet Service
First Killer Application for Multicasting
2. A Case Study in Japan
Example 1: Ether over EtherA Nation-wide Ethernet Service
First Killer Application for Multicasting
APNOMS 2006 9
Wide-area Ethernet Service= Cash Cow for Solution Business
Ref: Fuji Chimera Research Institute Inc., “Market Survey on Broadband Business in 2005”
Earnings in 100 M Yens
Earnings
# of Circuits
Circuits
Year
APNOMS 2006 10
The Dawn of Communication-and-broadcasting Convergence in 2000
Oct. 1999Wide-area LAN service started by CWC
Dec. 2000Q-in-Q Service introduced by TTNet (Poweredcom)
Dec. 2002Ether over Ether(EoE ) RFP was issued
Mar. 2004EoE put into operation
Nation-wide Communication Service based on Broadcast-based Routing Protocol!
APNOMS 2006 11
Principles of Wide-area Ethernet Service
Applying Ethernet, widely used in LAN and with inexpensive switches, to WAN, provide non-conventional services such as
i. Access network service having high affinity with Ethernet-based LANsii. Inexpensive communication service
Enhanced Ethernet technology in the following points Long distance transmission High reliability with redundancy VPN
From the service point view Protocol-transparent service above Multicast is inevitable for VLAN (The first killer application for
Multicast)
APNOMS 2006 12
Nation-wide Ethernet Service
Wide-area Ethernet Service Networks
User LAN CUser LAN B
User LAN A
C-SA
C-DA
B-SA
B-DA
Terminal a
Terminal b
Switch b
Switch a
Core Switch X Switch c
Tokyo
Osaka
Nagoya
APNOMS 2006 13
IEEE 802.1ad
Len/TypeC-DA C-SA S-TAG
C-TAGoptional
FCS
Len/TypeC-DA C-SA C-TAG
optionalFCS
6 6 4 2 46 - 1500 octets 6
4
(1) User Frame
(2) IEEE 802.1ad
Client Data
TPID
TCI
2 octets
PCP DE VID
3 1 12bits
TPID: Tag Protocol IDPCP: Priority Code PointDE: Drop EligibleVID: VLAN ID
C-DA: Customer Destination AddressC-SA: Customer Source AddressC-TAG:Customer VLAN TagS-TAG: Service VLAN TagLen/Type: Length or Type
Client Data
4
APNOMS 2006 14
IEEE 802.1ad vs. Ether over Ether
Base technologies of IEEE802.1ad s.t. Q-in-Q did not fit into the large scale deployments in Japan Not enough number of VLANs A number of troubles
FDB overflows in core switches Frequent loop conditions …..
Redesign of switches was urgent after the service started Ether over Ether = Development of new
switches based on a Japanese vendor
APNOMS 2006 15
Ether over Ether vs. IEEE 802.1ah
B-DA B-SA S-TAGEoE
TPIDFCS
(3) Ether over Ether
TTL
EID
2 1 1
I-TagB-DA B-SA B-TAG FCS
(4) IEEE 802.1ah6 octets
I-SIDI-TagTPID
PCP
RSV
2 octetsDEI
24bits3 1 4B-TAG: Backbone TAGI-TAG: Extended Service TagPCP: Priority Code PointDEI: Drop Eligible Indicator RSV:Reserved (0)I-SID:Instance-Service ID
Len/TypeC-DA C-SA
C-TAGoptional
Client Data
Len/TypeC-DA C-SA
C-TAGoptional
Client Data
B-DA: Backbone Destination Address B-SA: Backbone Source Address TTL: Time to Live EID: Extension tag ID
TPID
TCI
2 octets
PCP DE VID
3 1 12bits
APNOMS 2006 16
Ether over Ether vs. IEEE 802.1ah
IEEE 802.1ah Ether over Ether
Category Mac-in-Mac Mac-in-Mac
Loop Protection
None(Detection of Loop) TTL
Max # of VLANs
I-SID(24bits) =16,777,215 S-VID(12bits) + EID(8bits) =1,048,576
Max # of Broadcast Domains
B-VID(12bits)=4096 1,048,576
OAM Use IEEE 802.1ag. No OAM for MAC Flash of Edges
Ether over Ether Control Protocol.OAM for MAC Flash of Edges
Reactive Proactive
APNOMS 2006 17
Deployments of Ether over EtherJapan: Cash cow in enterprise market Service: NTT Communications, KDDI, JT, Other major
Communication Carriers Core switch products: Fujitsu, Hitachi Cable Used in part: ALAXALA Networks
Other countries: Nation-wide Ethernet is not popular (for metro-ethernet at best) Ether over Ether has not been used other than in Japan. The world big 3 switch vendors (Cisco Systems, Extreme
Networks, Foundry Networks) do not support EoE , but IEEE802.1ad.
No standardization efforts
APNOMS 2006 18
2. A Case Study in Japan
Example 2: 3GPP2 Mobile IP
First Carrier-Grade Specification
for Mobile IP
2. A Case Study in Japan
Example 2: 3GPP2 Mobile IP
First Carrier-Grade Specification
for Mobile IP
APNOMS 2006 19
3GPP2 Mobile IP
HLR
PDSNa
BSCa PCFa
BS
BSC
BS
PCF
RADIUSa
BSC PCFBS
BSC
BS
PCF
PDSNb
RADIUSc
HAc
CNMS
Access ProviderNetwork
Home Access Provider Network
Home IP Network (Private)
IPNetwork
HAb
RADIUSb
MS: Mobile StationRAN: Radio Access NetworkBS: Base StationBTS: Base Transceiver StationBSC: Base Station ControllerPCF: Packet Control FunctionPDSN: Packet Data Serving NodeRADIUS: Remote Authentication Dial-In User ServiceFA: Foreign AgentHA: Home Agent
MSC:Mobile Switching CenterHLR:Home Location Register
cdma2000 RAN
BTS MSCA1
SS7 Network
Modified from 3GPP2, “cdma2000 Wireless IP Network Standard: Introduction,” X.S0011-001-D, http://www.3gpp2. org/Public_html/specs/X.S0011-001-D_v1.0_060301.pdf, Feb.2006.
APNOMS 2006 20
(a) Protocol Reference Model for MIP4 Control and IKE
cdma2000Air Interface
PPP
IP
cdma2000Air Interface
PL
A-8
PL
A-8
PL
A-10
MS BTS,BSC PCF
PL
A-10
PL
L2
PDSN/FA
PPP
IP/IPsec
UDP
MIP4
UDP
MIP4 IKE
PL
L2
HA
IP/IPsec
UDP
MIP4IKE
Modified from 3GPP2, “cdma2000 Wireless IP Network Standard: Introduction,” X.S0011-001-D, http://www.3gpp2. org/Public_html/specs/X.S0011-001-D_v1.0_060301.pdf, Feb.2006.
APNOMS 2006 21
(b) Protocol Reference Model for MIP4 User Data
PDSN/FA HA
PL
A-10
PPP
PL
L2
MS BTS,BSC PCF
cdma2000Air Interface
PPP
cdma2000Air Interface
PL
A-8
PL
A-8
PL
A-10
IP IP/IPsec
IP
PL
L2
IP IP/IPsec
CN
PL
L2
IP
PL
L2
Modified from 3GPP2, “cdma2000 Wireless IP Network Standard: Introduction,” X.S0011-001-D, http://www.3gpp2. org/Public_html/specs/X.S0011-001-D_v1.0_060301.pdf, Feb.2006.
APNOMS 2006 22
A-8/A-9 and A-10/A-11
(c) BSC-PCF Interface (d) PCF-PDSN Interface
PL
L2
IP
GRE
PPP
PL
L2
IP
GRE
PPP
PL
L2
IP
UDP
Signalings
PL
L2
IP
UDP
Signalings
(c1) A-8 (c2) A-9 (d1) A-10 (d2) A-11
APNOMS 2006 23
What is Communication Business?
Privacy
Authentication
Authorization
Accounting
AAA
APNOMS 2006 24
Advantages of Fixed IP Address for Service Providers
Popular Better Security
Nonsense unless it is bound to some other tamperproof ID Customized services for each user
QoS Seamless hand-over between fixed and mobile networks
Terminals can be called in case that they are always-on
Not So Popular Scalable and inexpensive load-balancing of servers = Assign each
server to a set of addresses Design and control the congestion of server Scalable accounting system Protocol-transparent above L3
Load balancers independent on new service introductions Less operational cost Avoid payload analysis in load balancing, such as Cookie , SSL
Session ID, etc., from the point of secrecy of communication
APNOMS 2006 25
Advantages and Disadvantages Talked on MIP
Advantages: Mobility management by MIP = Session management by SIP Easy monitoring of mobile terminals = traffic controls Advantages of fixed IP address of Terminal More advantages in private address networks
Use terminals in visiting networks without configuration changes Duplicated IP addresses = enlarges the IP address space
Disadvantages: Difficult to optimize routes <= In fact it’s not a flaw. Mobility control is not sufficient to keep QoS <= Other IP
technologies also cannot No large scale deployments other than 3GPP2 No public information on usages of MIP by 3GPP2 providers
APNOMS 2006 26
Trend of Internet-accessible Cell Phones
I-mode: 4,428 MillionEzweb: 1,847 MillionVf-live: 1,282 Million
At the end of April, 2005Total: 7,559 Million(Total cell phones: 8,743 Million)
Ref.: Telecommunications Carriers Association
2002200120001999 2003
Ezweb, I-mode, Vodafone live(Million)
2004
Total subscription phones
2005
01020304050
60708090
100
6 9 12 3 6 9 12 3 6 9 12 3 6 9 12 3 6 9 12 3 6 9 12 3
i-mode
EZweb
Vf‐live
Class A Network Address 10.0.0.0/8 : 224=16,777,216
APNOMS 2006 27
Management of MIP addresses in 3GPP2
Principles: Reverse tunneling + PDSN(for PPP & FA) Assign I+j+k+l IP addresses (Simple IPv4(i), Simple IPv6( j),
MIP4(k), MIP6(l)) to each terminal(PPP) Support i=j=0 for FA-mode of MIP Possible to assign NAI or fixed IP address to terminal (MS)
based on hardware-defined IMSI (International Mobile Subscriber Identity)
=>=>=>Keep location privacy of terminalTerminals with the same private IP address belonging to different private networks co-exist under the same PDSN
APNOMS 2006 28
Possible Business Advantages of 3GP2 MIP- All Comes from PPP&FA with Reverse Tunneling -
As solution service Mobile terminals belonging to a enterprise network can be used in the
same conditions as they are the enterprise network No exits for the Internet in case of terminals given only MIP address
As cell phone service Extend the private network address to almost unlimited Address space = N x 10.0.0.0/8
Inter-provider service Easy roaming between providers without taking care of IP addresses
used Application can be served even to terminals in visiting networks
Easy realization of MVNO ( Mobile Virtual Network Operator) without taking care of IP addresses used
APNOMS 2006 29
Summary of 3GPP2 MIPFA mode vs. Co-located Care-of Address Mode
Faults in MIP standardization process FA mode has clear business needs No business needs for co-located care-of address mode No business needs for route optimization Reverse tunneling and PPP in FA mode was a key for business
requirements
Wasteful efforts for many technological issues which do not have clear business requirements?
Could we input proper business needs into the MIP standardization process?
APNOMS 2006 30
3. Roles of Academic Societies for Commercialization
3. Roles of Academic Societies for Commercialization
APNOMS 2006 31
Academic Publications
Ether over Ether 3GPP2 Mobile IP
Journal Papers 0 0
Technical Groups 0 0
Technical Commentaries 0.1(just as a technical term)
0
Technical Books 1(IEICE) 0
Academic societies are no more needed for technology development affecting the fate of the company?
APNOMS 2006 32
Lessons1. High possibility of service creations originated by
Japanese providers under competitive environmentsi. Operational divisions can develop services by themselvesii. Less influence by R&D division
a. No output for R&D
iii. Possibility of de facto service/technology development by providers other than NTT
iv. Increase of risks at development
2. Such de facto can appear earlier than corresponding international standards
3. No top vendors support Japan-local de facto standard① Standardizations are necessary but No incentive for
operational divisions to standardize② Who make efforts for standardizations?
APNOMS 2006 33
Towards Reconstruction of Proper R&D Model-- Avoid Research without needs & Development without seeds --
Plan
Do
Check
Action
Plan
Do
Check
Action
NeedsSeeds
OperationalDivision
R&DDivisions
Plan
Do
Check
Action
AcademicSocieties
Needs
Seeds
Role of R&D = Generalize problems in operational divisions, and publicize them to academic societies
APNOMS 2006 34
Thank You!R&D as well as operational divisions are
at the crossroads in communication industry.
Some relations to the stagnation of academic societies?
How to improve?
Thank You!R&D as well as operational divisions are
at the crossroads in communication industry.
Some relations to the stagnation of academic societies?
How to improve?