cellular’networks’and’mobile’ compung’ …coms6998-8/lectures/lec9-cell...previousstudies...
TRANSCRIPT
![Page 1: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/1.jpg)
Cellular Networks and Mobile Compu5ng COMS 6998-‐8, Spring 2012
Instructor: Li Erran Li ([email protected])
hKp://www.cs.columbia.edu/~coms6998-‐8/ 3/26/2012: Cellular Network and Traffic
Characteriza5on
![Page 2: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/2.jpg)
An Untold Story of Middleboxes in Cellular Networks
Zhaoguang Wang1
Zhiyun Qian1, Qiang Xu1, Z. Morley Mao1, Ming Zhang2 1University of Michigan 2MicrosoZ Research
![Page 3: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/3.jpg)
Background on cellular network
3
Internet Cellular Core Network
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 4: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/4.jpg)
Why carriers deploy middleboxes?
4
Internet Cellular Core Network
Private IP Public IP
IP address
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 5: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/5.jpg)
Problems with middleboxes
5
Internet Cellular Core Network
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
Policies ? Applica=on
performance ?
P2P ?
Smartphone energy cost
?
Courtesy: Z. Wang et al.
![Page 6: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/6.jpg)
Challenges and solu5ons
• Policies can be complex and proprietary √ Design a suite of end-‐to-‐end probes
• Cellular carriers are diverse √ Publicly available client Android app
• Implica5ons of policies are not obvious √ Conduct controlled experiments
6 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 7: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/7.jpg)
Related work
• Internet middleboxes study – [Allman, IMC 03], [Medina, IMC 04]
• NAT characteriza5on and traversal – STUN[MacDonald et al.], [Guha and Francis, IMC 05]
• Cellular network security – [Serror et al., WiSe 06], [Traynor et al., Usenix Security 07]
• Cellular data network measurement – WindRider, [Huang et al., MobiSys 10]
7 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 8: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/8.jpg)
Goals
• Develop a tool that accurately infers the NAT and firewall policies in cellular networks
• Understand the impact and implica5ons – Applica5on performance – Energy consump5on – Network security
8 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 9: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/9.jpg)
The NetPiculet measurement system
9
Internet Cellular Core Network
NetPiculet Server
NetPiculet Client
NetPiculet Client
NetPiculet Client
NetPiculet Client
Policies…
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 10: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/10.jpg)
Target policies in NetPiculet
10
Firewall IP spoofing TCP connec5on 5meout Out-‐of-‐order packet buffering
NAT
NAT mapping type Endpoint filtering TCP state tracking Filtering response Packet mangling
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 11: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/11.jpg)
Target policies in NetPiculet
11
Firewall IP spoofing TCP connec5on 5meout Out-‐of-‐order packet buffering
NAT
NAT mapping type Endpoint filtering TCP state tracking Filtering response Packet mangling
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 12: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/12.jpg)
Key findings
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 12
Firewall
Some carriers allow IP spoofing Create network vulnerability
Some carriers 5me out idle connec5ons aggressively Drain baFeries of smartphones
Some firewalls buffer out-‐of-‐order packet Degrade TCP performance
NAT One NAT mapping linearly increases port # with 5me Classified as random in previous work
Courtesy: Z. Wang et al.
![Page 13: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/13.jpg)
Diverse carriers studied
• NetPiculet released in Jan. 2011 – 393 users from 107 cellular carriers in two weeks
13 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
91%
9%
UMTS
EVDO
43%
24%
19%
10%
2% 2%
Europe
Asia
North America
South America
Australia
Africa
Technology Con=nent
Courtesy: Z. Wang et al.
![Page 14: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/14.jpg)
Outline
14
1 • IP spoofing
2 • TCP connec5on 5meout
3 • TCP out-‐of-‐order buffering
4 • NAT mapping
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 15: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/15.jpg)
Outline
15
1 • IP spoofing
2 • TCP connec5on 5meout
3 • TCP out-‐of-‐order buffering
4 • NAT mapping
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 16: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/16.jpg)
Why allowing IP spoofing is bad?
16
Internet Cellular Core Network
10.9.9.101
10.9.9.202
SRC_IP = 10.9.9.101 …
DST_IP = 10.9.9.101 …
DST_IP = 10.9.9.101 …
DST_IP = 10.9.9.101 …
DST_IP = 10.9.9.101 …
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 17: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/17.jpg)
Test whether IP spoofing is allowed
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 17
Internet Cellular Core Network
NetPiculet Server
NetPiculet Client
Allow IP spoofing!
10.9.9.101
SRC_IP = 10.9.9.202 PAYLOAD = 10.9.9.101
Courtesy: Z. Wang et al.
![Page 18: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/18.jpg)
4 out of 60 carriers allow IP spoofing
18 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
7%
93%
Allow Disallow
IP spoofing should be disabled
Courtesy: Z. Wang et al.
![Page 19: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/19.jpg)
Outline
19
1 • IP spoofing
2 • TCP connec5on 5meout
3 • TCP out-‐of-‐order buffering
4 • NAT mapping
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 20: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/20.jpg)
Why short TCP 5meout 5mers are bad?
20
Internet Cellular Core Network
KEEP-‐ALIVE KEEP-‐ALIVE KEEP-‐ALIVE Terminate Idle TCP Connec5on
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 21: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/21.jpg)
5min < Timer
Measure the TCP 5meout 5mer
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 21
Internet Cellular Core Network
NetPiculet Server
NetPiculet Client
5min < Timer < 10min
Time = 0 Time = 5 min Time = 10 min
Is alive?
Yes!
Is alive?
Courtesy: Z. Wang et al.
![Page 22: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/22.jpg)
Short 5mers iden5fied in a few carriers
< 5 min 5%
5 -‐ 10 min 10%
10 -‐20 min 8%
20 -‐ 30 min 11%
> 30 min 66%
22
4 carriers set 5mers less than 5 minutes
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 23: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/23.jpg)
Short 5mers drain your baKeries • Assume a long-‐lived TCP connec5on, a baKery of 1350mAh • How much baKery on keep-‐alive messages in one day?
23
20%
5 min
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 24: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/24.jpg)
Outline
24
1 • IP spoofing
2 • TCP connec5on 5meout
3 • TCP out-‐of-‐order buffering
4 • NAT mapping
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 25: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/25.jpg)
TCP out-‐of-‐order packet buffering
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 25
Internet Cellular Core Network
NetPiculet Server
NetPiculet Client
Buffering out-‐of-‐order packets
Packet 1 Packet 2 Packet 3 Packet 4 Packet 5 Packet 6
Courtesy: Z. Wang et al.
![Page 26: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/26.jpg)
Fast Retransmit cannot be triggered
26
1 2
Degrade TCP performance!
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
RTO
Courtesy: Z. Wang et al.
![Page 27: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/27.jpg)
TCP performance degrada5on
• Evalua5on methodology – Emulate 3G environment using WiFi – 400 ms RTT, loss rate 1%
27
+44%
Longer downloading
5me
More energy consump5on
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 28: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/28.jpg)
Outline
28
1 • IP spoofing
2 • TCP connec5on 5meout
3 • TCP out-‐of-‐order buffering
4 • NAT mapping
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 29: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/29.jpg)
NAT mapping is cri5cal for NAT traversal
29
A B NAT 1 NAT 2
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
Use NAT mapping type for port predic5on P2P
Courtesy: Z. Wang et al.
![Page 30: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/30.jpg)
What is NAT mapping type?
• NAT mapping type defines how the NAT assign external port to each connec5on
30 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
NAT
12 TCP connec5ons
…
Courtesy: Z. Wang et al.
![Page 31: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/31.jpg)
Behavior of a new NAT mapping type
31 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
• Creates TCP connec5ons to the server with random intervals
• Record the observed source port on server
Treated as random by exis5ng traversal techniques Thus impossible to predict port
NOT random! Port predic5on is feasible
Courtesy: Z. Wang et al.
![Page 32: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/32.jpg)
Lessons learned
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 32
Firewall
IP spoofing creates security vulnerability IP spoofing should be disabled
Small TCP 5meout 5mers waste user device energy Timer should be longer than 30 minutes
Out-‐of-‐order packet buffering hurts TCP performance Consider interac=on with applica=on carefully
NAT One NAT mapping linearly increases port # with 5me Port predic=on is feasible
Courtesy: Z. Wang et al.
![Page 33: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/33.jpg)
Conclusion
• NetPiculet is a tool that can accurately infer NAT and firewall policies in the cellular networks
• NetPiculet has been wildly deployed in hundreds of carriers around the world
• The paper demonstrated the nega5ve impact of the network policies and make improvement sugges5ons
33 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Z. Wang et al.
![Page 34: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/34.jpg)
Cellular Data Network Infrastructure Characteriza=on &
Implica=on on Mobile Content Placement
Qiang Xu*, Junxian Huang*, Zhaoguang Wang* Feng Qian*, Alexandre Gerber++, Z. Morley Mao*
*University of Michigan at Ann Arbor ++AT&T Labs Research
![Page 35: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/35.jpg)
Applica=ons Depending on IP Address • IP-‐based iden5fica5on is
popular – Server selec5on – Content customiza5on – Fraud detec5on
• Why? -‐-‐ IP address has strong correla5on with individual user behavior
35 Courtesy: Q. Xu et al. Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
![Page 36: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/36.jpg)
Cellular IP Address is Dynamic • Cellular devices are hard to geo-‐locate based on IP addresses – One Michigan’s cellular device’s IP is located to places far away
• /24 cellular IP addresses are shared across disjoint regions
36 Courtesy: Q. Xu et al. Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
![Page 37: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/37.jpg)
Problem Statement • Discover the cellular infrastructure to explain the diverse
geographic distribu5on of cellular IP addresses and inves5gate the implica5ons accordingly
– The number of GGSN data centers – The placement of GGSN data centers – The prefixes of individual GGSN data centers
37 Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) Courtesy: Q. Xu et al.
![Page 38: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/38.jpg)
Challenges • Cellular networks have limited visibility – The first IP hop (i.e., GGSN) is far away -‐-‐ lower aggrega5on levels of base sta5on/RNC/SGSN are transparent in TRACEROUT
– Outbound TRACEROUTE -‐-‐ private IPs, no DNS informa5on
– Inbound TRACEROUTE -‐-‐ silent to ICMP probing
• Cellular IP addresses are more dynamic [BALAKRISHNAN et al., IMC 2009] – One cellular IP address can appear at distant loca5ons – Cellular devices change IP address rapidly
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
38 Courtesy: Q. Xu et al.
![Page 39: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/39.jpg)
Solu=ons
• Collect data in a new way to get geographic coverage of cellular IP prefixes – Build Long-‐term and na5on-‐wide data set to cover major carriers and the majority of cellular prefixes
– Combine the data from both client side and server side
• Analyze geographic coverage of cellular IP addresses to infer the placement of GGSN data centers – Discover the similarity across prefixes in geographic coverage – Cluster prefixes according to their geographic coverage
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 39 Courtesy: Q. Xu et al.
![Page 40: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/40.jpg)
Previous Studies
• Cellular IP dynamics – Measured cellular IP dynamics at two loca5ons [Balakrishnan et al., IMC 2009]
• Network infrastructure – Measured ISP topologies using ac5ve probing via TRACEROUTE [Spring et al., SIGCOMM 2002]
• Infrastructure’s impact on applica5ons – Es5mated geo-‐loca5on of Internet hosts using network latency [Padmanabhan et al., SIGMETRICS 2002]
– On the Effec5veness of DNS-‐based Server Selec5on [Shaikh et al., INFOCOM 2001]
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 40 Courtesy: Q. Xu et al.
![Page 41: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/41.jpg)
Outline
• Mo5va5on • Problem statement • Previous Studies • Data Sets • Clustering Prefixes • Valida5ng the Clustering Results • Implica5on on mobile content placement
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 41 Courtesy: Q. Xu et al.
![Page 42: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/42.jpg)
... timestamp lat. long. address 1251781217 36.75 -119.75 166.205.130.244 1251782220 33.68 -117.17 208.54.4.78 ...
Data Sets } DataSource1 (server logs): a loca5on search server
} millions of records } IP address, GPS, and 5mestamp
} DataSource2 (mobile app logs): an applica5on deployed on iPhone OS, Android OS, and Windows Mobile OS } 140k records } IP address and carrier
} RouteViews: BGP update announcements } BGP prefixes and AS number
device: <ID:C7F6D4E78020B14FE46897E9908F83B> <Carrier: AT&T> address: <GlobalIP: 166.205.130.51> ...
...|95.140.80.254|31500|166.205.128.0/17|31500 3267 3356 7018 20057|...
...|95.140.80.254|31500|208.54.4.0/24|31500 3267 3356 21928|...
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 42 Courtesy: Q. Xu et al.
![Page 43: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/43.jpg)
Map Prefixes to Carriers & Geographic Coverage
• Correlate these data sets to resolve each one's limita5ons to get more visibility
address lat. long. 166.205.130.244 36.75 -119.75 208.54.4.11 33.68 -117.17
prefix 166.205.128.0/17 208.54.4.0/24
address carrier 166.205.130.51 AT&T 208.54.4.11 T-Mobile
prefix lat. long. 166.205.128.0/17 36.75 -119.75 208.54.4.0/24 33.68 -117.17
prefix carrier 166.205.128.0/17 AT&T 208.54.4.0/24 T-Mobile
prefix carrier lat. long. 166.205.128.0/17 AT&T 36.75 -119.75 208.54.4.0/24 T-Mobile 33.68 -117.17
DataSource1 RouteViews DataSource2
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 43 Courtesy: Q. Xu et al.
![Page 44: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/44.jpg)
Outline
• Mo5va5on • Problem statement • Previous Studies • Data Sets • Clustering Prefixes • Valida5ng the Clustering Results • Implica5on on mobile content placement
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 44 Courtesy: Q. Xu et al.
![Page 45: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/45.jpg)
Mo=va=on for Clustering -‐-‐ Limited Types of Geographic
Coverage PaFerns
• Prefixes with the same geographic coverage should have the same alloca5on policy (under the same GGSN)
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 45 Courtesy: Q. Xu et al.
![Page 46: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/46.jpg)
Cluster Cellular Prefixes • 1. Pre-‐filter out those prefixes with very few records (todo) • 2. Split the U.S. into N square grids (todo) • 3. Assign a feature vector for each prefix to keep # records in
each grid • 4. Use bisect k-‐means to cluster prefixes by their feature
vectors (todo)
} How to avoid aggressive filtering? } keep at least 99% records
} How to choose N? } # clusters is not affected by N while N >
15 && N < 150 } The geographic coverage of each cluster is
coarse-‐grained
} How to control the maximum tolerable SSE?
46 Courtesy: Q. Xu et al. Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
![Page 47: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/47.jpg)
Clusters of the Major Carriers
All 4 carriers cover the U.S. with only a handful clusters (4-‐8) • All clusters have a large geographic coverage • Clusters have overlap areas
– Users commute across the boundary of adjacent clusters – Load balancing
47 Courtesy: Q. Xu et al. Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
![Page 48: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/48.jpg)
Outline
• Mo5va5on • Problem statement • Previous Studies • Data Sets • Clustering Prefixes • Valida=ng the Clustering Results • Implica5on on mobile content placement
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 48 Courtesy: Q. Xu et al.
![Page 49: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/49.jpg)
Validate via local DNS Resolver (DataSource2)
• Iden5fy the local DNS resolvers – Server side: log the incoming DNS requests on the authorita5ve DNS resolver of eecs.umich.edu and record (id_5mestamp, local DNS resolver)
• Profile the geographic coverage of local DNS resolvers – Device side: request id_=mestamp.eecs.umich.edu and record the (id_5mestamp, GPS)
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 49 Courtesy: Q. Xu et al.
![Page 50: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/50.jpg)
Validate via Cellular DNS Resolver (Cont.)
• Clusters of Carrier A’s local DNS resolvers
• Clusters of Carrier A’s prefixes
50 Courtesy: Q. Xu et al. Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
![Page 51: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/51.jpg)
Clustering Results • Goal -‐-‐ “…discover the cellular infrastructure to explain
the diverse geographic distribu5on of cellular IP addresses…” – All 4 major carriers have only a handful (4-‐8) GGSN data centers
– Individual GGSN data centers all have very large geographic coverage
• Goal -‐-‐ “…inves5gate the Implica5ons accordingly…” – Latency sensi5ve applica5ons may be affected
• CDN servers may not be able close enough to end users • Applica5ons based on local DNS may not achieve higher resolu5on than GGSN data centers
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 51 Courtesy: Q. Xu et al.
![Page 52: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/52.jpg)
Outline
• Mo5va5on • Problem statement • Previous Studies • Data Sets • Clustering Prefixes • Valida5ng the Clustering Results • Implica=on on mobile content placement
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 52 Courtesy: Q. Xu et al.
![Page 53: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/53.jpg)
Rou=ng Restric=on: How to Adapt Exis=ng CDN service to Cellular?
• Where to place content? – Along the wireless hops: require infrastructure support
– Inside the cellular backhaul: require support from cellular providers
– On the Internet: limited benefit, but how much is the benefit?
• Which content server to select? – Based on geo-‐loca5on: finer-‐grained loca5on may not available
– Based on GGSN: loca5on of GGSN
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 53 Courtesy: Q. Xu et al.
![Page 54: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/54.jpg)
Server Selec=on (DataSource2) • Approximately locate the server with the shortest latency – Based on IP address – Based on applica5on level informa5on, e.g., GPS, ZIP code, etc.
• Compare the latency to the Landmark server (1) closest to device with the latency to the Landmark server (2) closest to the GGSN – Es5mate the loca5on of GGSN based on TRACEROUT
} Select the content server based on GGSN!
54 Courtesy: Q. Xu et al. Cellular Networks and Mobile Compu5ng (COMS 6998-‐8)
![Page 55: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/55.jpg)
Contribu=ons • Methodology
– Combine rou5ng, client-‐side, server-‐side data to improve cellular geo-‐loca5on inference
– Infer the placement of GGSN by clustering prefixes with similar geographic coverage
– Validate the results via TRACEROUTE and cellular DNS server.
• Observa5on – All 4 major carriers cover the U.S. with only 4-‐8 clusters – Cellular DNS resolvers are placed at the same level as GGSN data centers
• Implica5on – Mobile content providers should place their content close to GGSNs – Mobile content providers should select the content server closest to the GGSN
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 55 Courtesy: Q. Xu et al.
![Page 56: Cellular’Networks’and’Mobile’ Compung’ …coms6998-8/lectures/lec9-cell...PreviousStudies • Cellular’IP’dynamics’ – Measured’cellular’IP’dynamics’attwo’locaons’](https://reader031.vdocuments.us/reader031/viewer/2022022503/5aaf9df27f8b9aa8438d84bf/html5/thumbnails/56.jpg)
Ques5ons?
Cellular Networks and Mobile Compu5ng (COMS 6998-‐8) 56