Transcript ppt - Northwestern Networks Group

Measurement and Diagnosis of
Address Misconfigured P2P traffic
Zhichun Li, Anup Goyal, Yan Chen and Aleksandar
Kuzmanovic
Lab for Internet and Security Technology (LIST)
Northwestern Univ.
What is P2P address misconfiguration?
 Thousands of peers send P2P file downloading
requests to a “random” target (even not in the
P2P system) on the Internet
Peers
“random” target on the Internet
Address-misconfigured P2P traffic
2
Motivations
 P2P file sharing accounted for > 60% of traffic in USA
and > 80% in Asia
 P2P software DC++ has already been exploited by
attackers for DoS
 direct gigabit “junk” data per second to a victim host from
more than 150,000 peers
 End user perspective
 Involve innocent users in DDoS attacks unconsciously
 Anti-P2P arm-race
 Downloading performance
 ISP perspective
 Reduce unwanted traffic for “green” Internet
Get contacted by an ISP in Canada
 P2P developer perspective
 Identify the buggy software among a large number of
variances.
 Help design more robust P2P software
3
Outline
•
•
•
•
•
Motivation
Passive measurement results
P2PScope system design
Root cause diagnosis and analysis
Conclusion
4
Passive Measurement
• Honeynet/honeyfarm datasets
LBL
NU
GQ
Sensor
5 /24
10 /24
4 /16
Traces
901GB
916GB 49GB
Duration
47
16
26
months months days
• Events: # of unique sources > 100 in 6 hours
Scan traffic
removal
Target
identification
Event time
window extraction
5
Measurement Results
• Event characteristics:
– Usually involve thousands of peers on
average
– Duration: A few hours to up to a month
LBL
NU
eMule
143
416
BitTorrent
74
211
Gnutella
4
3
Soribada
6
0
Xunlei
12
0
VAgaa
1
1
6
Popularity
39%!
#connections(M)
• Growing Trend:
14
12
10
8
6
4
2
0
2004
2005
2006
2007
The total numbers of connections that match the P2P signatures.
• IP space: observed in three sensors in
five different /8 IP prefixes
7
Further Diagnosis
• Problems with passive measurement on
archived data
– Events have gone
– Hard to backtrack the propagation
– Root cause?
• Need a real-time backtracking and
diagnosis system!
8
Outline
•
•
•
•
•
Motivation
Passive measurement results
P2PScope system design
Root cause diagnosis and analysis
Conclusion
9
Design of P2PScope System
P2P-enabled
Honeynet
Backtracking
system
Root cause
inference
10100101011101
infohash; ‘abc.avi’
P2P payload
signature
based responder
Event
identification
Protocol parsing for
metadata
10
Design of P2P Doctor System
P2P-enabled
Honeynet
Backtracking
system
Root cause
inference
Server
...
Server
...
Server
Local
Crawler
Server
Server
Index Server (tracker)
Crawling
BT: top 100, eMule: 185
Peer Exchange
Protocol Crawling
DHT Crawling
11
Design of P2P Doctor System
P2P-enabled
Honeynet
Backtracking
system
Root cause
inference
• Track the information flow for suspicious
P2P software
• Track how honeynet IPs propagated in
P2P systems
• Peer routability checking
Totally
~7000
lines of Python, Perl and Bro
• Anti-P2P
analysis
• Hypothesis formulation and testing
12
Outline
•
•
•
•
•
Motivation
Passive measurement results
P2P Doctor system design
Root cause diagnosis and analysis
Conclusion
13
Diagnosis & Analysis
• Questions
– What is the root cause?
– Which peers spread misconfiguration?
– How is misconfiguration disseminated?
– How badly are individual clients affected?
• Results
– Data plane traffic radiation
– Detailed results focus on eMule and BitTorrent
14
Data Plane Traffic Radiation
1.2.3.4
Resource mapping
Peer
Exchange
Who has
avatar.avi?
DHT
Index
Server
1.2.3.4
15
eMule – Root Cause
• Byte ordering is the problem!
4.3.2.1
1.2.3.4
4.3.2.1
4.3.2.1
4.3.2.1
4.3.2.1
16
eMule – Root Cause
• Byte ordering is the problem!
– 61% of the reverse honeynet peers indeed
running eMule with the port number reported
– For the backtracked peers which is in the
unroutable IP space, 69.6% of them having
reverse IPs run eMule
• Locate bugs in source code
– At least aMule 2.1.0 (a popular eMule
alternative) has the byte order bug
17
eMule – Peers & Dissemination
• Which peers spread misconfiguration?
– 99.24% of misconfigured peers are normal peers
• How is the misconfiguration disseminated?
– Index Server? No
– Peer exchange? Yes
– DHT? No
• Percentage of bogus peers in eMule network?
– [12.7%, 25.0%] w/ a total of 37,079 backtracked
peers
18
BitTorrent – Root Cause I
• Anti-P2P companies deliberately inject
bogus peers!
– 20% of traffic we observed related to anti-P2P
peers
– Only return bogus peers or anti-P2P peers
– Using UTorrent peer exchange protocol to
disseminate
– Find a particular peer farm
• One /24 network, each IP run hundreds of peers
• Run Azureus 2.5.0.0 and IPs also run VMware
• Return peers even for non-existing file hashes.
19
BitTorrent – Root Cause II
• KTorrent also has a byte-order bug
– Discover using information flow tracking on KTorrent,
UTorrent and Azureus
– Identify the actual bug, report to KTorrent Developers
and get confirmed.
• Misconfiguration propagation
– [fully] KTorrent: all peers exchanged from others
– [partial] UTorrent: all peers that respond to TCP
handshaking
– [almost not] Azureus: all peers that respond to
BitTorrent handshaking.
20
Conclusions
• The first study to measure and diagnose largescale address misconfigured P2P traffic
• Find 39% Internet background radiation is caused
by address misconfiguration
– Popular in various P2P systems, increase 100% each
year for four years, and scattered in the IPv4 space
• For eMule, we found it is caused by network byte
order problem
• For BitTorrent
– Anti-P2P companies deliberately inject bogus peers
– KTorrent has a byte order bug
21
22