distributed denial of services the problem, its solutions, and their problems

52
08/02/01 S. Felix Wu --UCCS Visit 1 Services the Problem, its Solutions, and their Problems Dr. S. Felix Wu Computer Science Department University of California, Davis http:// www.cs.ucdavis.edu/~wu/ [email protected]

Upload: tavi

Post on 15-Jan-2016

27 views

Category:

Documents


0 download

DESCRIPTION

Distributed Denial of Services the Problem, its Solutions, and their Problems. Dr. S. Felix Wu Computer Science Department University of California, Davis http://www.cs.ucdavis.edu/~wu/ [email protected]. Denial of Service attack beyond Authenticity, Authority, and Privacy. victims. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 1

Distributed Denial of Servicesthe Problem, its Solutions, and their Problems

Dr. S. Felix Wu

Computer Science Department

University of California, Davishttp://www.cs.ucdavis.edu/~wu/

[email protected]

Page 2: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 2

Denial of Service attackbeyond Authenticity, Authority, and Privacy

Computer system

finite resources--bandwidth, connections,buffer space….

attacker

victims

consume all or mostof the resources!

Services are Denied!

Page 3: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 3

Distributed DoSyahoo, ebay, msn,...

Slave

Master

Slave

Slave

Slave

SlaveSlave

Attack traffic aggregated!

Denial of Service!

Hundreds/thousands of Slaves simultaneously launch attacks!

no service ordegraded service

Page 4: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 4

The Plain DDoS Model (1999-2000)

Masters

Slaves

Victim

... ISP

.com::.

Attackerssrc: randomdst: victim

1,500 bytes per pkt~10K bits per pkt~100K pkts per second

2000 slaves50 pkts per second per slave0.5M bits per second

Page 5: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 5

Reflector

• Use a legitimate network server/client as the reflector to avoid being traced. (stepping stone).

Reflector

VictimSlave

Service Request Packetsrc: Victimdst: Reflector

Service Reply Packet src: Reflector dst: Victim

Page 6: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 6

The Reflective DDOS Model (2000)

Masters

Slaves

Victim

... ISP

.com::.

Reflectors

Attackerssrc: victim

dst: reflector

src: reflectordst: victim

Page 7: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 7

Internet Source Accountability

UCD

AOL

UUNet

Headersrc: AOLdst:UCD

Payload……………..

A

B

Page 8: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 8

Possible Solutions

• Stop it!!– egress/ingress filtering– aggregated-flow anomaly-based rate limiting

• ISP, dot-COM,...

• Trace it!!– where are the slaves and masters?

• Law enforcement agencies,...

Page 9: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 9

Ingress/egress filteringboosting source accountability

drop it or not??

Is the source IP address of this incoming IP packetvalid from this particular network interface???

1. Static configuration2. Routing table reverse look-up3. Routing information analysis (BGP/OSPF/RIP)

Net: 169.237.6.*

207.12.1.56

filtering policies

Page 10: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 10

Aggregate-Based Congestion Controlavoiding micro-flow management

RED buffer(Random Early Dropping)

50%80%

good for aggressivebut responsive TCP flows...

Page 11: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 11

Aggregate-Based Congestion Controlavoiding micro-flow management

50%80%

rate limiters

HighbandwidthAG-Flow?

yes

no

High-BandwidthAG-Flow Analyzer

E.g., all ICMP packetstoward dst: 169.237.6.*.

(1). How to determine the signature of an AG-Flow??(2). How to set the limited rate for an AG-Flow??

Page 12: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 12

Packet Tracing

• A transit router puts a mark in the data packets themselves. (like UPS/FedEx)– find the space in the packet to perform the

mark?

• A transit router puts a mark outside of the data packets. (I have seen it!!)– find the bandwidth in the Internet?

Page 13: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 13

Statistical Packet Marking

Masters

Slaves

Victim

... ISP

.com::.

Attackerssrc: randomdst: victim

Page 14: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 14

Marking procedure at router R: for each packet w let x be a random number from [0..1) if x < p then write R into w.start and 0 into w.distance else if w.distance == 0 then write R into w.end increment w.distance

A5A5 R9R9R8R8

R4R4

R7R7R6R6

R3R3 R 5R 5

R2R2

R1R1

A6A6

ver hlen TOS Total Length

Identification flags offset

Time to live Protocol Header checksum

Source IP address

Destination IP address

offset Distance Edge fragment

0 2 3 7 8 15

Page 15: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 15

Problems with Packet Marking

• 16 bits is unreliable and restrictive.– partial IP header information– weak authentication– inefficiency

• can not handle reflective DDoS.– require modification of TCP protocol stack

(and specification) -- not sure exactly how to do it completely and correctly.

Page 16: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 16

Masters

Slaves

Victim

... ISP

.com::.

Reflectors

Attackerssrc: victim

dst: reflector

src: reflectordst: victim

???

???

Page 17: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 17

ICMP Traceback• For a very small probability (about 1 in 20,000),

each router will send the destination a new ICMP message indicating the previous hop for that packet.

• Net traffic increase at endpoint is probably acceptable.

iTrace it or not??

Page 18: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 18

Original iTrace

Masters

Slaves

Victim

... ISP

.com::.

Attackerssrc: randomdst: victim

Page 19: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 19

iTrace in Reflective DDOS

Masters

Slaves

Victim

... ISP

.com::.

Reflectors

Attackerssrc: victim

dst: reflector

src: reflectordst: victim

Page 20: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 20

Improved ICMP Traceback

• For a very few packets (about 1 in 20,000), each router will send the destination and the source a new ICMP message indicating the previous hop for that packet.

• Net traffic increase at endpoint is probably acceptable.

Page 21: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 21

Reflector

VictimSlave

Service Request Packetsrc: Victimdst: Reflector

Service Reply Packet src: Reflector dst: Victim

sourceTracebackMessages

Who has spoofed me??

Page 22: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 22

Improved iTrace

Masters

Slaves

Victim

... ISP

.com::.

Reflectors

Attackerssrc: victim

dst: reflector

src: reflectordst: victim

Page 23: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 23

VictimISP

Service Request Packetsrc: Victimdst: www.yahoo.com

sourceTracebackMessages

Is that really me???

How can I tell??

Page 24: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 24

Maybe it is my friend...

Masters

Slaves

Victim

... ISP

.com::.

Attackerssrc: randomdst: victim

Are you sure that thisis from a slave or not?

customers

Page 25: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 25

Emitting a “relatively small” amount

Masters

Slaves

Victim

... ISP

.com::.

Attackerssrc: randomdst: victim

Page 26: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 26

iTrace Probability: 1/20,000

Attack traffic

Background traffic

For a router with “lots” of background traffic, it will takea long time before we really generate a “useful” iTrace.

Page 27: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 27

A Statistic Problem with iTrace

• Routers closer to the victims have higher probability to generate iTrace packets toward the true victims in the first N iTrace messages generated.

• Routers closer to the DDoS slaves might have relatively small probability (smaller than the routers around the victims) to generate “useful” iTrace packets fast enough.

Page 28: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 28

“Usefulness”

• Useful:– It carries attack packets.

• Valuable:– It carries attack packets from a router that is

very close to the original slaves.– We have not received the same “kind” of iTrace

messages before.– The iTrace messages are received fast.

Page 29: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 29

Three Types of Nodes

• DDoS victim with the intention to trace the slaves.

• DDoS victim without the intention.

• non-DDoS victims (assuming they do not have the intention as well -- and very likely they hope they won’t receive ones).

Page 30: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 30

Intention-driven iTrace

• Different destination hosts, networks, domains/ASs have different “intention levels” in receiving iTrace packets.– We propose to add one “iTrace-intention” bit.

• Some of them might not care about iTrace, and some of them might not be under DDoS attacks, for example.

Page 31: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 31

Issues

• How to determine the intention bit

• How to distribute the intention bits to routers globally?

• How to use the intention bits at each router?

Page 32: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 32

0

16

48

32

20

36

28

24

40

44

52

56

60

232221

252627

293031

414243

535455

636261

373839

474645

595857

64

80

96

112

84

88

92858687

89 90 91

939495

100

104

108

116

120

124

101 102 103

105

106

107

109 110 111

117

118

119

121 122 123

125

126

127

Test-bed topology

133

49

17

6581

97113

Page 33: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 33

packet-forwardingtable

DecisionModule

iTraceGeneration (1/20000)

BGP routingtable

packets

iTracegenerationbit, (1/20000)

intentionbits

iTrace/Intention-Driven iTrace architecture

Page 34: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 34

Processing Overhead

Processing for each data packet:1. if the iTrace flag bit is 1,

(1). send an iTrace message for this data packet.(2). reset the iTrace bit to 0.

1/20K iTrace message trigger occurs:1. Select and Set one iTrace bit in the forwarding table.

Page 35: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 35

152.1.23.0/24 0169.20.3.0/24 0192.1.0.0/16 0

207.3.4.183/20 0152.1.0.0/16 0155.0.0.0/16 0

152.1.23.0/24 0169.20.3.0/24 0192.1.0.0/16 0

207.3.4.183/20 0152.1.0.0/16 1155.0.0.0/16 0

(1).BeforeiTracetrigger:

(2).AfteriTracetrigger:

I(n) iTrace bit

Page 36: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 36

152.1.23.0/24169.20.3.0/24192.1.0.0/16

207.3.4.183/20152.1.0.0/16155.0.0.0/16

(3).AfteriTracesent:

000000

I(n) iTrace bit

Page 37: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 37

Page 38: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 38

Usefulness in MSMV

0

Page 39: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 39

How to distribute I(n)?• YABE: (Yet Another BGP Extension)

– For every BGP route update, we include I(n) as a new string in the community attribute:

• 0x[iTrace-Intention]:0x[0-1] (optional & transitive)

– These I(n) values will be forwarded or even aggregated by the routers who understand this new community attribute.

• aggregation: I(new) = max {I(n)}

– Rate-Limiting on Intention Update:• should not be more frequent than Keep-Alive messages.

• should not trigger any major route computation.

Page 40: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 40

Signaling (BGP extension)

AS500

AS 120

AS200

AS300

AS250

AS600

AS800

AS900

AS700

AS 100

IDS

Intention-bitupdate request

BGP updateprefix: 900attribute: Intend to receive iTrace

Page 41: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 41

Summary

• Improve the probability of “useful” iTrace.

• Require some “minor” changes to the router forwarding process.

• Require a new BGP community string.

• The amount of generated iTrace messages should be no more than the current iTrace proposal.

Page 42: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 42

DECIDUOUS• Reliably identify the source(s) of attack

packets. (Tracing)– Intrusion Detection, Response, Source

Identification.

• Collaborating with Edge Routers or Security Gateways that support IPSEC or other types of Tunnels – Utilize the IPSEC framework– Requirements for IPSEC Policy System– Interacting with IDS and IRS/FW.

Page 43: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 43

Spoofed IP Address

NCSU

AOL

UUNet

Headersrc: AOLdst:NCSU

Payload……………..

A

B

Page 44: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 44

IPSec Tunnel

NCSU

AOL

UUNet

Headersrc: AOLdst: NCSU

Payload……………..

A

B

Header + IPSecsrc: A SPI=0x104dst: B

Page 45: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 45

IPSEC/AH, tunnel mode

Router orSecurityGateway

IPSECModule

freeSWAN& Pluto

Depending on the results from both IDS andIPSEC modules as well as the nature of thedetected attack itself, the Deciduous daemonwill decide dynamically where to setup SAs.

Attacker’sTarget

Intrusion DetectionSystem

IPSECModule

IPsecPHIL/API

DeciduousDaemon

Every single SA that has beenor has not been used by the attackpacket will provide some locationinformation about the true source.

Page 46: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 46

Collaboration

Internet Core

NCSUISP

Attacker’sTarget

Intrusion DetectionSystem

IPSECModule

IPsecPHIL/API

DeciduousDaemon

Page 47: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 47

Tunnel Path NCSUISP

Attacker’sTarget

Intrusion DetectionSystem

IPSECModule

IPsecPHIL/API

DeciduousDaemon

DeciduousDaemon

Internet Core

Phase II-SA

Page 48: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 48

DECIDUOUS Testbed at SHANG LAB

Stone163

Stone 4

Sun 2

Hychang23

Redwing164

Squeeze175

Bone 177

Norwork1661 1

eth0152,1.75.163

5

2

4

3

eth1 eth2 eth1

eth0192.168.1.4

152.1.75.177eth0

10.0.0.0255..0.0.0

eth0152.1.75.166

eth0152.1.75.164

eth1

172.16.0.0255.255.0.0

192.168.5.0255.255.255.0

eth0192.168.1.2

eth2eth1

192.168.4.0255.255.255.0

eth1

eth1

192.168.2.0255.255.255.0

eth0152.1.75.175

eth2 eth1

192.168.1.3eth0

192.168.3.0255.255.255.0

eth2

• Simple Single Source• Simple multiple Sources• Coordinated Multiple Sources

Page 49: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 49

Results

Page 50: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 50

Magic Marks: concept

src/dst IP addresses the rest…..

an outgoing packet

src/dst IP addresses 128 bit digest

HMACselector

16 bit mark

src/dst IP addresses the rest…..16 bit markiTracemessage

either a SRC itraceor DST itrace...

Privatekey

Page 51: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 51

Magic Marks: design

src/dst IP addresses the rest…..

an outgoing packet

Src IP addressplus

N bits (N=8) ofthe dst IP address

128 bit digest

HMACselector

16 bit marks

Privatekey

Pre-compute theMarking tablewith 2N entries!

Mark Table look-up

Page 52: Distributed Denial of Services the Problem, its Solutions, and their Problems

08/02/01 S. Felix Wu --UCCS Visit 52

A scenariosrc/dst IP addresses the rest…..16 bit mark

dst iTracemessage

src/dst IP addresses the rest…..16 bit markverifymessage 16 bit mark

src

response (Y/N)