Methods and apparatus for protecting against overload conditions on nodes of a distributed network
First Claim
1. A method of responding to an overload condition at a network element (“
- victim”
) in a set of one or more potential victims on a network, the method comprising the steps of;
A. responsively to an indication of an anomalous traffic condition, initiating diversion of traffic destined for the victim by a first set of one or more network elements external to the set of one or more potential victims to a second set of one or more network elements external to the set of one or more potential victims,B. the element(s) of the second set filtering traffic diverted in step A (“
diverted traffic”
) and selectively passing a portion thereof to the victim,wherein said filtering step includes detecting any of (i) a traffic pattern that differs from an expected pattern and (ii) traffic volume that differs from an expected volume, said expected pattern and said expected volume being determined during a period in which the victim is not at an overload condition, andwherein said detecting step includes determining whether any of the traffic pattern and volume varies statistically significantly from any of the expected pattern and volume, respectively.
4 Assignments
0 Petitions
Accused Products
Abstract
Methods and apparatus for protecting against and/or responding to an overload condition at a node (“victim”) in a distributed network divert traffic otherwise destined for the victim to one or more other nodes, which can filter the diverted traffic, passing a portion of it to the victim, and/or effect processing of one or more of the diverted packets on behalf of the victim. Diversion can be performed by one or more nodes (collectively, a “first set” of nodes) external to the victim. Filtering and/or effecting traffic processing can be performed by one or more nodes (collectively, a “second set” of nodes) also external to the victim. Those first and second sets can have zero, one or more nodes in common—or, put another way, they may wholly, partially or not overlap. The methods and apparatus have application in protecting nodes in a distributed network, such as the Internet, against distributed denial of service (DDoS) attacks.
51 Citations
39 Claims
-
1. A method of responding to an overload condition at a network element (“
- victim”
) in a set of one or more potential victims on a network, the method comprising the steps of;A. responsively to an indication of an anomalous traffic condition, initiating diversion of traffic destined for the victim by a first set of one or more network elements external to the set of one or more potential victims to a second set of one or more network elements external to the set of one or more potential victims, B. the element(s) of the second set filtering traffic diverted in step A (“
diverted traffic”
) and selectively passing a portion thereof to the victim,wherein said filtering step includes detecting any of (i) a traffic pattern that differs from an expected pattern and (ii) traffic volume that differs from an expected volume, said expected pattern and said expected volume being determined during a period in which the victim is not at an overload condition, and wherein said detecting step includes determining whether any of the traffic pattern and volume varies statistically significantly from any of the expected pattern and volume, respectively. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 24, 25)
- victim”
-
17. A network element for use in protecting against an overload condition on a network, the network element comprising:
-
an input for receiving traffic diverted from the network, the traffic comprising flows of data packets having respective source addresses; a statistics module that is arranged to perform a statistical analysis of the diverted traffic so as to detect an anomalous pattern of a flow associated with at least one of the source addresses; a filter, which is operative, responsively to detection of the anomalous pattern, to block at least a portion of the data packets having the at least one of the source addresses; and an output coupled to the input for selectively passing on to further elements in the network traffic not blocked by the filter, wherein said statistical analysis comprises detecting any of (i) a traffic pattern that differs from an expected pattern and (ii) traffic volume that differs from an expected volume, said expected pattern and said expected volume being determined during a period in which the victim is not at an overload condition, and determining whether any of the traffic pattern and volume varies statistically significantly from any of the expected pattern and volume, respectively. - View Dependent Claims (18, 19)
-
-
20. A system for use in protecting against an overload condition on a network, the system comprising:
-
one or more network elements (“
guards”
) disposed on the network, each network element havingan input for receiving traffic from the network, a filter coupled to the input, the filter selectively blocking traffic originating from a source suspected as potentially causing the overload condition, a statistics module that is coupled to the filter and that identifies the traffic statistically indicative of having originated from the source suspected as potentially causing the overload condition, and an output coupled to the input for selectively passing on to further elements in the network traffic not blocked by the filter, one or more further network elements (“
diverters”
) disposed on the network and in communication with the guards, the further network elements selectively initiating, responsively to detection of an anomalous traffic condition, diversion to at least one of the guards traffic otherwise destined for a still further network element (“
victim”
) in a set of one or more potential victims on the network,wherein said statistics module performs statistical analysis comprising detecting any of (i) a traffic pattern that differs from an expected pattern and (ii) traffic volume that differs from an expected volume, said expected pattern and said expected volume being determined during a period in which the victim is not at an overload condition, and determining whether any of the traffic pattern and volume varies statistically significantly from any of the expected pattern and volume, respectively. - View Dependent Claims (21, 22, 23)
-
-
26. A method of responding to an overload condition at a network element (“
- victim”
) in a set of one or more potential victims on a network, the method comprising the steps of;A. responsively to an indication of an anomalous traffic condition, initiating diversion of traffic destined for the victim by a first set of one or more network elements external to the set of one or more potential victims to a second set of one or more network elements external to the set of one or more potential victims, B. the element(s) of the second set filtering traffic diverted in step A (“
diverted traffic”
) and selectively passing a portion thereof to the victim,wherein the initiating step includes effecting a path of traffic that differs from a path that traffic would otherwise take to the victim, wherein the first set of one or more network elements comprises network switches having respective ports, comprising at least one switch that is configured to route the traffic to the victim through a first port while the victim is not under attach, and wherein effecting the path comprises instructing the at least one switch to route the traffic destined for the victim through a second port, to which at least one of the network elements in the second set is coupled, wherein said filtering step includes detecting any of (i) a traffic pattern that differs from an expected pattern and (ii) traffic volume that differs from an expected volume, said expected pattern and said expected volume being determined during a period in which the victim is not at an overload condition, and wherein said detecting step includes determining whether any of the traffic pattern and volume varies statistically significantly from any of the expected pattern and volume, respectively.
- victim”
-
27. A method of responding to an overload condition at a network element (“
- victim”
) in a set of one or more potential victims on a network, the method comprising;diverting to a guard machine traffic destined for the victim, the traffic comprising flows of data packets having respective source addresses; performing a statistical analysis of the diverted traffic at the guard machine so as to detect an anomalous pattern of a flow associated with at least one of the source addresses; and responsively to detecting the anomalous pattern, preventing at least a portion of the data packets having the at least one of the source addresses from reaching the victim while passing to the victim at least some of the data packets from other source addresses, wherein said performing step includes learning an expected traffic pattern of the flows while said victim is not under attack and is not in an overload condition, and detecting an attack by determining that the anomalous pattern differs statistically significantly from the expected traffic pattern. - View Dependent Claims (28, 29, 30, 31, 32, 33, 34, 35)
- victim”
-
36. A method of responding to an overload condition at a network element (“
- victim”
) in a set of one or more potential victims on a network, the method comprising;coupling the victim to receive traffic from the network via a first port of a network switch; actuating the network switch to divert the traffic destined for the victim to a second port to which a guard machine is coupled; filtering the diverted traffic using the guard machine; and selectively passing at least a portion of the filtered traffic from the guard machine to the victim, wherein said filtering comprises performing statistical analysis comprising detecting any of (i) a traffic pattern that differs from an expected pattern and (ii) traffic volume that differs from an expected volume, said expected pattern and said expected volume being determined during a period in which the victim is not at an overload condition, and determining whether any of the traffic pattern and volume varies statistically significantly from any of the expected pattern and volume, respectively. - View Dependent Claims (37, 38, 39)
- victim”
Specification