Methods, systems, and computer readable media for implementing indirect general packet radio service (GPRS) tunneling protocol (GTP) firewall filtering using diameter agent and signal transfer point (STP)

Abstract

A method for implementing indirect GTP firewall filtering includes using a signaling message routing node to dynamically populate an indirect GTP-C firewall filtering database with IMSIs and VPLMN IDs extracted from mobility management signaling messages for updating the locations of outbound roaming subscribers. The method further includes receiving a CCR-I message generated in response to a GTP-C message. The method further includes extracting an IMSI and a VPLMN ID from the CCR-I message. The method further includes accessing the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message. The method further includes determining that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database. The method further includes determining that a VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message. The method further includes, in response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, rejecting the CCR-I message.

Description

TECHNICAL FIELD

The subject matter described herein relates to implementing firewall functionality for GTP core (GTP-C) signaling traffic. More particularly, the subject matter described herein relates to methods, systems, and computer readable media for implementing indirect GTP firewall filtering to prevent fraud-based attacks using one or more Diameter agents and an STP without intercepting GTP-C roaming signaling.

BACKGROUND

GTP is a group of IP-based communications protocols used to carry GPRS traffic within global system for mobile communications (GSM), universal mobile telecommunication system (UMTS), long term evolution (LTE), and 5G networks. GTP-C is used within the evolved packet core (EPC) network for signaling between serving gateways (SGWs) and packet gateways (PGWs). GTP-C control plane messages are exchanged between SGWs and PGWs to communicate serving gateway capability information to the PGW, to create update, and delete GTP tunnels, and for path management.

Because the PGW is used for Internet traffic, it can be subject to fraud-based attacks from nodes that are impersonating SGWs serving outbound roaming subscribers. An outbound roaming subscriber is a subscriber of a service provider's network that is roaming in another service provider's network. Outbound roaming subscribers can be distinguished from inbound roaming subscribers where a subscriber of another network is roaming in a service provider's home network. Signaling relating to outbound mobile subscribers is particularly subject to fraud-based attacks because an attacker impersonating a serving gateway or mobility management entity (MME) serving a particular subscriber can impersonate the subscriber using the subscriber's international mobile subscriber identity (IMSI), which may not be difficult to obtain. Using the IMSI of a real subscriber, an attacker can establish GTP sessions with a packet gateway and, at a minimum, deny service to real subscribers. The attacker may also obtain subscriber information from the subscriber's home network. One possible way to guard against such attacks is to implement GTP-C firewall functionality at the PGW of the home network. However, implementing GTP-C firewall functionality at the PGW may be burdensome on the network operator in light of the number of PGWs that may be deployed on the network and also on the processing resources of the PGW. For example, the PGW, if equipped to screen GTP-C messages, may have to contact the home subscriber server (HSS) to verify if subscriber is roaming out and then determine whether or not to allow a GTP-C session from a particular MME or serving gateway (SGW) of that roaming network. Such processing would be burdensome on both the PGW and the HSS. The PGW would be required to intercept the GTP-C signaling, query the HSS, receive the response from the HSS, and determine whether to allow the GTP session based on the response. This would be non-standard PGW behavior, as there is no existing standard-defined interface between the PGW and the HSS. The HSS would be required to process queries and responses for every GTP-C-session from every PGW in the network.

Accordingly, there exists a need for implementing GTP firewall functionality without intercepting GTP-C roaming signaling and in a manner that reduces the processing burden on core network nodes.

SUMMARY

A method for implementing indirect GTP firewall filtering includes using a signaling message routing node to dynamically populate an indirect GTP-C firewall filtering database with IMSIs and VPLMN IDs extracted from mobility management signaling messages for updating the locations of outbound roaming subscribers. The method further includes receiving a CCR-I message generated in response to a GTP-C message. The method further includes extracting an IMSI and a VPLMN ID from the CCR-I message. The method further includes accessing the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message. The method further includes determining that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database. The method further includes determining that a VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message. The method further includes, in response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, rejecting the CCR-I message.

According to another aspect of the subject matter described herein, using a signaling message routing node to dynamically populate the indirect GTP-C firewall filtering database includes, at the signaling message routing node, receiving a Diameter update location request (ULR) message, extracting an IMSI and VPLMN ID from the Diameter ULR message, temporarily storing the IMSI and the VPLMN ID extracted from the Diameter ULR message, determining that updating of the location of the subscriber is successful, and in response to determining that the updating of the subscriber's location is successful, associating the VPLMN ID extracted from the Diameter ULR message with the IMSI extracted from the Diameter ULR message in the indirect GTP-C firewall filtering database.

According to yet another aspect of the subject matter described herein, the signaling message routing node comprises a Diameter edge agent. (DEA)

According to yet another aspect of the subject matter described herein, the signaling message routing node comprises a Diameter relay agent (DRA).

According to yet another aspect of the subject matter described herein, dynamically populating the indirect GTP-C firewall filtering database includes, at the signaling message routing node, receiving a mobile application part (MAP) update location request message, extracting an IMSI and VPLMN ID from the MAP update location request message, temporarily storing the IMSI and the VPLMN ID extracted from the MAP update location request message, determining that the updating of the subscriber's location is successful, and, in response to determining that the updating of the subscriber's location is successful, associating the VPLMN ID with the IMSI extracted from the MAP update location request message in the indirect GTP-C firewall filtering database.

According to yet another aspect of the subject matter described herein, the signaling message routing node comprises a signal transfer point (STP).

According to yet another aspect of the subject matter described herein, the indirect GTP-C firewall filtering database is implemented on a computing platform separate from the signaling message routing node.

According to yet another aspect of the subject matter described herein, the indirect GTP-C firewall filtering database is co-located with the signaling message routing node.

According to yet another aspect of the subject matter described herein, the indirect GTP-C firewall filtering database is located on a computing platform separate from the signaling message routing node and from a home location register (HLR) or home subscriber server (HSS).

According to yet another aspect of the subject matter described herein, the method for indirect GTP-C firewall filtering includes dynamically populating the GTP-C firewall filtering database with international mobile equipment identifiers (IMEs) extracted from mobility management signaling messages, extracting an IMEI value from the CCR-I message, and using the IMEIs in the GTP-C firewall filtering database to screen the CCR-I message.

According to yet another aspect of the subject matter described herein, a system for implementing indirect general packet radio service (GPRS) tunneling protocol (GTP) firewall filtering. The system includes an indirect GTP core (GTP-C) firewall filtering database. The system further includes at least one signaling message routing node configured to dynamically populate the indirect GTP-C firewall filtering database with international mobile subscriber identifiers (IMSIs) and visited public land mobile network identifiers (VPLMN IDs) extracted from mobility management signaling messages for updating locations of outbound roaming subscribers, receive a credit control request-initial (CCR-I) message generated in response to a GTP-C message, extract an IMSI and a VPLMN ID from the CCR-I message, access the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message, determine that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database, determine that a VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, and, in response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, reject the CCR-I message.

According to yet another aspect of the subject matter described herein, the at least one signaling message routing node is configured to dynamically populate the indirect GTP-C firewall filtering database by receiving a Diameter update location request (ULR) message, extracting an IMSI and VPLMN ID from the Diameter ULR message, temporarily storing the IMSI and the VPLMN ID extracted from the Diameter ULR message, determining that the updating of the subscriber's location is successful, and, in response to determining that the updating of the subscriber's location is successful, associating the VPLMN ID extracted from the Diameter ULR message with the IMSI extracted from the Diameter ULR message in the indirect GTP-C firewall filtering database.

According to yet another aspect of the subject matter described herein, the at least one signaling message routing node comprises a Diameter edge agent (DEA).

According to yet another aspect of the subject matter described herein, the at least one signaling message routing node comprises a Diameter relay agent (DRA).

According to yet another aspect of the subject matter described herein, the at least one signaling message routing node is configured to dynamically populate the indirect GTP-C firewall filtering database by receiving a mobile application part (MAP) update location request message, extracting an IMSI and VPLMN ID from the MAP update location request message, temporarily storing the IMSI and the VPLMN ID extracted from the MAP update location request message, determining that the updating of the subscriber's location is successful, and, in response to determining that the updating of the subscriber's location is successful, associating the VPLMN ID with the IMSI extracted from the MAP update location request message in the indirect GTP-C firewall filtering database.

According to yet another aspect of the subject matter described herein, the at least one signaling message routing node comprises a signal transfer point (STP) for dynamically populating the GTP-C firewall filtering database and a Diameter agent for receiving the CCR-I message generated in response to the GTP-C message, extracting the IMSI and the VPLMN ID from the CCR-I message, accessing the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message, determining that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database, determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, and, in response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, rejecting the CCR-I message.

According to yet another aspect of the subject matter described herein, the indirect GTP-C firewall filtering database is co-located with the signaling message routing node.

According to yet another aspect of the subject matter described herein, the indirect GTP-C firewall filtering database is located on a computing platform separate from the signaling message routing node and from a home location register (HLR) or home subscriber server (HSS).

According to yet another aspect of the subject matter described herein, the at least one signaling message routing node is configured to dynamically populate the GTP-C firewall filtering database with international mobile equipment identifiers (IMEIs) extracted from mobility management signaling messages, extract an IMEI value from the CCR-I message, and use the IMEIs in the GTP-C firewall filtering database to screen the CCR-I message.

According to yet another aspect of the subject matter described herein, a non-transitory computer readable medium having stored thereon executable instructions that when executed by a processor of a computer control the computer to perform steps is provided. The steps include using a signaling message routing node to dynamically populate an indirect general packet radio service (GPRS) tunneling protocol core (GTP-C) firewall filtering database with international mobile subscriber identifiers (IMSIs) and visited public land mobile network identifiers (VPLMN IDs) extracted from mobility management signaling messages for updating locations of outbound roaming subscribers. The steps further include receiving a credit control request-initial (CCR-I) message generated in response to a GTP-C message. The steps further include extracting an IMSI and a VPLMN ID from the CCR-I message. The steps further include accessing the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message. The steps further include determining that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database. The steps further include determining that a VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message. The steps further include, in response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, rejecting the CCR-I message.

The subject matter described herein may be implemented in hardware, software, firmware, or any combination thereof. As such, the terms “function” “node” or “module” as used herein refer to hardware, which may also include software and/or firmware components, for implementing the feature being described. In one exemplary implementation, the subject matter described herein may be implemented using a computer readable medium having stored thereon computer executable instructions that when executed by the processor of a computer control the computer to perform steps. Exemplary computer readable media suitable for implementing the subject matter described herein include non-transitory computer-readable media, such as disk memory devices, chip memory devices, programmable logic devices, and application specific integrated circuits. In addition, a computer readable medium that implements the subject matter described herein may be located on a single device or computing platform or may be distributed across multiple devices or computing platforms.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a network diagram illustrating exemplary messaging for an outbound roaming subscriber in a 4G network performing an update location procedure with the home network;

FIG. 2 is a network diagram illustrating exemplary messaging for validating a GTP-C create session request for an outbound roaming subscriber in the 4G network;

FIG. 3 is a network diagram illustrating exemplary messaging associated with an outbound roaming subscriber in a second generation/third generation (2G/3G) network performing an update location procedure with a home network;

FIG. 4 is a network diagram illustrating exemplary messaging associated with validating a GTP PDP context create request for an outbound roaming subscriber in a 2G/3G network;

FIG. 5 is a block diagram illustrating an exemplary architecture of a DEA/DRA/STP node for creating a GTP-C screening database and for using the database to implement GTP-C firewall functionality without intercepting GTP-C signaling;

FIG. 6 is a flow chart illustrating an exemplary process for dynamically populating entry to a GTP-C screening database; and

FIG. 7 is a flow chart illustrating an exemplary process for implementing GTP-C firewall functionality without intercepting GTP-C roaming signaling.

DETAILED DESCRIPTION

The subject matter described herein implements GTP-C firewall functionality using an indirect GTP-C firewall filtering database populated by a DEA, DRA, and/or STP and using the database to screen GTP-C traffic without intercepting GTP-C roaming signaling. Such a database provides for GTP-C-signaling-based fraud detection when an attacker tries to masquerade as a node serving a legitimate outbound roaming subscriber. Because the solution does not require interception of GTP-C roaming signaling traffic, PGW implementation is simplified. The solution described herein provides the capability to indirectly detect fraudulent GTP-C roaming signaling traffic at DEAs and DRAs based on Diameter messages sent to the DEAs and DRAs in response to GTP-C roaming signaling traffic. The DEAs and DRAs may receive create connection request messages generated in response to GTP-C session creation requests sent by attackers. The DEAs and DRAs may utilize subscriber PLMN information obtained either from Diameter location update signaling transactions maintained in an indirect GTP-C firewall filtering database or subscriber roaming information obtained from SS7 signal messaging traffic stored by an STP in the indirect GTP-C firewall filtering database.

As indicated above, GTP-C traffic is used for session management, information management, and location management, which enables UEs to access the internet. By providing an efficient screening mechanism for such traffic, core network security is enhanced and in a more efficient way than implementing such screening at the PGW based on GTP-C traffic.

FIG. 1 is a network diagram illustrating exemplary network nodes and messaging associated with an outbound roaming subscriber performing an update location procedure. Referring to FIG. 1, when a subscriber roams to a visited network the mobility management entity (MME) or serving GPRS support node (SGSN) 100 will initiate a Diameter update location transaction with the core network. In the illustrated example, MME/SGSN 100 sends an update location request to the subscriber's home network. The update location request is intended to update the location of the subscriber with a home subscriber server (HSS) 102 in the home network. In particular, according to third generation partnership project (3GPP) TS 29.272, the update location procedure is used between the MME and the HSS and between the SGSN and the HSS to update location information in the HSS. Table 1 shown below illustrates exemplary parameters that may be included in an update location request message.

TABLE 1
Update Location Request AVPs
	Mapping to
Information	Diameter Attribute
element name	Value Pair (AVP)	Cat.	Description
IMSI	User-Name	M	This information element shall contain the
	(See Internet		user IMSI, formatted according to
	Engineering Task Force		3GPP TS 23.003 [3], clause 2.2.
	(IETF) Request for
	Comments (RFC) 673
	3 [61])
Supported	Supported-	O	If present, this information element shall
Features	Features		contain the list of features supported
(See			by the origin host.
3GPP TS 29.2
29 [9])
Terminal	Terminal-	O	This information element shall contain
Information	Information		information about the user's mobile
(See 7.3.3)			equipment. Within this Information Element,
			only the IMEI and the Software-
			Version AVPs shall be used on the S6a/S6d
			interface.
ULR Flags	ULR-Flags	M	This Information Element contains a bit
(See 7.3.7)			mask. See 7.3.7 for the meaning of
			the bits.
Visited PLMN	Visited-PLMN Id	M	This IE shall contain the mobile country
Id			code (MCC) and the mobile network code
(See 7.3.9)			(MNC), see 3GPP TS 23.003 [3]. It may
			be used to apply roaming based features.
Equivalent	Equivalent-	O	This Information Element shall contain the
PLMN List	PLMN-List		equivalent PLMN list of which the
(See 7.3.151)			MME/SGSN requests the corresponding
			closed subscriber group (CSG) Subscription
			data.
RAT Type	RAT-Type	M	This Information Element contains the radio
(See 7.3.13)			access type the UE is using. See
			clause 7.3.13 for details.
SGSN number	SGSN Number	C	This Information Element contains the
(See 7.3.102)			integrated services digital
			network(ISDN)number of the SGSN, see
			3GPP TS 23.003 [3], It shall be present
			when the message is sent on the S6d
			interface and the SGSN supports LCS
			(using MAP based Lg interface) or short
			message service (SMS) functionalities or
			the Gs interface.
			It may be present when the message is
			sent on the S6a interface and the
			requesting node is a combined
			MME/SGSN.
Homogeneous	Homogeneous-	O	This Information Element, if present,
Support of IMS	Support-of-		indicates whether or not “IP multimedia
Voice Over PS	IMS-Voice-		subsystem (IMS) Voice over
Sessions	Over-PS-Sessions		”PS Sessions is supported homogeneously
			in all TAs or RAs in the serving
			node (MME or SGSN or combined
			MME/SGSN).
			The value “SUPPORTED” indicates that
			there is support for “IMS Voice over
			PS Sessions” in all TAs or RAs.
			The value “NOT_SUPPORTED” indicates
			that there is not support for “IMS
			Voice over PS Sessions” in any of the TAs
			or RAs.
Visited gateway mobile	GMLC Address	C	This Information Element shall contain, if
location center (V-			available, the IPv4 or IPv6 address
GMLC)			of the V-GMLC associated with the serving
Address			node.
Active access point	Active-access point	O	This Information Element, if present,
name (APN)	name (APN)		contains the list of active APNs stored by
			the MME or SGSN, including the identity of
			the packet date network gateway (PDN
			GW) assigned to each
			APN. For the case of explicitly subscribed
			APNs, the following information
			shall be present:
			Context-Identifier: context id of subscribed
			APN in use
			Service-Selection: name of subscribed
			APN in use- MIP6-Agent-Info: including
			PDN GW identity in use for subscribed APN
			Visited-Network-Identifier: identifies the
			PLMN where the PDN GW was
			allocated
			For the case of the Wildcard APN, the
			following information shall be present:
			Context-Identifier: context id of the
			Wildcard APN
			Specific-APN-Info: list of APN-in use and
			related PDN GW identity when the
			subscribed APN is the wildcard APN
			It may be present when MME or SGSN
			needs to restore PDN GW data in
			HSS due to a Reset procedure.
User equipment (UE)	UE-SRVCC Capability	C	This information element shall indicate if the
single radio voice call			UE supports or does not support
continuity (SRVCC)			the SRVCC capability and shall be present
Capability			if the MME or the SGSN supports
			SRVCC and this information is available to
			the MME or the SGSN.
MME Number	MME-Number for-mobile	C	This Information Element contains the ISDN
for MT SMS	terminated short		number of the MME to route SMS
	message service (MT-		to the UE through the MME, see 3GPP TS
	SMS)		23.003 [3].
			It shall be present when the MME supports
			SMS in MME and wishes to
			provide SMS in MME.
SMS Register	SMS-Register-	C	This information element is used to inform
Request	Request		the HSS if the MME or the SGSN
			needs to be registered for SMS, prefers not
			to be registered for SMS or has
			no preference. It shall be present when the
			MME supports SMS in MME and
			requests to be registered for SMS. It shall
			be present when the SGSN
			supports “SMS in SGSN” as defined in
			clause 5.3.18 in 23.060 [12], and
			requests to be registered for SMS.
SGs MME	SGs-MME Identity	O	This information element is used to inform
Identity			the HSS of the MME identity that
			the MME will use over the SGs interface.
			This information element shall be
			present, if the MME supports this
			information element and if the MME identity
			used over SGs is different from the MME
			Diameter identity used over S6a.
Coupled	Coupled-	O	This information element contains the
node's	Node-		Diameter identity of the coupled node
Diameter	Diameter-ID		(i.e. MME's Diameter identity for the SGSN
identity			and SGSN's Diameter identity for
			the MME) when the message is sent by the
			combined MME/SGSN.
			This information element may be present
			when the message is sent on the
			S6a/S6d interface and the requesting node
			is a combined MME/SGSN.
Adjacent	Adjacent-	O	This information element, if present, shall
PLMNs	PLMNs		contain the list of PLMNs where an
			UE served by the MME/SGSN is likely to
			make a handover from the PLMN
			where the MME/SGSN is located. This list
			is statically configured by the
			operator in the MME/SGSN, according to
			the geographical disposition of the
			different PLMNs in that area, the roaming
			agreements, etc . . .
Supported	Supported-	O	If present, this Information Element shall
Services	Services		contain attribute value pairs (AVPs)
(3GPP TS 29.			indicating details of the
336 [54])			services supported by the MME/SGSN.

In Table 1, it can be seen that the update location request message includes as a mandatory parameter the IMSI of the subscriber and the visited PLMN identifier. These parameters may be used to perform indirect GTP-C firewall filtering, as will be described in detail below. Another parameter that may be more information element that may be used to perform indirect GTP-C firewall filtering is the IMEI.

DEA 104 receives the update location request message from MME/SGSN 100 and may store the IMSI, visited PLMN (VPLMN) ID, and IMEI temporarily until the information is validated by the HSS. One reason that DEA 104 may not store these parameters in its Indirect GTP-C firewall filtering database initially is that the update location request message may be initiated by an attacker. Only after successful validation by the HSS will DEA 104 store the parameters identifying the location of an outbound roaming subscriber in its GTP-C screening database.

In step 2 in the call flow diagram, DEA 104 forwards the ULR message to core Diameter relay agent (DRA) 106. In step 3, core DRA 106 forwards the S6A ULA message to HSS 102.

HSS 102 may validate the ULR message and, if validation is successful, update the subscriber's location in a database maintained by HSS 102. In step 4, HSS 102 sends an update location answer (ULA) message indicating successful updating of the subscriber's location to core DRA 106. In step 5, core DRA 106 forward the S6A ULA message to DEA 104. In step 6, DEA 104 updates an indirect GTP-C firewall filtering database 108 with the IMSI, VPLMN ID, and optionally, the IMEI previously extracted from the ULR message. As will be described in more detail below, the records in indirect GTP-C firewall filtering database 108 will be used to screen GTP-C traffic without requiring interception of the GTP-C traffic. In step 7, DEA 104 forwards the ULA message to MME/SGSN 100.

It should also be noted that FIG. 1 also includes STP 110. STP 110 may also dynamically populate indirect GTP-C firewall filtering database 108 with subscriber location information obtained from SS7 signaling messages.

Table 2 shown below illustrates an example of an entry that may be populated in indirect GTP-C firewall filtering database 108 after the call flow illustrated in FIG. 2.

TABLE 2
Example Indirect GTP-C Firewall Filtering Database Record
	IMSI	VPLMN ID	IMEI
	IMSI_1	Visited Network X	IMEI_1

In Table 2, the record includes the IMSI for the outbound roaming subscriber, the identity of the visited network (i.e., the VPLMN ID), and the IMEI.

Once the database is populated with subscriber information, the database can be used to screen GTP-C messages. FIG. 2 is a network and message flow diagram illustrating the use of indirect GTP-C firewall filtering database 108 to indirectly screen GTP-C messages. Referring to FIG. 2, in step 1, an attacker masquerades as a serving gateway serving a fictitious outbound mobile subscriber. The purpose of such an attack may be to create a session between the attacker and a core network node such as packet gateway 202 to send attack traffic to the core network. The attacker initiates the attack by sending a GTP-C create session request message with an IMSI and a VPLMN ID of the serving network. In one example, the IMSI may be the IMSI of a subscriber that is actually provisioned in the network and the VPLMN ID may be a false network corresponding to the attacker, rather than the network currently serving the outbound roaming subscriber. The GTP-C create session request is sent to the home network packet gateway 202 as part of a packet data network session establishment procedure. The GTP-C create session request may include a user location information element which stores the VPLMN ID. If the GTP-C create connection request is a legitimate create connection request from a real outbound roaming subscriber, the VPLMN ID may be the ID of the VPLMN in which the subscriber is roaming. However, if the GTP-C create connection request originates from an attacker the VPLMN ID may be the VPLMN that identifies the network in which the attacker is located. An additional information element that may be included in the GTP-C create connection request is the IMSI, which contains the 15 digit identifier of the UE. In this example, it is assumed that the attacker has obtained the IMSI of a real outbound roaming subscriber and inserts a false VPLMN ID in the GTP-C create session request.

In step 2 of the message flow diagram PGW 202 receives the GTP-C create session request and, in response, formulates and sends a credit control request-initial (CCR-I) message addressed to policy and charging rules function (PCRF) 204. The CCR-I message is sent from the PGW to the PCRF in order to request policy and charging control rules for a bearer and to provision IP flow mobility routing rules. The CCR-I message contains a subscription-ID attribute value pair (ADP), which stores the IMSI of the subscriber. The CCR-I message also includes a 3GPP-location-info AVP, which stores an indication of the current location of the subscriber, such as the VPLMN ID of the network serving the subscriber. In this example, the VPLMN ID may be one inserted by SGW 200, rather than the actual VPLMN ID serving the subscriber.

In step 3 in the message flow diagram, DRA 106 performs a lookup in indirect GTP-C firewall filtering database 108 using the IMSI received in the CCR-I message. In this example, it is assumed that a record is present in indirect GTP-C firewall filtering database 108 and that a VPLMN ID is present in the record. Accordingly, DRA 106 retrieves the VPLMN ID corresponding to the IMSI from indirect GTP-C firewall filtering database 108. In step 4 in the message flow diagram, DRA 106 compares the VPLMN ID extracted from the database record with the VPLMN ID received in the CCR-I message. In this example, it is assumed that the VPLMN ID stored for the IMSI in database 108 is different from the VPLMN ID received in the CCR-I message. Accordingly, in step 5, DRA 108 sends a message to PGW 202 indicating a mismatch with the VPLMN ID or, alternatively, record not found if there is no record present in database 108 corresponding to the IMSI. The message from core DRA 106 to PGW 202 may be a credit control answer-initial (CCA-1) message with a result code indicating that the GTP-C create session request should be rejected. In step 6 of the message flow diagram, PGW 202 creates and sends a GTP-C create session response to SGW 200 with an error code indicating APN access denied-no subscription.

Thus, using the steps in FIG. 2, subscriber roaming data obtained from Diameter update location transactions is used to indirectly screen for fraudulent GTP-C traffic. Such an approach is advantageous over implementing screening directly at PGW 202 where the PGW 202 is required to interrogate the HLR or the HSS to obtain the subscriber's location or to store the subscriber's location locally at the PGW. In such an example, a DEA and/or a DRA can perform GTP-based fraud detection without intercepting GTP messages. The solution illustrated in FIGS. 1 and 2 avoids the need for a dedicated GTP firewall from mobile network operators. Instead, DEA 104 and/or DRA 106 indirectly implements a GTP firewall by blocking create connection request traffic that is generated in response to GTP-C create session request traffic.

Although in the example illustrated in FIGS. 1 and 2, the DEA dynamically populates indirect GTP-C firewall filtering database 108 with the subscriber's current location and DRA 106 uses the record in database 108 to perform the GTP-C firewall functionality, the subject matter described herein is not limited to such an implementation. In an alternate implementation, DEA 104 can dynamically populate indirect GTP-C firewall filtering database 108 with subscriber location information obtained from an update location transaction, and DEA 104 can screen the create connection traffic using the records stored in database 108.

In yet another alternate implementation, database 108 can be populated using subscriber location information obtained by a signal transfer point, such as STP 110. FIG. 3 is a network diagram illustrating exemplary messaging associated with populating indirect GTP-C firewall filtering database 108 using subscriber location information obtained by STP 110. Referring to FIG. 3, in line 1 of the message flow diagram, when a subscriber roams into a visited 2G or 3G network, the visitor location register (VLR) 300 serving a roaming subscriber in a visited network sends a mobile application part (MAP) update location request to a home location register (HLR) 302 in the subscriber's home network. The map update location request message includes the IMSI of the subscriber and the VPLMN ID identifying visited network x where the subscriber is currently roaming. STP 110 receives the map update location request and temporarily stores the IMSI and the VPLMN ID for the transaction. In line 2 of the message flow diagram, STP 110 forwards the MAP update location request to HLR 302. HLR 302 validates the update location request and, if validated, updates the subscriber's location in its local location database.

In line 3 of the message flow diagram, HLR 302 sends a MAP update location response indicating successful updating of the subscriber location to VLR 300. HLR 302 forwards the update location response to STP 110.

In step 4 of the message flow diagram, STP 110 updates indirect GTP-C firewall filtering database 108 with the IMSI and the VPLMN ID previously stored by STP 110 in response to receiving the map update locations request. Updating indirect GTP-C firewall filtering database 108 may include determining whether a record exists corresponding to the IMSI. If a record exists corresponding to the IMSI, STP 110 may update the VPLMN ID in the record with the VPLMN ID received in the map update location request message. If database 108 does not include a record corresponding to the IMSI, STP 110 may create a new record in the database mapping the IMSI to the VPLMN ID received in the update location request message. STP 110 may also store the IMEI in the record. After being updated or newly created the record may appear as illustrated above in Table 1.

Like the example illustrated in FIG. 1, STP 110 only populates database 108 with the mapping between the VPLMN ID and the IMSI upon receiving confirmation from the HLR that the update location transaction was successful. The reason for waiting until receiving confirmation that the update location transaction is successful is the reduce the likelihood that an attacker can impersonate VLR 300 and populate database 108 with false location information for the subscriber.

In addition, in the example illustrated in FIG. 3, database 108 is populated by an STP. It is understood that database 108 may be populated by a physical or a virtual STP without departing from the scope of the subject matter described herein. Such an STP may be a physical on premises node residing in a service provider's network or a virtual STP that resides in a network cloud that is hosted by the service provider or by a cloud service provider.

FIG. 4 is a network and message flow diagram illustrating exemplary messaging used to screen GTP-C traffic for outbound roaming subscribers in 2G/3G network. Referring to FIG. 4, an attacker masquerades as an SGSN 400 serving an outbound mobile subscriber in a 2G or 3G visited network. The attacker starts the attack in step 1 by sending a GTP-C create PDP context request message to GGSN 402 located in the subscriber's home network. The GTP-C create PDP context request includes the IMSI and VPLMN ID that in this example identifies the network of the attacker, rather than the VPLMN ID currently serving the subscriber corresponding to the IMSI.

GGSN 402 receives the GTP-C create PDP context request message and formulates and sends a CCR-I message to PCRF 204. The CCR-I message includes the IMSI, the VPLMN ID, and the IMEI from the GTP-C message. GGSN 402 forwards the CCR-I message to DRA 106.

In step 4 of the message flow diagram, DRA 106 compares the VPLMN ID stored in indirect GTP-C firewall filtering database 108 corresponding to the IMSI with the VPLMN ID extracted from the CCR-I message. In this example, it is assumed that there is a mismatch between the IMSI and the VPLMN ID. Accordingly, in line 5, DRA 106 rejects the CCR-I message by sending a CCA-I message with a 5XXX response code to GGSN 402. In response to receiving the CCA-I message with the rejection response code, GGSN 402 sends a create PDP context response to attacker 400 indicating APN access denied-no subscription. Accordingly, using the steps illustrated in FIG. 4, a database populated by an STP can be used to screen attackers masquerading as SGSNs in visited 2G or 3G networks.

FIG. 5 is a block diagram illustrating an exemplary internal architecture for DEA/DRA/STP 104, 106, or 110 for implementing the subject matter described above with regard to FIGS. 1 through 4. In the example illustrated in FIG. 5, DEA/DRA/STP 104, 106, or 110 includes Diameter routing as well as SS7 routing capabilities. It is understood that the functions could be implemented in the same real or virtual network node or could be implemented in separate network nodes. Referring to the architecture illustrated in FIG. 5, DEA/DRA/STP 104, 106, or 110 includes a plurality of message processors (MPs) 500, 502, 504, and 506. Each message processor 500, 502, 504, and 506 includes at least one processor 508 and memory 510. Each message processor 500, 502, 504, and 506 may be implemented using a printed circuit board and corresponding traces for interconnecting the components mounted on the circuit board. Message processors 500, 502, 504, and 506 may communicate with each other using an internal communications medium 512, which in one example is an Ethernet communications medium.

Message processor 500 implements DEA and/or DRA functionality. Accordingly, message processor 500 includes a Diameter protocol stack 514 that implements diameter connection and routing functionality. Thus, Diameter stack 514 may initiate or respond to Diameter connections with diameter peers and route messages based on diameter layer information it the messages.

Message processor 502 implements SS7 routing functionality. Accordingly, message processor 502 includes an SS7 protocol stack 516 for routing SS7 messages based on message transfer part (MTP) level 3 information in the messages. SS7 stack 516 may also implement SIGTRAN protocols for carrying SS7 messages over IP networks.

Message processor 504 and 506 each implement GTP-C firewall functionality using indirect GTP-C firewall filtering database 108. In the illustrated example, message processors 504 and 506 may be identically provisioned with a copy of GTP firewall database 108, and message processors 500 and 502 may load balance messages requiring GTP-C firewall screening between message processors 500 and 506. In addition, each message processor 504 and 506 includes an indirect GTP-C firewall filtering database controller/screener 518 for screening create connection request messages using subscriber location information populated in database 108 and for dynamically populating database 108 using information received from Diameter or SS7 update location messages for roaming subscribers.

FIG. 6 is a flow chart illustrating an exemplary process for dynamically populating the indirect GTP-C firewall filtering database. Referring to FIG. 6, in step 600, a mobility management for updating of a location of an outbound roaming subscriber is received. For example, if the network uses Diameter to update the location of a subscriber, the mobility management message may be an update location request message from an MME or SGSN serving a roaming mobile subscriber for updating the subscriber's location stored in the HSS. If the network uses SS7 messaging to update the subscriber's location, the mobility management message may be a MAP update location request from a VLR for updating of the subscriber's location with an HLR.

In step 602, the process includes extracting and temporarily storing the VPLMN ID and the IMSI from the message. For example, DEA 104, DRA 106, or STP 110 may temporarily store the IMSI and the VPLMN ID of the subscriber extracted from the Diameter or SS7 update location request in memory local to DEA 104, DRA 106, or STP 110. In step 604, it is determined whether the update location was successful. For example, DEA 104, DRA 106, or STP 110 may receive an update location answer or response message from an HLR or HSS indicating successful or unsuccessful updating of a subscriber's location with the HLR or HSS. If the update location answer or request message indicates that the updating of the subscriber's location was not successful, control proceeds to step 606 where the IMSI and the VPLMN ID store in step 602 are discarded.

If, in step 604, it is determined that the update location transaction was successful, control proceeds to step 608 where the indirect GTP-C firewall filtering database is accessed using the IMSI. For example, DEA 104, DRA 106, or STP 110 may access indirect GTP-C firewall filtering database 108 using the IMSI extracted from an update location message.

In step 610, it is determined whether a record is present in the database. If a record is present in the database, control proceeds to step 612 where the record is updated with the VPLMN ID from the message. If a record is not present, control proceeds to step 614 where a new record is added mapping the IMSI to the VPLMN ID from the message.

FIG. 7 illustrates an exemplary process where using a dynamically populated indirect GTP-C firewall filtering database to implement GTP-C fraud detection and screening. Referring to FIG. 7, in step 700, a CCR-I message generated in response to a GTP-C create context request message is received. For example, DEA 104 or DRA 106 may receive a CCR-I message relating to a GTP-C create context request from a legitimate subscriber or from an attacker.

In step 702, the process includes extracting the IMSI, VPLMN ID, and IMEI from the message. In step 704, the indirect GTP-C firewall filtering database is accessed using the IMSI.

In step 706, if a record is present, control proceeds to step 708 where it is determined whether the VPLMN ID from the message matches the VPLMN ID in the database record. If the VPLMN ID does not match the VPLMN ID in the database record, control proceeds to step 710 where the CCR-I message is rejected. If the VPLMN ID in the message matches the VPLMN ID stored for the IMSI in the database, control proceeds to step 712 where the CCR-I message is forwarded to the PCRF. For example, DEA 104 or DRA 106 may forward the CCR-I message to PCRF 204. PCRF 204 may determine the appropriate policy for the session and respond with a credit control request-answer (CCR-A) message indicating successful establishment of the session. DEA 104 or DRA 106 may forward the CCR-A message to the gateway GPRS support node (GGSN) that sent the CCR-I message. The GGSN may respond to the GTP-C message indicating successful creation of a PDP context.

Thus, using the process described herein, a dynamically populated indirect GTP-C firewall filtering database may be accessible by DRA, a DEA, and/or an STP and used to indirectly implement a GTP-C firewall. Such an implementation does not require that the GTP-C signaling traffic be intercepted or that the PGW implement GTP-C screening functionality. In addition, because the indirect GTP-C firewall filtering database is dynamically provisioned based on answer messages received from a subscriber's home HLR or HSS, the labor required to populate the database is reduced over manual population methods.

The disclosure of each of the following references is incorporated herein by reference in its entirety:

REFERENCES

• 1. 3GPP TS 29.002, “Technical Specification Group Core Network and Terminals; Mobile Application Part (MAP) specification,” (Release 15) V15.5.0 (2019-06).
• 2. 3GPP TS 29.212, “Technical Specification Group Core Network and Terminals; Policy and Charging Control (PCC); Reference points,” (Release 16) V16.1.0 (2019-09).
• 3. 3GPP TS 29.272, “Technical Specification Group Core Network and Terminals; Evolved Packet System (EPS); Mobility Management Entity (MME) and Serving GPRS Support Node (SGSN) related interfaces based on Diameter protocol,” (Release 16) V16.0.0 (2019-09).

It will be understood that various details of the presently disclosed subject matter may be changed without departing from the scope of the presently disclosed subject matter. Furthermore, the foregoing description is for the purpose of illustration only, and not for the purpose of limitation.

Claims

1. A method for implementing indirect general packet radio service (GPRS) tunneling protocol (GTP) firewall filtering, the method comprising: using a signaling message routing node to dynamically populate an indirect GTP core (GTP-C) firewall filtering database with international mobile subscriber identifiers (IMSIs) and visited public land mobile network identifiers (VPLMN IDs) extracted from mobility management signaling messages for updating locations of outbound roaming subscribers;receiving a credit control request-initial (CCR-I) message generated in response to a GTP-C message;extracting an IMSI and a VPLMN ID from the CCR-I message;accessing the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message;determining that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database;determining that a VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message; andin response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, rejecting the CCR-I message.

2. The method of claim 1 wherein using a signaling message routing node to dynamically populate the indirect GTP-C firewall filtering database includes, at the signaling message routing node: receiving a Diameter update location request (ULR) message for updating a location of a subscriber with a home network of the subscriber;extracting an IMSI and VPLMN ID from the Diameter ULR message;temporarily storing the IMSI and the VPLMN ID extracted from the Diameter ULR message;determining that the updating of the location of the subscriber with the home network of the subscriber is successful; andin response to determining that the updating of the subscriber's location is successful, associating the VPLMN ID extracted from the Diameter ULR message with the IMSI extracted from the Diameter ULR message in the indirect GTP-C firewall filtering database.

3. The method of claim 2 wherein the signaling message routing node comprises a Diameter edge agent.

4. The method of claim 2 wherein the signaling message routing node comprises a Diameter relay agent (DRA).

5. The method of claim 1 wherein dynamically populating the indirect GTP-C firewall filtering database includes, at the signaling message routing node: receiving a mobile application part (MAP) update location request message for updating a location of a subscriber with a home network of the subscriber;extracting an IMSI and VPLMN ID from the MAP update location request message;temporarily storing the IMSI and the VPLMN ID extracted from the MAP update location request message;determining that the updating of the location of the subscriber with the home network of the subscriber is successful; andin response to determining that the updating of the location of the subscriber with the home network of the subscriber is successful, associating the VPLMN ID with the IMSI extracted from the MAP update location request message in the indirect GTP-C firewall filtering database.

6. The method of claim 5 wherein the signaling message routing node comprises a signal transfer point (STP).

7. The method of claim 6 wherein the indirect GTP-C firewall filtering database resides on a computing platform separate from the signaling message routing node.

8. The method of claim 1 wherein the indirect GTP-C firewall filtering database is co-located with the signaling message routing node.

9. The method of claim 1 wherein the Indirect GTP-C firewall filtering database is located on a computing platform separate from the signaling message routing node and from a home location register (HLR) or home subscriber server (HSS).

10. The method of claim 1 further comprising: dynamically populating the GTP-C firewall filtering database with international mobile equipment identifiers (IMEIs) extracted from mobility management signaling messages;extracting an IMEI value from the CCR-I message; andusing the IMEIs in the GTP-C firewall filtering database to screen the CCR-I message.

11. A system for implementing indirect general packet radio service (GPRS) tunneling protocol (GTP) firewall filtering, the system comprising: at least one memory;an indirect GTP core (GTP-C) firewall filtering database; andat least one signaling message routing node configured to dynamically populate the indirect GTP-C firewall filtering database with international mobile subscriber identifiers (IMSIs) and visited public land mobile network identifiers (VPLMN IDs) extracted from mobility management signaling messages for updating locations of outbound roaming subscribers, receive a credit control request-initial (CCR-I) message generated in response to a GTP-C message, extract an IMSI and a VPLMN ID from the CCR-I message, access the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message, determine that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database, determine that a VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, and, in response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, reject the CCR-I message.

12. The system of claim 11 wherein the at least one signaling message routing node is configured to dynamically populate the indirect GTP-C firewall filtering database by: receiving a Diameter update location request (ULR) message;extracting an IMSI and VPLMN ID from the Diameter ULR message;temporarily storing the IMSI and the VPLMN ID extracted from the Diameter ULR message;determining that the updating of the subscriber's location is successful; andin response to determining that the updating of the subscriber's location is successful, associating the VPLMN ID extracted from the Diameter ULR message with the IMSI extracted from the Diameter ULR message in the indirect GTP-C firewall filtering database.

13. The system of claim 12 wherein the at least one signaling message routing node comprises a Diameter edge agent.

14. The system of claim 12 wherein the at least one signaling message routing node comprises a Diameter relay agent (DRA).

15. The system of claim 11 wherein the at least one signaling message routing node is configured to dynamically populate the indirect GTP-C firewall filtering database by: receiving a mobile application part (MAP) update location request message;extracting an IMSI and VPLMN ID from the MAP update location request message;temporarily storing the IMSI and the VPLMN ID extracted from the MAP update location request message;determining that the updating of the subscriber's location is successful; andin response to determining that the updating of the subscriber's location is successful, associating the VPLMN ID with the IMSI extracted from the MAP update location request message in the indirect GTP-C firewall filtering database.

16. The system of claim 15 wherein the at least one signaling message routing node comprises a signal transfer point (STP) for dynamically populating the GTP-C firewall filtering database and a Diameter agent for receiving the CCR-I message generated in response to the GTP-C message, extracting the IMSI and the VPLMN ID from the CCR-I message, accessing the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message, determining that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database, determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, and, in response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, rejecting the CCR-I message.

17. The system of claim 11 wherein the indirect GTP-C firewall filtering database is co-located with the at least one signaling message routing node.

18. The system of claim 11 wherein the indirect GTP-C firewall filtering database is located on a computing platform separate from the at least one signaling message routing node and from a home location register (HLR) or home subscriber server (HSS).

19. The system of claim 11 wherein the at least one signaling message routing node is configured to: dynamically populate the indirect GTP-C firewall filtering database with international mobile equipment identifiers (IMEIs) extracted from mobility management signaling messages;extract an IMEI value from the CCR-I message; anduse the IMEIs in the GTP-C firewall filtering database to screen the CCR-I message.

20. A non-transitory computer readable medium having stored thereon executable instructions that when executed by a processor of a computer control the computer to perform steps comprising: using a signaling message routing node to dynamically populate an indirect general packet radio service (GPRS) tunneling protocol core (GTP-C) firewall filtering database with international mobile subscriber identifiers (IMSIs) and visited public land mobile network identifiers (VPLMN IDs) extracted from mobility management signaling messages for updating locations of outbound roaming subscribers;receiving a credit control request-initial (CCR-I) message generated in response to a GTP-C message;extracting an IMSI and a VPLMN ID from the CCR-I message;accessing the indirect GTP-C firewall filtering database using the IMSI extracted from the CCR-I message;determining that a record corresponding to the IMSI is present in the indirect GTP-C firewall filtering database;determining that a VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message; andin response to determining that the VPLMN ID in the record does not match the VPLMN ID extracted from the CCR-I message, rejecting the CCR-I message.