The present invention concerns the field of firewalls in a computer system, and more specifically the configuration of firewalls.
A firewall is a machine or group of machines that makes it possible to protect the junction between an internal network and an external network like the Internet against unauthorized, or even malicious, intrusions. It is noted that the Internet consists of a set of interconnected networks and machines around the world, allowing users throughout the world to share information.
The term “machine” in the present specification represents a very broad conceptual unit that includes hardware and/or software. The machines can be very diverse, such as workstations, servers, routers, specialized machines and gateways between networks.
All of the messages flowing between the internal and external network must pass through the firewall, which examines each message and blocks those that do not comply with given access control rules. The firewall is one element of a global security policy, integrated into an increasingly rich applicative environment and designed to protect computer resources.
Firewalls are used, in particular, to prevent unauthorized Internet users from accessing internal networks connected to the Internet, to give a user of an internal network secure access to the Internet, to separate a company's public machines allowing access to the Internet from its internal network, so as to create a partition in a given network so as to protect the partitioned segments of internal networks.
The firewall is embodied, for example, by a dedicated machine that controls access to the various machines of a given internal network.
To do this, the firewall controls which machines and/or which users and/or which services or applications of an internal network can access which machines and/or which users and/or which services or applications of an external network and vice versa.
Machines belonging to the Internet use the TCP/IP protocol. The firewall filters TCP/IP communications. The firewall manipulates applicative data, information transmitted in the part reserved for data in the headers of TCP/IP datagrams.
The filtering criteria are, to give a non-limiting example:
The complexity of a firewall configuration is illustrated by the following example, which can be applied to most of the partitioned architectures in enterprise networks.
Let us consider the case of an enterprise network comprising n firewalls named NW1, . . . , NWn connected to subnetworks.
We would like to apply a security policy according to which, in each subnetwork CCi, a workstation (client station) Ci is authorized to access a server Si located in a subnetwork SSi. The subnetworks CCi and SSi are connected to one and the same firewall NWi.
This example can, of course, be extended to include several workstations that are authorized to access several servers.
With conventional firewall configuration systems, administrators work in two ways:
Simplifying the configuration is a priority for a firewall administrator.
The current known solutions for attempting to resolve the problem of complexity in the configuration are the following.
There is a known system marketed under the name Net Partitioner and produced by the Solsoft company.
The Net Partitioner device allows the administrator to graphically represent his entire network, with the installation of the firewalls and the various servers and workstations that belong to it. The machines are represented by icons and their interconnections by lines connecting them.
The administrator also defines, in the form of arrows, the ways in which the machines can access other machines and the applications they host.
This solution makes it possible to define groups of computers, as well as rules for controlling access between these groups. On the other hand, the rules define the access of all the elements of a group to all the elements of another group, which complicates the configuration procedure.
The description of the system, (i.e. all of the machines present in the form of icons and their interconnections in the form of lines), and the specification of the rules applied to the system and represented in the form of arrows, are combined in the same graphical interface. The more machines, and the more connections between these machines, the system comprises, the more difficult it is for the administrator to describe the system via the interface.
Moreover, the Net Partitioner device does not provide for any transfer of rules from said device to the firewalls in question, or for any retrieval of the new security policy. The administrator himself must configure each of the firewalls from the results obtained by the Net Partitioner device.
Therefore, this solution does not make it possible to simplify the configuration procedure.
One object of the present invention is to simplify the configuration of a large number of firewalls.
In this context, the present invention offers a method for configuring a firewall in a computer system comprising objects, the objects for which an access control policy is established being called resources, characterized in that it groups the objects of the system into protection domains, each firewall ensuring the protection of an internal domain relative to an external domain, and applies to the firewall in question a rule for controlling access between a source resource and a destination resource only if said source and destination resources belong to the same protection domain.
The present invention also relates to the system for implementing said method.
Other characteristics and advantages of the invention will emerge in light of the following description, given as an illustrative and non-limiting example of the present invention, in reference to the attached drawings in which:
As shown in
The computer system 2 is distributed and comprises objects 3, users and firewalls 1. An object 3 is a very broad conceptual unit that includes hardware and/or software. The objects 3 can be very diverse, such as networks, subnetworks, workstations, servers, routers, specialized machines and gateways between networks, and applications. Only the components of the objects 3 of the system 2 that are characteristic of the present invention will be described, the other components being known to one skilled in the art. The objects 3 between which access control rules constituting the security policy of the system 2 are defined are called resources 4.
As represented in
Each of the two internal 5 and external 6 protection domains is constituted by zones 8 comprising one or more networks or subnetworks 9 of machines. A zone 8 is a part of the system 2 that is separated from the rest of the system by one or more firewalls. The zones 8 are connected to the firewall 1 in question by several network interfaces 10. The administrator 7 determines, for each zone 8 connected to each firewall, whether the zone 8 is inside the protection domain 5 of the firewall (internal zone) or whether it is outside it (external zone), i.e., whether it is directly protected by the firewall or whether it is a zone for providing a connection between the firewalls, or between the various protection domains, which is essentially the same thing.
In the exemplary embodiment illustrated in
A subnetwork of the “demilitarized zone” type is a buffer subnetwork, creating a sort of screen between an internal and external network in order to reinforce its protection.
Each firewall 1 is connected to a zone 8 of the external domain 6 comprising a so-called backbone network 13. The zone 8 of the external domain 6 comprising the network 13 is called the backbone zone. The backbone zone 8 constitutes the connection of the internal domain 5 to the rest of the network in question, and represents the outside of the domain 5 in question.
According to one development of the invention, the backbone zone 8 comprises a central configuration machine 14 from which the global configuration of the system 2 is performed. The global configuration of the system 2 can be performed, for example, as explained in the French patent application no. 2,802,662 filed on Dec. 2, 1999 by the present Applicant, the title of which is “METHOD AND DEVICE FOR CENTRALIZED FIREWALL CONFIGURATION IN A COMPUTER SYSTEM”, granted 21 Jun. 2001. The central configuration machine 14 offers a graphical interface 15 that allows the administrator 7 to perform said configuration. The graphical interface 15 is illustrated in
The present invention is described below in the embodiment of the system illustrated in
In the embodiment illustrated in
The administrator defines the properties of the firewall 1 in the following way. The firewall NW1 has three network interfaces 10, mentioned in the “Name” column 22 with the zones 8 indicated in the “Zone” column 23: a network interface NW1 with the zone of the subnetwork 11, a network interface NW1_dmz with the zone of the subnetwork DMZ1, and a network interface NW1_backbone with the backbone zone. The properties are similar for the firewalls NW2 and NW3. An “Address” column 24 in the table 21 indicates the addresses of the network interfaces whose names are located on the same lines.
An “Is External” column 25 of the zone table 21 makes it possible to specify, for each network interface 10, whether said network interface is attached to a zone 8 outside the protection domain 5 (the value “true”) or inside the protection domain (the value “false”).
In the example in question, the network interfaces NW1_dmz and NW1 are attached to zones 8 (subnetworks DMZ1, I1) inside the protection domain 5, while the network interface NW1_backbone (backbone network) is outside the protection domain (configuration similar for the firewalls NW2 and NW3).
Each firewall provides access control for both the communications between the domains 5 and the communications between the zones 8 inside the domain 5 for which it is responsible. One part of the security policy concerns access control between the domains; another part of the security policy concerns access control between the zones inside the domain controlled by the firewall.
The invention consists of defining an operation for factoring the access control rules constituting the access control policy so as to minimize the number of filtering rules to be declared by the administrator.
To this end, the administrator 7 joins into the same groups the objects 3 of the system 2 (in the example illustrated, workstations and servers) for which the same security policy is applied. In the example illustrated in
The administrator 7 can, for example, group the machines C1, C2, C3 into a group of workstations 26 and the machines S1, S2, S3 into a group of servers 27.
The invention consists of declaring, among the types of groups defined by the administrator, access control rules whose scope is limited to each firewall or extended to the system 2. The administrator specifies for the access control rules whether the scope is local to the firewall or global.
A rule of local scope defines the access relationships between the resources 4 of two groups, said resources belonging to the same protection domain. The local scope makes it possible to limit the rule to accesses inside the protection domain 5.
In the example mentioned above, a rule of local scope defines an access relationship of the group (C1, . . . , Cn) to the group (S1, . . . , Sn) involving an access from the resource Ci to the resource Si, without establishing a relationship of Ci to Sj, with j different from i. When there are several workstations and servers as see above, the principle is the same: the rule of local scope defines an access relationship of the group (C11, C12, . . . , C1K, . . . , Cn1, Cn2 . . . ) to the group (S11, S12, . . . , S1m, . . . , Sn1, Sn2 . . . ) using an access from the resource Cik to the resource Sim, without establishing a relationship of Cik to Sjm, with j different from i, no matter what k and m are.
A rule of global scope defines the possible access relationships between two groups in the system 2 as a whole.
A rule of global scope is saved and can always be used by the administrator to handle general cases of the security policy. Rules of global scope govern the access relationships of the group (C1, . . . , Cn) to the group (S1, . . . , Sn) and establish all the relationships of Ci to Sj, for i and j varying from 1 to n. When there are several workstations and servers as seen above, the rule of global scope defines an access relationship of the group (C11, C12, . . . , C1K, . . . , Cn1, Cn2 . . . ) to the group (S11, S12, . . . , S1m, . . . , Sn1, Sn2 . . . ) using an access from the resource Cik to the resource Sim no mater what i, j and m are.
The “local” or “global” scope attribute of each rule is attached to each rule in such a way that each firewall individually knows the scope of the rules.
In the embodiment illustrated in
As shown in
The administrator then defines, in the rule window 18, the rules of local or global scope. In the example illustrated in
The rules table 29 comprises a “Name” column 31 for identifying the access control rule, a “Source” column 32 for designating the source group of the rule, and a “Destination” column 33 for designating the destination group of the rule.
The scope of the rule is defined in a “Scope” column 34 and can have the values “LOCAL” for a local scope or “GLOBAL” for a global scope. In the example illustrated, the scope of the rule has the default value “GLOBAL.”
The method according to the present invention works in the following way:
When the firewall applies the access control (for example during an attempt to establish a connection), the firewall 1 analyzes the scope attribute of the rule governing the control of the current access.
If the rule is of global scope, it is applied without any additional control: access is authorized or denied based on the instructions given by the rule. This is a standard firewall operation.
If the scope of the rule is local, the firewall determines the incoming and outgoing network interfaces 10 for the current traffic and analyzes whether these network interfaces are attached to the internal 5 or external 6 domain.
If both the incoming and outgoing network interfaces 10 are attached to the internal domain 5, the current traffic is within the firewall's protection domain 5; the rule is therefore applied and the access is authorized or denied based on the instructions given by said rule.
If one of the two network interfaces 10 is attached to the external domain 6, the current traffic is not within the firewall's protection domain 5; the rule in question is not applicable for the profile of the current traffic.
In the example illustrated, no firewall connecting the domains D1, D2, D3 to one another has been provided. The invention is not concerned with linked domains. The interfaces associated with linked domains are automatically attached to an external domain, which means that the “Is External” column has the true value.
In the example illustrated in
During an access from the subnetwork I1 to the subnetwork DMZ1, the firewall NW1 determines that the traffic enters through the network interface 10 NW1 and leaves through the network interface 10 NW1
During an access from the subnetwork I1 to the subnetwork DMZ2, the firewall NW1 determines that the traffic enters through the network interface NW1 and leaves through the network interface NW1
In the same way, the firewall NW2 detects that the traffic in question enters through the network interface NW2
The present invention relates to the method for configuring a firewall 1 in a computer system 2 comprising objects 3, the objects 3 for which an access control policy is established being called resources 4, characterized in that it groups the objects 3 of the system into protection domains 5, 6, each firewall 1 ensuring the protection of an internal domain 5 relative to an external domain 6, and applies to the firewall in question a rule for controlling access between a source resource 4 and a destination resource only if said source and destination resources belong to the same protection domain 5 or 6.
The method determines the protection domain of the resources 4 by means of the network interfaces 10 of the firewall in question, interfaces through which the communications pass in order to reach said resources.
The method defines the zones 8 comprising networks or subnetworks; it associates the network interfaces 10 of the firewalls to which said zones are connected with an internal or external domain; it determines the incoming and outgoing network interfaces 10 of the current traffic; it analyzes whether said network interfaces are attached to an internal or external domain; it applies the rule only if both network interfaces are attached to the same internal domain 5, which corresponds to the fact that the resources belong to the same protection domain.
The method composes the groups of objects 3 for which the access control policy is identical and applies the rule between each of the resources of a source group and a destination group.
The method characterizes the rule with a local or global scope, and it applies the rule to the resources in question only if said resources belong to the same protection domain 5 or 6 when the scope of the rule is local, and applies the rule to all of the resources in question when the scope of the rule is global.
The present invention also concerns the device for implementing the method described above.
The present invention also relates to the device for configuring a firewall 1 in the computer system 2, characterized in that it comprises the central configuration machine 14 that makes it possible to group the objects 3 of the system into protection domains, each firewall 1 ensuring the protection of an internal domain 5 relative to an external domain 6, and to apply to the firewall in question a rule for controlling access between a source resource 4 and a destination resource only if said source and destination resources belong to the same protection domain 5 or 6.
The device comprises the graphical interface 15 from which an administrator 7 can enter the protection domains 5 and 6 and the access control roles.
The graphical interface allows the administrator 7 to define a local or global scope for the access control rule, and the machine 14 applies the rule to the resources in question only if said resources belong to the same protection domain 5 or 6 when the scope of the rule is local, and applies the rule to all of the resources in question when the scope of the rule is global.
While this invention has been described in conjunction with specific embodiments thereof, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art. Accordingly, the preferred embodiments of the invention as set forth herein, are intended to be illustrative, not limiting. Various changes may be made without departing from the true spirit and full scope of the invention as set forth herein and defined in the claims.
Number | Date | Country | Kind |
---|---|---|---|
99 16118 | Dec 1999 | FR | national |
Number | Name | Date | Kind |
---|---|---|---|
6182226 | Reid et al. | Jan 2001 | B1 |
6212558 | Antur et al. | Apr 2001 | B1 |
6880089 | Bommareddy et al. | Apr 2005 | B1 |
20020184525 | Cheng | Dec 2002 | A1 |
20040213258 | Ramamoorthy | Oct 2004 | A1 |
Number | Date | Country | |
---|---|---|---|
20020129142 A1 | Sep 2002 | US |