Title of Invention

A COMPUTER NETWORK AND A METHOD FOR PROVIDING NETWORK SERVICES TO MULTIPLE NETWORK CLIENT COMPUTERS

Abstract

A computer network includes one or more service computers configured to provide multiple network services via the network, and one or more connection devices that allow multiple network client computers to access the services via the network. The network also includes a single routing computer that serves as a firewall through which all traffic between the network services and the network client computers must pass.

Full Text

The present invention relates to a computer network in combination with one or more service computers configured to provide network services. The present invention also relates to a method for providing netv/ork services to multiple network client computers via s computer network. Administrators o f local area networks (LANs) frequently purchase network services, such as Internet access or network management services, from entities known as network service providers. To receive these network services, a customer must connect its LAN to the network service provider's LAN, which presents several security concerns. Among these concerns are unwanted infiltration of the network service provider's LAN by its customers and by other outside entities, unwanted access to a customer's LAN by other customers, and unwanted interception of mformation passing between the network service provider and its customers. The network service provider often deals with security concerns such as these by erecting firewalls or by providing dedicated connections to its service machines. In general, these solutions require a dedicated firewall for each customer or a private connection for each customer to each service machine. Several firewalling techniques are known in the art, see, for example, Hubbsird, S.D. and Sager, J.C. "Firewalling the Net", BT Technology Journal, vol 15, no. 2, April 1997, pages 94-106, and Chapman, D.B. and Zwicky, E.D., "Building Internet Firewalls", O'Reilly and Associates, Inc., 1995. Hubbard et al. disclose four firewalling techniques that can be used alone or in combination to restrict access between a trusted internal network and a hostile external network. These firewalling techniques include packet filtering or screening, application proxies, circuit-level gateway, and address or port translation. Packet screens operate at the level of the internet protocol (IP) packet. Packet screens are placed between the internal network and the external network and are programmed to selectively route packets to or from the internal network. Application proxies work at the level of the application protocol, rather than at the level of the network and transport protocols. An example of an application proxy is a World Wide W.eb (WWW) proxy that is able to decode the contents of a WWW packet so that it can be forwarded by the proxy to its ultimate destination. Application proxies are processes that are run on a machine known as a bastion host. Chapman et al. define a bastion host as "a computer system that must be highly secured because it is vulnerable to attack, usually because it is exposed to the Internet and is a main point of contact for users of internal networks." Chapman et al. disclose that multiple bastion hosts may be used in a firewall configuration. The circuit-level gateway is an extreme form of proxy server. It is similar to packet screen in that it works at the transport protocol level rather than the application protocol level. Circuit-level gateways set up "virtual circuits" between the internal and external hosts without interpreting the application protocol. In address or port translation, a gateway is placed at the interface of the internal network and the external network. When a packet destined for the outside network from a machine on the internal network arrives at the gateway, the gateway assigns the internal host an IP address taken from its global allocation pool. The gateway remembers the mapping and sends the packet onwards to its destination with the new source IP address. When the response is received, the gateway checks its mapping and rewrites the packet with its correct internal destination IP address before sending it out on to the internal network. Summarv of the Invention The invention is a computer network in combination with one or more service computers configured to provide multiple network services and firewalls for governing flow of traffic to and from the network services. The computer network comprises one or more connection devices that allow multiple network client computers to access the services.via the network and a logical connection device that creates logical broadcast domains among the network services and the network client computers, wherein at least two firewalls each implement a static route policy that governs flow of traffic between the network services and the network client computers and each of the firewalls provides independent routing within the network services. Each embodiment of the invention may provide one or more of several advantages. For example, a single security policy for a computer network may be distributed across multiple firewalls, which in turn may be managed remotely from virtually anywhere within the network. Multiple network services may be provided via the network, and multiple customers of the network service provider may access the network to receive these services. The network may support a variety of connection technologies, including frame relay, asynchronous transfer mode (ATM), Point-To-Point Protocol (PPP), Integrated Services Digital Network (ISDN) and Internet connections, which reduces or eliminates the need for customers to reconfigure their LANs. Moreover, the network may support multiple customers and multiple connectivity options with minimal network components and physical connections. Virtual LANs (VLANs) may be used to allow software and hardware reconfigurations of the network, including the relocation of service machines, without requiring customers to reconfigure their LANs and without revising the security policy. Static routing policies may be used in the firewalls to simplify management an improve security of the network. Accordingly, the present invention provides a computer network in combination with one or more service computers configured to provide network services, the computer network comprising one or more connection devices that allow a plurality of network client computers to access the network services via the computer network, wherein each of the plurality of network client computers is associated with a physical group of computers, and a logical connection device coupled between the service computers and the plurality of network client computers and configured to create logical broadcast domains among the network services and the plurality of network client computers, wherein each logical broadcast domain comprises a grouping of selected network services and at least one of the plurality of network client computers, wherein the logical broadcast domain is implemented using at least two routing devices each implementing a single static route policy that governs flow of traffic between the network services and the plurality of network client computers and prevents unauthorized access to the computer network, wherein each of the routing devices provides independent routing to the network services wherein at least two firewalls each implement a static route policy that governs flow of traffic between network service and the network client computers and each of the firewalls provide independent routing within the network service. The present invention also provides a method for providing network services to multiple network client computers via a computer network, the method comprising allowing the network client computers to access the network services via one or more connection devices in the computer network, creating logical broadcast domains among the network services and the network client computers, requiring all traffic within the network service to pass through a routing device that implements a static routing policy, and requiring all traffic between the network services and the network client computers to pass through two routing devices that each implement a static route policy. With reference to the accompanying drawings, in which : FIG. 1 is a block diagram of a network over which a network service provider delivers services securely to multiple customers. FIG. 2 is a block diagram of a network server computer used to implement a firewall. Detailed Description Referring to FIG. 1, a network service provider maintains a secure customer access network (SCAN) 10 that allows unrelated customers at remote physical sites 12, 14, 16, 18 to receive network services 20, such as network management, trouble ticketing and Internet access, through a common, secure architectural interface. The SCAN 10 employs network switching, virtual local area network (VLAN) and firewall technology to provide the network services 20 in a secure manner to a virtually unlimited number of customers using only a minimum of network components. The SCAN 10 is capable of providing the network services 20 over a wide variety of connections, including frame relay (FR) , private leased line, asynchronous transfer mode (ATM), Point-to-Point Protocol (PPP) and Integrated Services Digital Network (ISDN) connections. The SCAN 10 also includes a router 23, e.g., a Cisco 7000 router, that maintains a full Internet route table and that serves as a full Border Gateway Protocol (BGP) peer to several Internet service providers (ISPs) . This router 23 allows a SCAN customer to access the network services 20 and the customer's own network via the Internet 25 and to access the Internet 25 from the customer's network via the SCAN 10. Each customer site 12, 14, 16, 18 typically includes a computer subnetwork, e.g., a local area network (LAN) 15, and a router 17 capable of connecting the LAN 15 to other computer networks, including the Internet 25. Some of the network services 20 provided by the SCAN 10 may be implemented as executable programs running on programmable computers 11, 13, e.g., network server computers, in a subnetwork maintained by the network service provider. In general, each computer 11, 13 in the subnetwork is dedicated to providing one of the network services. The computers 11,13 may operate under different operating systems, e.g,, Unix and Windows NT, or they all may run under the same operating system. The network services may be provided as described in "Providing Network Services Through a Common Interface" in International Publication WO 99/46692. Some of the network services 20 process and generate information that is proprietary to individual customers, so the SCAN 10 must ensure that information exchanged between the group of network services 20 and any given customer cannot be accessed by anyone other than that customer. The network service provider also may want to prevent unauthorized communications between customers through the SCAN 10. To do so, the SCAN 10 recognizes each subnetwork of computers as a unique physical domain, or group, each of which must be protected from users in the other groups. In particular, the SCAN 10 treats each customer site 12, 14, 16, 18 as a unique group and treats any other subnetwork maintained by the network service provider, e.g., a subnetwork connecting administrative personnel, as a separate group, and it treats the Internet router 23 as a separate group. A logical connection device or LAN switch 30, e.g., a Xylan OmniSwitch, allows the SCAN 10 to create broadcast domains, known as virtual LANs (VLANs), among the various physical domains (groups). Each VLAN represents a logical connection created by the LAN switch 30 between computers located in different physical domains. In other words, the LAN switch 30, through VLANs, allows computers in"different logical networks to communicate with each other, via connections made either internally within the LAN switch 30 or externally through a security device, such as a firewall. Thus, the LAN switch 3 0 allows users at the customer sites 12, 14, 16/ 18 to access the network services 10 maintained by the network provider. The SCAN 10 may use several policies to define VLANs, including any of the following*. (1) a port-based policy, which assigns computers in the various groups to VLANs based on the physical ports to which they attach in the SCAN 10; (2) a media access control (MAC) address-based policy, which defines VLANs based on the physical layer addresses of the computers in the various groups; and (3) an Internet Protocol (IP) address policy, which defines VLANs based on the network layer addresses of the computers in the various groups. Even though each computer may belong to only one group, which is determined by the computer's physical location, each computer may belong to multiple VLANs. The SCAN 10 protects the groups and the information flowing between groups by requiring all group-to-group communications to pass through at least one of four firewalls 22, 24, 26, 28, each of which implements a single, static routing policy. The SCAN 10 further protects the group of network services by requiring all communications between this group and any other group to pass through two of the four firewalls. Referring also to FIG. 2, each of the firewalls may be implemented as a network server computer 50, e.g., a Sun Sparc workstation, running an executable program 68, such as Checkpoint's "Firewal1-1" software, that has been loaded from a fixed storage medium, e.g., a hard disk 66, into the computer's system memory 54. Each firewall routes TCP/IP (Transmission Control Protocol/Internet Protocol) data packets according to a static routing policy defined by a route table 70, which also may be stored in the hard disk 66. Packets that meet all of the conditions prescribed in a firewall' s route table are forwarded by the firewall to the appropriate destination; packets that do not meet the prescribed conditions are discarded. Implementing a static routing policy in each firewall ensures that no one other that the network service provider can change the routing policy within the SCAN 10, which in turn ensures that the network service provider' s, customers receive secure access to the SCAN 10. Also, because the routing policy is static, the firewalls do not propagate any routing information to the customer sites 12, 14, 16, 18 or to the Internet 25. Instead, the routers 17 in the customer sites are configured with static routes to the SCAN 10, as discussed below. Referring again to FIG. 1, each of the four firewalls 22, 24, 26, 28 implements a single route policy that pertains only to transactions involving certain physical domains (groups) . The first firewall 22 (Firewall A) protects the group of network services 20 from unwanted penetration by users in other groups by inspecting all traffic passing to and from the network services 20, including communications among the network services 2 0 themselves. The route policy implemented in the first firewall 22 includes two rulesets (listed as source/destination/service) : (1) a "service group/any/any" ruleset, which allows the network service provider to access any other group using any of the network services; and (2) a "customer/service group/service" ruleset, which allows any bona fide customer to access the group of network services 20 using any service for which the customer subscribes. The other three firewalls 24, 26, 28 serve to ensure that only bona fide customers of the network service provider are able to access the SCAN 10. Each of these firewalls provides access only to those customers that meet certain service-subscription criteria. For example, the second firewall 2 4 (Firewall B) provides access only to customers that access the network services 2 0 directly through the SCAN 10 and that do not receive Internet access from any source other than the network service provider. This type of customer is equipped with a static route to the second firewall 24 in its internal router. Customers that receive Internet service from a source other than the network service provider, i.e., through any source other than the SCAN 10, cannot access the second firewall 24. This limitation ensures that customers receiving Internet service from another source cannot send traffic improperly through the second firewall 24 instead of through the Internet. The route policy implemented in the second firewall 24 includes three rulesets: (1) a "service group/any/any" ruleset, which allows the network service provider to access any other group using any of the network services; (2) a "customer/firewall A/service" ruleset, which allows the customer to access, through the first firewall 22, the group of network services using any service for which the customer subscribes; and (3) a "customer/Internet/services" ruleset, which allows the customer to access the Internet via the Internet router 23 using any service for which the customer subscribes. The third ruleset may be defined to limit Internet access to certain users within the customer's physical domain or to certain services, e.g., e-mail only. The third firewall 2 6 (Firewall C) provides access only to customers that access the network services 20 directly through the SCAN 10 and that receive Internet service from a source other than the network service provider. This type of customer is equipped with a static route to the third firewall 26 in its internal router. The route policy implemented in the third firewall 26 includes two rulesets; (1) a "service group/any/any" ruleset, which allows the network service provider to access any other group using any of the network services; and (2) a "customer/firewall A/service" ruleset, which allows the customer to access, through the first firewall 22, the group of network services using any service for which the customer subscribes. The fourth firewall (Firewall D) 28 provides access to customers that receive any of the network services 2 0 via the Internet 25. The route policy implemented in the fourth firewall 28 includes two rulesets; (1) a "service group/any/any" ruleset, which allows the network service provider to access any other group using any of the network services; and (2) a "customer-via-Internet/firewall A/service" ruleset, which allows the customer to access, through the first firewall 22, the group of network services using any service for which the customer subscribes. Customers that access the SCAN 10 via the Internet 25 may experience reduced transmission bandwidth and extra delays beyond the control of the network service provider. As mentioned above, the customers of the network service provider can connect to the SCAN 10 using a variety of connection technologies, including frame relay, ATM, PPP, ISDN and Internet connections. For example, the first customer site (Customer I) 12 in FIG. 1 accesses the SCAN 10 through a PPP line 34 that terminates directly at either the second firewall 24 or the third firewall 26, depending on whether the customer 12 receives Internet service from any source other than' the network service provider, as discussed above. A PPP link typically involves a dedicated physical connection, or physical port, at a firewall and therefore requires the customer to lease the link from the network service provider. Each PPP link also accounts for one IP interface at the firewall. The second customer site (Customer II) 14 in FIG. 1 accesses the SCAN 10 through a link 36 to a frame relay 32 that terminates directly at either the second firewall 24 or the third firewall 26. In general, each link 38 from a frame relay to a firewall requires a dedicated synchronous port on the firewall, but since each frame relay link 38 can support multiple permanent virtual circuits (PVCs), and therefore multiple customers, each firewall can support a considerable number of customers via frame relay connections. The third customer site (Customer III) 16 in FIG. 1 accesses the SCAN 10 through an ISDN line 40 that terminates at an ISDN server 42, e.g., a Cisco 4500M server, within the SCAN 10. The ISDN server 42 connects physically to the LAN switch 3 0, which in turn forms a logical connection between the ISDN server 4 2 and either the second firewall 24 or the third firewall 26, depending upon whether the customer 16 receives Internet access from any source other than the network service provider. Because the ISDN server 4 2 is used to terminate ISDN links to the SCAN 10, the second and third firewalls each need dedicate only one IP interface to service all customers with ISDN links. The ISDN server 42 may provide additional security for the SCAN 10 by ending each ISDN call as soon as it begins, using the standard ISDN "Caller ID" feature to reestablish a connection with the caller, and then using an authentication protocol, such as the "Challenge Handshake Authentication Protocol" (CHAP), to verify that the caller is a bona fide customer. The fourth customer site (Customer IV) 18 in FIG. 1 accesses the SCAN 10 via a link 44 to the Internet 25. This customer also may access the SCAN 10 in other ways, e.g.. through a frame relay or PPP connection, via either the second firewall 24 or the third firewall 26, as discussed above. Referring again to FIG. 2, each firewall is implemented as a programmable computer 50 having, among other things, a central processing unit (CPU) 52, a memory controller 54, and a system memory 56 coupled to a system bus 58. The system memory 56 may include a random access memory (RAM) and a non-volatile memory e.g., a writable read-only memory such as a flash ROM. The computer 50 also includes a fixed storage medium, such as a hard disk 66, and a hard disk controller 64 coupled via an input/output (I/O) bus 62, which in turn is coupled to the CPU bus 58 by a bus interface device 60. The computer 50 may be preprogrammed, e.g. , in ROM, to serve as a firewall, or it may be programmed by loading an executable program 68 from a storage medium, such as the hard disk 66, a floppy disk or a CD-ROM, into system memory 56. The executable program 68 accesses a route table 70, which may be stored on the hard disk 50, to determine how to route information through the firewall. The computer 5 0 also includes a network interface controller 72 coupled to the I/O bus 62 which enables the computer 50 to connect to one or more computer networks. Other embodiments are within the scope of the following claims. WE CLAIM : 1. A computer network (10) in combination with one or more service computers (11, 13) configured to provide network services (20), the computer network (10) comprising one or more connection devices (34, 36, 40, 44) that allow a plurality of network client computers (12,14,16,18) to access the network services (20) via the computer network (10), wherein each of the plurality of network client computers (12, 14, 16, 18) is associated with a physical group of computers, and a logical connection device (30) coupled between the service computers (11, 13) and the plurality of network client computers (12, 14, 16, 18) and configured to create logical broadcast domains among the network services (20) and the plurality of network client computers (12, 14, 16, 18), wherein each logical broadcast domain comprises a grouping of selected network services (20) and at least one of the plurality of network client computers (12, 14, 16, 18), wherein the logical broadcast domain is implemented using at least two routing devices (17, 22) each implementing a single static route policy that governs flow of traffic between the network services (20) and the plurality of network client computers (12, 14, 16, 18) and prevents unauthorized access to the computer network (10), wherein each of the routing devices (17, 22) provides independent routing to the network services (20) wherein at least two firewalls each implement a static route policy that govems flow of traffic between network service (20) and the network client computers (12, 14, 16, 18) and each of the firewalls provide independent routing within the network service (20). The computer network as claimed in claim 1, wherein the logical connection device (30) is configured to allow access via a public frame relay (32). The computer network as claimed in claim 1, wherein the logical connection device (30) is configured to allow access via a PPP link (34). The computer network as claimed in claim 1, wherein the logical connection device (30) is configured to allow access via an ISDN link (40). The computer network as claimed in claim 1, wherein the logical connection device (30) is configured to allow access via the internet (25). The computer network as claimed in claim 1, wherein the static route policy is defined by a static route table. The computer network as claimed in claim 1, wherein the plurality of network cUent computers (12, 14, 16, 18) belong to multiple logical broadcast domains. The computer network as claimed in claim 1, wherein the logical broadcast domains allow communication between the plurality of network client computers (12,14,16,18) intemally within a switch (30). The computer network as claimed in claim 1, wherein the logical broadcast domains allow communication between the plurality of network client computers (12, 14, 16, 18) externally through the routing device (17, 22). A method for providing network services (20) to multiple network client computers (12, 14, 16, 18) via a computer network (10), the method comprising allowing the network client computers (12, 14, 16, 18) to access the network services (20) via one or more connection devices (34, 36, 40, 44) in the computer network (10), creating logical broadcast domains among the network services (20) and the network client computers (12, 14, 16, 18), requiring all traffic within the network service (20) to pass through a routing device (17, 22) that implements a static routing policy, and requiring all traffic between the network services (20) and the network client computers (12, 14, 16, 18) to pass through two routing devices (17, 22) that each implement a static route policy. The method as claimed in claim 10, further comprising allowing the network client computers (12, 14, 16, 18) to access the network via a public frame relay (32). The method as claimed in claim 10, further comprising allowing the network client computers (12, 14, 16,18) to access the network via a PPP link (34). The method as claimed in claim 10, further comprising allowing the network client computers (12, 14, 16, 18) to access the network via an ISDN link (40). The method as claimed in claim 10, further comprising allowing the network client computers (12, 14, 16, 18) to access the network via the internet (25). The method as claimed in claim 10, wherein the static route policy is defined by a static route table. The method as claimed in claim 10, wherein the network client computers (12, 14, 16, 18) belong to multiple logical broadcast domains. The method as claimed in claim 10, further comprising allowing communication between network client computers (12, 14, 16, 18) intemally within a switch (30) by the logical broadcast domains. The method as claimed in claim 10, further comprising allowing communication between network client computers (12, 14, 16, 18) externally through the routing device (17, 22) by the logical broadcast domains. A computer network substantially as herein described with reference to the accompanying drawings. A method for providing network services substantially as herein described with reference to the accompanying drawings.

Documents:

in-pct-2000-520-che-abstract.pdf

in-pct-2000-520-che-assignment.pdf

in-pct-2000-520-che-claims filed.pdf

in-pct-2000-520-che-claims grand.pdf

in-pct-2000-520-che-correspondence others.pdf

in-pct-2000-520-che-correspondence po.pdf

in-pct-2000-520-che-description complete filed.pdf

in-pct-2000-520-che-description complete grand.pdf

in-pct-2000-520-che-drawings.pdf

in-pct-2000-520-che-form 1.pdf

in-pct-2000-520-che-form 19.pdf

in-pct-2000-520-che-form 26.pdf

in-pct-2000-520-che-form 3.pdf

in-pct-2000-520-che-form 5.pdf

in-pct-2000-520-che-pct.pdf