Automatic discovery of switch devices in a network

Details Automatic discovery of switch devices in a network Automatic discovery of switch devices in a network

2000-03-31

2004-09-21

Kizou, Hassan (Department: 2662)

Multiplex communications

Network configuration determination

Using a particular learning algorithm or technique

C370S390000

Reexamination Certificate

active

06795403

ABSTRACT:

FIELD OF THE INVENTION
The present invention generally relates to data processing. The invention relates more specifically to methods and apparatus providing automatic discovery of switch devices in a computer network.
BACKGROUND OF THE INVENTION
Computer networks are widely used in business, education and government to interconnect end stations, such as workstations and personal computers, and to share resources, such as storage devices and servers. Networks may carry data, voice, video, or hybrid messages and information (“traffic”). A network may be organized as a local area network (LAN), a wide area network (WAN), an internetwork that connects one or more LANs or WANs, a campus network, enterprise network, etc. Intermediate network devices such as routers, bridges, switches and gateways are used to route traffic between end stations, between networks, and from one network device to another. Basic information about networks and internetworking is provided in D. Comer, “Computer Networks and Internets” (2d ed. 1999) and “Internetworking Technologies Handbook” (Cisco Systems, Inc., 1999).
Large networks may include thousands of end stations and network devices. In these networks, management of the network devices, their logical interconnections, and their internal configurations becomes a problem. As one response, the network management station (NMS) has been developed. Typically a network management station is a workstation or personal computer that runs a software program known as a network management system. The network management system generally enables an administrator to view a current configuration of a network, re-configure devices or their attributes, monitor network performance, etc.
Network management systems often are used with network devices that communicate using Simple Network Management Protocol (SNMP) and that store configuration information in one or more Management Information Base (MIB) variable values in the network device. The SNMP MIB information stored in a network device enables a NMS to determine the nature of the device more easily. However, not all network devices or switch devices, support SNMP.
For proper operation and to report accurate data, an NMS normally determines the logical and physical configuration of a managed network periodically, e.g., when the NMS initializes or starts operation. One aspect of determining the configuration of a network involves determining what devices are in the network, a process known as “device discovery.” Some network management stations can carry out automatic discovery of certain network devices, such as routers and workstations, and thereby provide the ability to manage the discovered devices remotely from the network management system console. An example of a NMS with this capability is a workstation that runs Cisco Resource Manager Essentials, commercially available from Cisco Systems, Inc., San Jose, Calif.
Several approaches may be used to carry out automatic discovery of various network devices. For example, routing table lookup can be used to discover routers, sequential ICMP Echo or Address Resolution Protocol (ARP) table lookup may be used to discover IP hosts, and Service Advertisement Protocol (SAP) table lookup can be used to discover IPX nodes. ARP is described in Request For Comments (RFC) 826, “Address Resolution Protocol.”
However, there is no standard approach to discover network switch devices (“switches” or “bridges”). Generally, switches are computer-based systems that connect and permit communication between two or more different network media, e.g., Ethernet, Token Ring, FDDI, etc. An example of a switch is the Catalyst 5000, commercially available from Cisco Systems, Inc. Some NMS products, such as Novell Managewise, cannot discover switches. Other network management systems require the network administrator to provide, manually, seed information pertaining to the switches, e.g., an IP address of the switch. An example of an NMS that requires such seed information is Optivity, from Nortel Networks.
Other systems carry out automatic discovery by sending ICMP Echo requests to all the IP addresses in the IP subnet range, or selectively to some active nodes by looking at the ARP cache on the hosts or on the routers. This approach is roughly equivalent to “pinging” all possible IP addresses in the IP subnet range. If a device responds, and it supports SNMP, then the NMS may also issue a query to the device for the value of any SNMP MIB variables that are specific to switches, such as the “sysType” MIB variable that is defined in MIB-II. For example, the NMS may issue such a query to all the hosts that have responded to the ICMP Echo requests, to determine if any such responding host is a switch.
Unfortunately, such approaches essentially involve blind or brute force searching, and introduce significant ICMP message traffic and SNMP message traffic into the network. For example, there may be hundreds of messages required to identify one switch, e.g., if the subnet range is 255 IP addresses, then 255 packets may be required. Another disadvantage is that the discovery time is too long, on the order of several hours in a complex network. Further, if a switch does not support SNMP, it is not discovered using these approaches because it cannot respond to SNMP requests. And even if a device does support SNMP, different device vendors implement different MIBs, so there is no assurance that a switch can be identified.
Many computer-based network routing devices (“routers”) support a form of remote network monitoring that is implemented in a Remote Network Monitoring (RMON) agent in the router. The RMON agent is a software element that is executed by or in the router normally used for traffic analysis or packet capturing. The RMON agent can monitor network traffic on network segments that directly connect interfaces of the routers. The RMON agent can examine all packets that travel across a physical segment that is being monitored. Based on the contents of the packets, the RMON agent can create network traffic statistic values and store them in the form of SNMP RMON MIB values that are used by NMS stations for network traffic analysis. Presently RMON has no applicability to device discovery.
The Spanning Tree Algorithm, as specified in IEEE Specification 802.1D, is implemented in almost all network switches. The Spanning Tree Algorithm defines a protocol that is used for communications between switches for the purpose of eliminating packet looping in a switched network.
FIG. 1
is a simplified block diagram of a network that illustrates use of Spanning Tree. In
FIG. 1
, four (4) switches
100
A,
100
B,
100
C,
100
D are interconnected by internetwork links
102
AB,
102
AC,
102
BD,
102
CD. Switch
100
D is further coupled to router
104
, which routes data traffic to a local area network (LAN)
106
. Each switch
100
A,
100
B,
100
C,
100
D is associated with a separate local area network
105
A,
105
B,
105
C,
105
D to which the switch directs or bridges traffic from LAN
106
.
In this configuration, a logical loop exists among switches
100
A,
100
B,
100
C,
100
D, and such a loop can cause a network to crash as packets travel endlessly around the loop. The Spanning Tree algorithm will prevent such crashes by logically severing one of the links
102
AB,
102
AC,
102
BD,
102
CD and thereby breaking the loop. Using Spanning Tree, the switches are logically organized in a multiway tree, in which one switch is a root node of the tree, and all other switches are leaf nodes of the tree. In this logical arrangement, loops are eliminated.
Switches operating under Spanning Tree will transmit Bridge Protocol Data Units (“BPDUs”) to communicate with other switches. The BPDUs are transmitted periodically, e.g., every thirty (30) seconds. The BPDUs allow each switch to calculate a spanning tree by communicating information about what switches are in the network.
When a switch initializes, it assumes that itself is the root node of a spanning tree, and periodically transmits BPDUs on each of its

Affiliated with

Also associated with

Rating

Automatic discovery of switch devices in a network does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Automatic discovery of switch devices in a network, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic discovery of switch devices in a network will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3199739