Electrical computers and digital processing systems: multicomput – Computer-to-computer data routing – Least weight routing
Reexamination Certificate
1998-07-22
2001-03-20
Kim, Kenneth S. (Department: 2183)
Electrical computers and digital processing systems: multicomput
Computer-to-computer data routing
Least weight routing
C709S223000, C709S241000
Reexamination Certificate
active
06205465
ABSTRACT:
A microfiche appendix containing computer program listings Appendix 1 and Appendix 2 is included with this application. The microfiche appendix has a total of 22 frames on a total of one sheet of microfiche.
FIELD OF THE INVENTION
The present invention generally relates to data processing. The invention relates more specifically to a mechanism that enables and manages parallel execution of computing processes.
BACKGROUND OF THE INVENTION
Computer networks have become ubiquitous in the home, office, and industrial environment. As computer networks have grown ever more complex, automated mechanisms for organizing and managing the networks have emerged. These mechanisms are generally implemented in the form of one or more computer programs, and are generically known as network management systems or applications.
FIG. 1
is a simplified diagram of a network
100
that is managed by a network management station
10
. The network
100
comprises one or more network devices
102
, such as switches, routers, bridges, gateways, and other devices. Each network device
102
is coupled to another network device
102
, or to one or more end stations
120
. Each end station
120
is a terminal node of the network
100
at which some type of work is carried out. For example, an end station
120
is a workstation, a printer, a server, or similar device.
Each network device
102
executes a network-oriented operating system
110
. An example of a network-oriented operating system is the Internetworking Operating System (IOS) commercially available from Cisco Systems, Inc. Each network device
102
also executes one or more applications
112
under control of the operating system
102
. The operating system
102
supervises operation of the applications
112
and communicates over network connections
104
using an agreed-upon network communication protocol, such as Simplified Network Management Protocol (SNMP).
Each device
102
stores information about its current configuration, and other information, in a Management Information Base (MIB)
114
. Information in the MIB
114
is organized in one or more MIB variables. The network management station
10
can send fetch and set commands to the device
102
in order to retrieve or set values of MIB variables. Examples of MIB variables include sysObjectID.
Preferably the network management station
10
is a general-purpose computer system of the type shown and described further herein in connection with FIG.
10
. The network management station
10
executes one or more software components that carry out the functions shown in block diagram form in FIG.
1
. For example, the network management station
10
executes a basic input/output system (BIOS)
20
that controls and governs interaction of upper logical layers of the software components with hardware of the network management station. An example of a suitable BIOS is the Phoenix ROM BIOS. The network management station
10
also executes an operating system
30
that supervises and controls operation of upper-level application programs. An example of a suitable operating system is the Microsoft Windows NT® operating system.
The network management station
10
executes an asynchronous network interface
50
or ANI under control of the operating system
30
. The ANI
50
provides an interface to the network
100
and communicates with the network using SNMP or another agreed-upon protocol. The ANI
50
provides numerous low-level services and functions for use by higher-level applications.
The network management station
10
executes a network management system
40
that interacts with a database
60
containing information about the managed network
100
. The network management system
40
is an example of a network management application. Using a network management application, a manager can monitor and control network components. For example, a network management application enables a manager to interrogate devices such as host computers, routers, switches, and bridges to determine their status, and to obtain statistics about the networks to which they attach. The network management application also enables a manager to control such devices by changing routes and configuring network interfaces. Examples network management applications are CiscoWorks, CiscoWorks for Switched Internetworks (CWSI), and CiscoView, each of which is commercially available from Cisco Systems, Inc.
Contemporary information processing involves execution by a processor or computer of a computer program, process, or routine, all of which are called “processes” in this document. In many contexts, execution of a process by a processor may be delayed when the process is required to wait for an external process or device to carry out some other task. When no such delays occur, it is has been recognized that significant processing time is saved, and the use of processing resources is maximized, by executing several processes concurrently or in parallel.
For example, consider a network management system that is used to monitor and manage the operation of a computer network that comprises numerous network devices. The network devices comprise switches, routers, and other devices that connect to the external world and are also called “managed devices”. In such systems, the network management application program often must wait for a managed device carry out another task and to respond to the network management application. It is desirable to configure the system so that it communicates with several devices during one general time period, so that the network management system communicates with a second device while it is waiting for the first device to become available.
Prior parallel processing approaches do not address significant problems that arise in the network environment. For example, problems arise when parallel processing is applied in a network environment that uses a shared database of network device information. In particular, it is possible that one process could change the shared database at the same time that a second process is attempting to change the shared database.
Further, in network management systems, there is a continuing need to modify the network management system and its associated database to accommodate new network devices and new services for the network devices. Often, the network management system is modified by installation of patches, upgrades, and other modifications at the customer's site. This is called field modification or field extensibility. A parallel processing mechanism, in this context, must be able to adapt to new processes and apply parallel processing to new data set definitions that are installed in the field.
Generally, the use of a shared data model, and field extensibility, create three major problems.
First, determining the order of execution of components of a process is a problem. Because the components of the system are independently developed but share access to a common data model, the problem of when code is executed arises in ways it does not arise in monolithically developed code. In particular, in monolithic code, the execution order of components is explicitly expressed in static sequencing (particular components are explicitly stated and invoked at the right time as determined by the coder). In our case, because code for various execution threads is provided by independent developments, execution order can not be explicitly stated. Thus a mechanism for execution order determination had to be added.
Second, providing for parallel execution of the independently provided components is a problem. Usually, components are executed in parallel by the developer carefully constructing appropriate synchronization and thread initiation mechanisms for custom tailored code. In our case, the code is not custom tailored for parallel execution since even the order of execution is not known when the code is written. Thus, we need a mechanism which minimally effects the writing of the code but which permits for the code to operate correctly at the right time.
Third, providing
Schleimer Stephen I.
Schoening Charles B.
Smith, Jr. Richard J.
Cisco Technology Inc.
Hickman Palemo Truong & Becker LLP
Kim Kenneth S.
Palermo Christopher J.
LandOfFree
Component extensible parallel execution of multiple threads... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Component extensible parallel execution of multiple threads..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Component extensible parallel execution of multiple threads... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2536749