Electrical computers and digital data processing systems: input/ – Input/output data processing – Peripheral monitoring
Reexamination Certificate
2000-09-07
2002-04-09
Lee, Thomas (Department: 2182)
Electrical computers and digital data processing systems: input/
Input/output data processing
Peripheral monitoring
C709S203000, C709S219000, C711S003000, C714S100000
Reexamination Certificate
active
06370599
ABSTRACT:
BACKGROUND OF THE INVENTION
1. The Field of the Invention
The present invention relates generally to methods for increasing the efficiency, speed and/or throughput of a computer system. More specifically, the invention relates to methods for offloading computing tasks that are typically performed by a host processor in software, to a specific hardware component, thereby freeing up host processor resources and increasing the overall efficiency of the computer system.
2. The State of the Art
A functional computer system generally consists of three fundamental components. The first component is the host computer and its associated peripheral hardware components. The host computer typically includes a central processing unit (CPU), which is interconnected via a bus with, for instance, system memory such as RAM or ROM. A system will also include a number of peripheral hardware devices, depending on the functionality needed, such as magnetic or optical disk storage devices, a keyboard or other input device, a display or other output device and communication equipment, such as a modem and/or a network interface card (NIC). Another fundamental computer component is the application software. Such software includes the familiar word processing applications, spread sheet applications, database applications, communications and network applications and so forth.
The final component of a modem, functional computer system is an operating system. The computer operating system performs many functions such as allowing a user to initiate execution of an application program. In addition, modern operating systems also provide an interface between application software and the host computer and its peripheral hardware. Thus, while it was once commonplace for an application program to directly access computer system hardware, modem operating systems provide standardized, consistent interfaces that allow user applications to interface with or access computer hardware peripherals in a standardized manner. To provide a consistent interface, operating system architectures are increasingly designed so that there may be several software layers between the actual hardware peripheral and the application program. For example, an application may make a call into the operating system. The operating system, in turn, may utilize the services provided by a hardware device driver layer. The device driver layer would then interface directly with the specific hardware peripheral. A primary advantage of such a layered approach is that layers may be added or replaced without impacting the other layers.
As will be appreciated, the complexity and sophistication of such operating systems, application software, and networking and communications continues to increase. This of course results in more functional and useful computer systems. However, this increased functionality is not without a cost. More feature rich operating systems and software applications often result in an increase in the processor overhead as a result of the additional duties that must be performed by a processor/CPU when executing such system functions and/or applications, this phenomenon is especially apparent in connection with particular types of applications, such as network communication-type software applications. With the high bandwidth media that are increasingly prevalent, network speeds often match or exceed the CPU processor speed and memory bandwidth of the host computer. As such, to efficiently communicate over such networks, the CPU utilization and memory bandwidth used of the network-connected host computer must be minimized.
In addition, network applications further burden the host processor due to the layered architecture used by most, such as the seven-layer OSI model, or the layered model, or the layered model used by the Windows NT operating system. As is well known, such a model is used to describe the flow of data between the physical connection to the network and the end-user application. The most basic functions, such as putting data bits onto the network cable, are performed at the bottom layers, while functions attending to the details of applications are at the top layers. Essentially, the purpose of each layer is to provide services to the next higher layer, shielding the higher layer from the details of how services are actually implemented. The layers are abstracted in such a way that each layer believes it is communicating with the same layer on the other computer that is being communicated with via the network.
As will be appreciated, the various functions that are performed on a data packet as it proceeds between layers can be software intensive, and thus can demand a substantial amount of CPU processor and memory resources. For instance, in the Windows NT networking model, certain functions that are performed on the packet at various layers are extremely CPU intensive, such as packet checksum calculation and verification; encryption and decryption of data; message digest calculation and TCP segmentation. As each of these functions are performed, the resulting demands on the CPU/memory can greatly effect the throughput and performance of the overall computer system.
Although software applications and operating system functions are placing greater demands on computer system resources at the same time the capability, efficiency, and throughput of many computer hardware peripherals—such as network interface cards (NICs)—are also increasing. These computer system peripherals are often equipped with a dedicated processor and memory, and typically are capable of performing very sophisticated and complex computing tasks—tasks that are otherwise performed by the computer system processor in software. For instance, may NICs are capable of independently performing tasks otherwise performed by the CPU in software at an appropriate network layer, such as checksum calculation/verification; data encryption/decryption; message digest calculation; TCP segmentation; and others. As such, there is an advantage in offloading such CPU intensive tasks to a peripheral hardware device. This would reduce processor utilization and memory bandwidth usage in the host computer, and thereby increase the efficiency, speed and throughput of the overall system.
However, the processing capabilities of different peripheral devices vary widely. Thus, there needs to be an efficient method by which a computer system/operating system can identify the processing capabilities of such peripheral devices, and then assign and offload specific processing tasks to the device when needed. Also, it would be desirable if the tasks could be identified and assigned dynamically, depending on the then current needs of the processor. This would allow the computer system processor to take advantage of the capabilities of a hardware peripheral on an as-needed basis.
SUMMARY AND OBJECTS OF THE INVENTION
The foregoing problems in the prior state of the art have been successfully overcome by the present invention, which is directed to a system and method for offloading functions and tasks that were previously performed at a processor-software level, to an appropriate hardware peripheral connected to the computer system. The invention is particularly useful in connection with the offloading of tasks to network interface card (NIC) peripheral devices, which can often perform many of the tasks otherwise performed by the computer CPU in software.
In a preferred embodiment of the invention, a software implemented method and protocol is provided that allows, for instance, the operating system (OS) to “query” the device drivers (often referred to as “MAC” drivers) of any hardware peripherals (such as a NIC) that are connected to the computer system. The various device drivers each respond by identifying their respective hardware peripheral's processing capabilities, referred to herein as “task offload capabilities.” In the preferred embodiment, once the task offload capabilities of each particular peripheral have been identified, the OS can then enable selected peripherals to p
Anand Sanjay
Brandon Kyle E.
Hyder Jameel
Srinivas Nk
Elamin Abdelmoniem
Lee Thomas
Microsoft Corporation
Workman & Nydegger & Seeley
LandOfFree
System for ascertaining task off-load capabilities of a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System for ascertaining task off-load capabilities of a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System for ascertaining task off-load capabilities of a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2820139