This application claims benefit to provisional application 60/020,106 filed Jun. 20, 1996.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to the field of data communications. In particular, the present invention provides a system for establishing a master-slave relationship, at the physical layer, on a point-to-point link in a peer-to-peer network.

2. Background

Peer-to-Peer networks utilize a communication protocol such as Carrier Sense Multiple Access/Collision Detection (CSMA/CD). The philosophy of this type of network is that every device coupled to the network determines, on an independent basis, when to send information to another device coupled to the network. To summarize this process, when a device has data for transmission to another device on the network, the device checks the medium across which it intends to transmit the data for any existing traffic. If the device senses traffic on the medium, it waits until the medium is free of traffic before initiating transmission of its data. If the device does not sense traffic on the medium, then it attempts to send data across the network to the other device. If the sending device detects a collision during the transmission, it suspends transmission and waits a randomly chosen amount of time before attempting to resend the data by repeating the above process.

As the utilization of a peer-to-peer network increases, the number of collisions on the network increases, thereby reducing the overall throughput of the network. To address this problem, technology has developed various types of hardware capable of higher data transmission rates and has developed more efficient network software. Higher data transmission rates reduce the time period during which the data is on the medium for a given amount of data, thereby reducing the probability of a collision with data from another device.

Network software typically includes several layers of protocol. For example, in an Ethernet local area network, the network interface comprises: the physical layer, the media access control sublayer and the logical link control sublayer. The physical layer's function is to take data packets given to it from the media access control layer and place the packets on the medium to be delivered to a destination device. The carrier sense, collision detection functions to support the arbitration of access to the medium by the attached device.

Each device in a peer-to-peer network is a peer to other devices in the network. Thus, no device is considered to be “superior” to any other device on the network. As such, a master-slave relationship between devices is not typically supported in a peer-to-peer network environment. It is therefore desirable to provide a system capable of establishing a master-slave relationship (when such need arises), at the physical layer, on a point-to-point link in a peer-to-peer network.

SUMMARY OF THE INVENTION

The present invention provides improved performance in a peer-to-peer network by permitting the establishment of a master-slave relationship between any two devices at each end of a point-to-point link. By establishing a master-slave relationship, the master may provide control signals or other data to the slave for use by the slave.

An embodiment of the present invention provides a system for establishing a master-slave relationship at the physical layer between a first device and a second device in a peer-to-peer network connected by a point-to-point link. The system generates a first information which is associated with the first device. Additionally, a second information is generated which is associated with the second device. The first information is compared with the second information to generate a comparison result. Based on the comparison result, one device is designated as a master and the other device is designated as a slave.

Another feature of the invention provides that the generation of the first information and the second information is performed by the first device and second device, respectively.

A specific aspect of the invention transmits the first information from the first device to the second device and transmits the second information from the second device to the first device. Each device then compares the first information to the second information to determine which device is master and which device is slave. Since each device has the same data and the same criteria, their determination is expected to agree.

An embodiment of the invention assigns a particular value to each device. If the assigned values are unequal, then the assigned values are used to determine which device is master and which device is slave. If the assigned values are equal, then the first device selects a first random number and the second device selects a second random number. The first selected number is compared with the second selected number to generate a comparison result. Based on the comparison result, one device is designated as master and the other device is designated as slave.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention is illustrated by way of example in the following drawings in which like references indicate similar elements. The following drawings disclose various embodiments of the present invention for purposes of illustration only and are not intended to limit the scope of the invention.

FIG. 1

illustrates an exemplary network structure upon which the present invention can be implemented.

FIG. 2

illustrates a comparison of protocol layers between the ISO standard and the IEEE 802.3 standard.

FIG. 3

is a flowchart illustrating a method for designating one device as a master and one device as a slave.

FIG. 4

is a flowchart which more specifically illustrating a method for overriding a device's assigned designation as master or slave.

FIG. 5

illustrates a pair of devices coupled to provide a clock loopback function.

DETAILED DESCRIPTION

The following detailed description sets forth numerous specific details to provide a thorough understanding of the invention. However, those skilled in the art will appreciate that the invention may be practiced without these specific details. In other instances, well-known methods, procedures, components, and circuits have not been described in detail so as not to obscure the invention. Referring now to the drawings, in which like numerals represent like elements throughout the several figures, aspects of the present invention and the preferred operating environment will be described.

The present invention provides a system for establishing a master-slave relationship, at the physical layer, on a point-to-point link in a network. Thus, one device on the point-to-point link is designated a master and the other device is designated a slave. This master-slave relationship may be used to provide control signals, clock signals, or other data from the master to the slave.

In a particular example, the master-slave relationship may be used to provide a clock loopback function on a point-to-point link between the master device and the slave device. The clock loopback function allows higher data transmission rates, thereby reducing the probability of collisions. If one of the devices or ends on the link is designated as a master and the other as a slave at the physical layer, then the master can provide the clocking mechanism for both devices while the slave loops the clock signal back to the master. This configuration allows higher transmission rates on cabling that previously could not transmit data at these rates because of a high noise-to-signal ratio (due to the poor signaling characteristics of the wire). Additional details regarding clock loopback is provided below with reference to FIG.

5

.

FIG. 1

depicts an exemplary network in which an embodiment of the present invention may be implemented. The network includes several End Devices

115

,

120

,

125

,

130

, and

135

(also referred to End Devices A through E, respectively). The End Devices may be any device capable of being coupled to a network, such as computer workstations. The End Devices

115

-

135

are connected to Central Device Units

105

and

110

using point-to-point links

145

a

-

145

e.

Each Central Device Unit

105

,

100

may also be coupled to one or more additional Central Device Units, thereby interconnecting all of the devices coupled to the Central Device Units. Although not shown in

FIG. 1

, End Devices may also be coupled directly to one another.

An example of transmitting data on the network of

FIG. 1

is as follows: End Device A has data to transmit to End Device E. To accomplish this transmission, End Device A transmits data along link

145

a

to Central Device Unit

105

. Central Device Unit

105

then transmits the data across link

150

to Central Device Unit

110

. Central Device Unit

110

transmits the data along link

145

e

to the its destination at End Device E.

The present invention permits the establishment of a master-slave relationship between any two devices on a point-to-point link in a network. This includes the link between any of the Central Device Units and an End Device (links

145

a

-

145

e

), between two End Devices (link not shown) or between two Central Unit Devices (link

150

). A device in the context of this specification will be defined as either an End Device or a Central Device Unit that connects the end devices. Furthermore, although only five End Devices and two Central Device Units are shown, this is only for discussion purposes since any number of End Devices and Central Device Units may be incorporated into the network. In a particular embodiment of the invention, the End Devices and the Central Device Units are part of an Ethernet local area network (not shown).

FIG. 2

depicts a comparison of protocol layers between the International Standards Organization Open Systems Interconnect (ISO OSI) and Institute of Electrical and Electronic Engineers, Inc. (IEEE) 802.3 data communication standards as well as a breakdown of the physical layer in IEEE Std 802.3u—1995.

FIG. 2

depicts two of the ISO OSI layers, the physical layer

210

and the Data Link layer

205

. These are the two layers that perform the services required within a local area network.

FIG. 2

also illustrates the comparable IEEE 802.3 protocol layers, commonly referred to as Ethernet, comprising: the Logical Link Control sublayer

215

; the Media Access Control sublayer

220

; and the Physical layer

225

.

The Ethernet protocol layers are defined by a standard developed by IEEE and promulgated as IEEE 802.3. In IEEE 802.3u, the physical layer is divided into the following sublayers: the Physical Coding sublayer (PCS)

230

; the Physical Medium Attachment sublayer (PMA)

235

; the Physical Medium Dependent sublayer (PMD)

240

; and the Auto-Negotiation sublayer

245

. In one embodiment of the invention, the Auto-negotiation sublayer of IEEE 802.3 is utilized to perform the functions described below. This Auto-Negotiation sublayer provides a service within the physical layer for devices to detect the modes of operation supported by a device at the other end of a point-to-point link, determine common abilities and configure joint operation. Using the Auto-Negotiation sublayer, devices may establish services to each other that other devices on the network need not be aware of or even support. For additional information regarding IEEE 802.3 and the Auto-Negotiation sublayer, refer to “Media Access Control (MAC) Parameters, Physical Layer, Medium Attachment Units and Repeater for 100 Mb/s Operation,” IEEE Std 802.3u—1995, Clause 28.

A clocking loopback function may be provided at the physical layer to make digital signal processing (DSP) more efficient, thereby allowing higher data transmission rates on cables or wires not previously capable of handling these rates. By looping back the same clock signal, both the master device and the slave device use the same clock signal. This common clock improves the ability of the DSP circuits to remove noise from the signals and, therefore, improve the flow of data. Noise removal is especially important when transmitting signals at high speed across lower quality wires or cabling; i.e., wires or cabling having a higher noise-to-signal ratio.

In a specific example, the present invention may be used in networks having previously installed cables or wires that are otherwise unable to take advantage of higher transmission rates due to the poor, or low quality, signaling characteristics of the cabling or wire. Prior to the present invention, users or administrators of these networks wishing to improve transmission rates may be required to install new cables or wires at considerable expense. In addition to the cost of the new wire or cable itself, the cable may be required to be installed in existing walls or structures. By using the invention described herein, existing cables having a poor transmission quality may be used without requiring the installation of new, higher quality, wires or cables.

FIG. 5

depicts a clock loopback function where a Device

500

(“Device A”) is designated as the master and Device

505

(“Device B”) is designated as the slave. Device A generates, or receives from an external source, its timing signal (clock signal). Device A uses this clock signal to transmit data using a Transmitter

515

across link segment

530

to Device B. A Receiver

525

in Device B receives Device A's data on link segment

530

. Device B then loops back the clock signal received from Device A and uses Device B's Transmitter

520

to transmit it's data to Device A across link segment

535

. Device A's Receiver

510

receives data with the clock signal used by Device A's Transmitter

515

. Since Device A and Device B have established a master-slave relationship, both devices know which device will be using their own clock signal and which device will be looping back the other's clock signal.

Link segments

530

and

535

shown in

FIG. 5

illustrate two separate data flows in opposite directions. This arrangement may be accomplished by using a multi-conductor twisted-pair cable or a fiber optic pair. Those skilled in the art will appreciate that various other types of wires and cables may be used as link segments

530

and

535

in FIG.

5

.

FIG. 3

illustrates a procedure for designating one device as a master and one as a slave. The method of

FIG. 3

starts at step

305

and continues to step

310

where the first device is assigned as a master or a slave. The procedure then continues to step

315

where the second device is assigned as either a master or slave.

In one embodiment of the invention, the assignments in steps

310

and

315

are performed during each device's initialization to the network. For example, as End Device A (

FIG. 1

) begins an initialization sequence to add itself to the network, End Device A sets itself as a master or a slave based on its default designation. This default designation as master or slave may be based on a preassigned status set within the device. This default designation may be modified by the network administrator. When Central Device Unit

105

is initialized, it sets itself as a master or slave based on its default designation.

Following the assignment as master or slave to both of the devices, the procedure of

FIG. 3

continues at step

320

where the first device and the second device pass their assigned values to each other. In essence, each device transmits its default designation to the other device.

At step

325

, having passed assigned values, a comparison is performed to determine if there is a conflict (i.e., both devices assigned as slaves or both devices assigned as masters). This comparison may be performed by both devices simultaneously.

If the assignments are not the same at step

325

, the procedure stops at step

340

at which time the devices will continue with their network initialization or begin communicating in the standard network environment. Otherwise, the procedure continues to step

330

where one of the devices is selected to change its assignment. Additional details regarding the changing of assignments is provided below with respect to FIG.

4

. After selecting the device assignment to override at step

330

, step

335

continues by changing the assignment of the selected device such that it is different from the other device's assignment. The procedure of

FIG. 3

then stops at step

340

, and permits the devices to continue network initialization and begin regular network communication.

FIG. 4

illustrates in greater detail the selection and changing procedure steps (

330

and

335

) of FIG.

3

. The procedure of

FIG. 4

starts at

405

. At this point there has already been a determination that a conflict exists in the assignment of master and slave status to the two devices. To resolve the conflict, the procedure continues to step

410

where the first device randomly selects (or generates) a number or other information. At step

415

, the second device randomly selects (or generates) a number or other information. The first and second random numbers must be independent from one device to another to prevent the two devices from generating the same series of random numbers. Additionally, the random numbers are preferably selected such that it is equally likely that either device will be selected as master. These random number selections may be generated by any process available to the device.

A particular embodiment of the invention is described as generating random numbers to determine which device assignment will be overridden. However, those skilled in the art will appreciate that tokens or other information may be generated by each device and compared to determine which device assignment is overridden. Any token or information may be used, provided that the tokens or information may be distinguished from one another using a comparison mechanism.

The maximum value of the random number (e.g., number of bits) is a design choice recognizing that the larger the number the less probability that both numbers will be the same. Consequently, the larger the number, the more time and resources will be devoted to handling this random function as it will need to be transmitted to the other device. For example, one may not want to choose to design a random number that takes up four bytes of data because this data will need to be transferred to the other device and four bytes will take up more transmission time than a one byte random number. Similarly, one may not want to choose a random number that only takes up one bit because this increases the probability of both devices generate the same random number.

After both devices have selected a random number, the method continues to step

420

where both devices transmit their random number to the other device. At step

425

, both devices compare the two random numbers. If the random numbers are equal, then the method repeats steps

410

through

425

. If the random numbers are not equal, then, at step

430

, both devices determine which device had the higher number. If the first device has a higher number, it's designated as master at step

435

, regardless of how it was initially assigned. In this situation, the second device is designated as slave at step

440

, regardless of how it was initially assigned.

If, at step

430

, the first device was not higher than the second device, then the procedure continues to step

445

where the second device is designated as master and the first device is designated as slave. Again the designations in step

445

and

450

are performed regardless of how the devices were initially assigned.

One embodiment of the present invention includes using the Auto-Negotiation sublayer of the IEEE 802.3u standard. Referring to the network shown in

FIG. 1

, End Device A is added to the network so that it can communicate with the other network devices. Once the medium

145

a

is connected, End Device A begins to initialize itself onto the network. As part of the initialization process of IEEE 802.3u to establish the physical link, the Auto-Negotiation sublayer in the End Device A and Central Device Unit

105

will each transmit Fast Link Pulse (FLP) bursts on the link between the two devices. Within these FLP bursts include bits which allow devices to transmit operational modes supported by the devices. In this example, within a FLP burst, a device will set bits signifying whether it can support a master-slave function, whether its default designation is a master or slave and several bits devoted to a random number in case of a conflict arises where both devices' default designations are the same (i.e., both devices are designated as preferring master or preferring slave).

If a master-slave relationship is to be developed between End Device A and Central Device Unit

105

, both devices send one or more FLP bursts to the other, carrying the information designated above. Therefore, End Device A and Central Device Unit

105

each have a complete set of information from both devices including whether both devices can support the master-slave function, the other device's default designation, and each other's random number. This information is used to resolve any conflicts in the default designation. Therefore, if either device lacks the ability to support a master-slave relationship, then no such relationship is established. If, on the other hand, both End Device A and Central Device Unit

105

have the master-slave function bit set, signifying that each device supports the master-slave function, but the devices have different defaults (i.e., one is a master and the other is a slave) then each device uses its default designation.

However, if both devices support the master-slave relationship and both devices have the same default designation, then they compare their own random number with the random number sent from the other device in the FLP burst, and the device with the highest value becomes the master and the device with the lower value becomes the slave. If the random numbers are the same, another random number is generated by each device and sent to the other in another FLP burst. Another comparison is performed and the procedure continues until the random numbers are not equal and a master and a slave has been established.

From the above description and drawings, it will be understood by those skilled in the art that the particular embodiments shown and described are for purposes of illustration only and are not intended to limit the scope of the invention. Those skilled in the art will recognize that the invention may be embodied in other specific forms without departing from its spirit or essential characteristics. References to details of particular embodiments are not intended to limit the scope of the claims.