Multiplex communications – Channel assignment techniques – Using a separate control line or bus for access control
Reexamination Certificate
1997-11-06
2001-01-23
Nguyen, Chau (Department: 2739)
Multiplex communications
Channel assignment techniques
Using a separate control line or bus for access control
C370S445000
Reexamination Certificate
active
06178176
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates generally to network computing. In particular the present invention relates to methods and apparatuses for synchronizing a stack bus on a Fast Ethernet repeater stack.
The growth of local-area networks (LANs) has been driven by the introduction of Ethernet technology as well as the availability of powerful, affordable personal computers and workstations. As a result, applications that once were possible only on mainframe computers are now running on LANs. Network speed and availability are critical requirements. However, existing applications and a new generation of multimedia, groupware, imaging, and database products can tax a network running at Ethernet's traditional speed of 10 megabits per second (Mbps). Moreover, with more applications requiring faster LAN speeds for acceptable performance, network managers increasingly find that high-performance computation platforms and mission-critical applications can overwhelm a 10 Mbps network. Network managers therefore are increasingly are implementing high-speed LAN technology.
Fast Ethernet
For organizations with existing Ethernet installations, increasing the network speed to 100 Mbps is preferable to investing in a completely new LAN technology. This user preference has driven the industry's decision to specify a higher-speed Ethernet that operates at 100 Mbps. This higher-speed Ethernet is known as Fast Ethernet.
In July 1993, a group of networking companies joined to form the Fast Ethernet Alliance. The charter of the group was to draft the 802.3u 100BaseT specification of the Institute of Electrical and Electronics Engineers (IEEE) and to accelerate market acceptance of Fast Ethernet technology. The final IEEE 802.3 specification was approved in June 1995. Among the other goals of the Fast Ethernet Alliance are: To maintain the Ethernet transmission protocol Carrier Sense Multiple Access Collision Detection (CSMA/CD); to support popular cabling schemes; and to ensure that Fast Ethernet technology will not require changes to the upper-layer protocols and software that run on LAN workstations. For example, no changes are necessary to Simple Network Management Protocol (SNMP) management software or Management Information Bases (MIBs) in order to implement Fast Ethernet.
Other high-speed technologies, such as 100VG-AnyLAN and Asynchronous Transfer Mode (ATM), achieve data rates in excess of 100 Mbps by implementing different protocols that require translation when data moves to and from 10BaseT. Protocol translation requires changing the frame, which often causes delays in frame transmission through layer
2
(data-link layer) LAN switches. Data can move between Ethernet and Fast Ethernet, on the other hand, without requiring protocol translation or software changes, because Fast Ethernet maintains the 10BaseT error control functions as well as the frame format and length.
In many cases, organizations can upgrade to 100BaseT technology without replacing existing wiring. Options for 100BaseT media are the same as those for 10BaseT. They include shielded and unshielded twisted pair (STP and UTP) and fiber. the Media Independent Interface (MII) provides a single interface that can support external transceivers for any of the 100BaseT physical sublayers.
CSMA/CD
Carrier sense-collision detection is widely used in LANs. Many vendors use this technique with Ethernet and the IEEE 802.3 specification. A carrier sense LAN considers all stations as peers; the stations contend for the use of the channel on an equal basis. Before transmitting, the stations monitor the channel to determine if the channel is active (that is, if another station is sending data on the channel). If the channel is idle, any station with data to transmit can send its traffic onto the channel. If the channel is occupied, the stations must defer to the station using the channel.
FIG. 1
depicts a carrier sense-collision detection LAN. Network devices
102
,
104
,
106
, and
108
are attached to a network bus
110
. Only one network device at a time is allowed to broadcast over the bus, since if more than one device were to broadcast at the same time, the combination of signals on the bus would likely not be intelligible. For example, assume network devices
102
and
104
want to transmit traffic. Network device
108
, however, is currently using the channel, so network devices
102
and
104
must “listen” and defer to the signal from network device
108
, which is occupying the bus. When the bus goes idle, network devices
102
and
104
can then attempt to acquire the bus to broadcast their messages.
Because network device
102
's transmission requires time to propagate to other network devices, these other network devices might be unaware that network device
102
's signal is on the channel. In this situation, network device
102
or
104
could transmit its traffic even if network device
108
had already seized the channel after detecting that the channel was idle. This problem is called the collision window. The collision window is a factor of the propagation delay of the signal and the distance between two competing stations. Propagation delay is the delay that occurs before a network device can detect that another network device is transmitting.
Each network device is capable of transmitting and listening to the channel simultaneously. When two network device signals collide, they create voltage irregularities on the channel, which are sensed by the colliding network devices. The network devices then turn off their transmission and, through an individually randomized wait period, attempt to seize the channel again. Randomized waiting decreases the chances of another collision because it is unlikely that the competing network devices generate the same wait time.
It is important that the total propagation delay not exceed the amount of time that is required to send the smallest size data frame. This allows devices to discard data corrupted by collisions by simply discarding all partial frames. It is therefore not desirable for entire frames of data to be sent before a collision is detected. Carrier sense networks are usually implemented on short-distance LANs because the collision window lengthens as the channel gets longer. Longer channels provide opportunity for the more collisions and can reduce through-put in the network. Generally, a long propagation delay coupled with short frames and high data transfer rates give rise to a greater incidence of collisions. Longer frames can mitigate the effect of long delay, but they reduce the opportunity for competing stations to acquire the channel.
The IEEE 802.3 specification sets a standard minimum frame size of 64 bytes (512 bits). Therefore, it order for a network to comply with the standard, a station on the network must not be able to transmit 64 bytes of data before a collision is detected.
Although Fast Ethernet maintains CSMA/CD, the Ethernet transmission protocol, it reduces the transmission time for each bit by a factor of 10. Thus, the Fast Ethernet signal speed increases tenfold, from 10 Mbps to 100 Mbps. Therefore, the propagation delay for each part of the network, also referred to as the part's “latency,” must be reduced if the 64 byte 802.3 specification collision detection standard is to be satisfied. Latency is typically expressed in terms of bit time, or the amount of data that could be transmitted on the network during the period which it takes a signal to propagate through a network device.
Repeaters
While some Ethernet applications connect numerous network devices to a network bus that is literally a cable connecting the network devices, it is often more desirable to connect network devices using a repeater or hub. It should be noted that in the following description the term “hub” and the term “repeater” are used interchangeably. The repeater manages collision detection for the network devices so that the network devices need only broadcast messages without detecting collisions. The repeater notifies a net
Cavaro Mark D.
Voloshin Moshe
Cisco Technology, Inc
Lee Chiho Andrew
Nguyen Chau
Ritter, Van Pelt & Yi LLP
LandOfFree
Synchronous stack bus for fast Ethernet repeater does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Synchronous stack bus for fast Ethernet repeater, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Synchronous stack bus for fast Ethernet repeater will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2489304