Electrical computers and digital processing systems: multicomput – Computer-to-computer data routing – Alternate path routing
Reexamination Certificate
1999-05-26
2004-04-06
Coulter, Kenneth R. (Department: 2142)
Electrical computers and digital processing systems: multicomput
Computer-to-computer data routing
Alternate path routing
C709S241000, C709S241000, C709S235000, C709S241000, C709S243000, C345S215000, C370S360000, C370S400000
Reexamination Certificate
active
06718393
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to the field of data transfer. More particularly, the present invention relates to a system and method that dynamically distributes data traffic load among multiple channels to improve overall performance of the system.
2. Description of Art Related to the Invention
With the continual emergence of smaller, faster and more powerful computers, many businesses have implemented or are in the process of implementing distributed networks, such as local area networks (“LANs”). A “LAN” is a high-speed communications network which electrically couples a number of nodes together. A “node” is defined as a computer operating as a personal computer, workstation or server or any other electronic network connective device. As a result, distributed networks allow each user control over his or her own computer as well as access to other nodes coupled to the network. One well-known type of LAN is commonly referred to as an “Ethernet Network”. Switches are also used in Ethernet Networks resulting in a network called a “Switched Ethernet Network”.
As shown in
FIG. 1
, a Switched Ethernet Network
100
is a non-priority based network which typically includes a plurality of nodes
110
which are electrically coupled to at least one server
120
normally through a plurality of client communication channels
130
, a plurality of server communication channels
140
and a switching mechanism
150
having a number of ports
151
and
152
dedicated for each of the client and server communication channels
130
and
140
, respectively. Server
120
is a node, loaded with software such as network application software and the like, that supports multiple server communications channels
140
. Client and server communication channels
130
and
140
propagate information in parallel so that the nodes
110
and the server
120
can communicate.
Referring to
FIG. 2
, in a network environment, server
120
is implemented with network operating system (“NOS”) software
125
such as, for example, NETWARE™ by Novell Corporation of Provo, Utah, running on a host processor. Of course, other NOS software may be implemented within server
120
. Upon receiving requests for data from one of the plurality of nodes, NOS software
125
, in operation, provides data packets to adaptive driver software
126
for transmission through one of server communication channels
140
in accordance with an IEEE 802.3 standard (ANSI/IEEE std. 802.3, 1982 first edition). Each of server communication channels
140
supports a limited bandwidth currently ranging from 10 million bits per second (“Mbps”) up to approximately 100 Mbps for FastEthernet physical medium.
Referring to both
FIGS. 1 and 2
, in order to increase the bandwidth available for data transmission, server communication channels
140
from the. server
120
may include multiple physical adapters (e.g., Network Interface Cards “NICs”
141
), each assigned a unique identifier that functions as its address, which is typically referred to as a media access control (“MAC”) address. These NICs
141
are controlled by adaptive driver software
126
which communicates with both NOS software
125
and NICs
141
. Adaptive driver software
126
“virtualizes” these server communication channels
140
by presenting them to NOS software
125
as a single server communication channel having the collective bandwidth of the combined server communication channels
140
. For example, the Switched Ethernet Network
100
would have a “virtual” server communication channel equivalent to the number of server communication channels times the bandwidth of a single server communication channel “BW
scch
”.
In operation, when a data packet is provided to adaptive driver software
126
for transmission, adaptive driver software
126
, based on substantially predetermined parameters, select one of server communication channels
140
for use. By choosing different server communication channels
140
for different data packets, adaptive driver software
126
achieves “load balancing”, a scheme where adaptive driver software
126
operates so as to distribute the data traffic load generally equally among the server communication channels
140
.
Currently, Switched Ethernet Networks have been configured to operate in accordance with two types of conventional load balancing techniques. A first type of conventional load balancing technique involves adaptive driver software ascertaining a destination address of the data packet before it is transferred through one of the server communication channels. If the destination address is associated with a node that has not been assigned a server communication channel, adaptive driver software assigns that node to utilize a specific server communication channel. Thereafter, the adaptive driver software sequentially assigns a different channel to the next unassigned node until all nodes coupled to the network have been assigned substantially equally among the server communication channels. However, this technique is highly static in nature because channel assignment for the node is maintained throughout the duration of the communication session between the newly assigned node and the server. Moreover, although conventional network may achieve almost perfect load balancing if the data traffic load is evenly distributed, such situations are highly unusual in normal operation.
Another “load balancing” technique currently utilized by adaptive driver software is to assign one data packet at a time to different server communication channels in which the channel assignment is extinguished after every data packet transmission being one type of “purely dynamic” loading. Although this technique reduces disparate data traffic loads over server communication channels, it may result in a number of disadvantages. A first disadvantage is that for a “burst” data transfer (e.g., sending successive packets to a single node), there is a possibility that the data may become out of order at its destination. The reason is that the switching mechanism
150
may include first-in, first-out registers (“FIFOs”) to receive data transmitted through ports
152
. Over time, these FIFOs will likely be storing different amounts of data which would affect the order of data received by the destination node. A second disadvantage is that the switching mechanism would likely experience a data overflow because data is being transmitted to the destination node at a rate much greater than that supported by its client communication channels
130
coupling the node to switching mechanism
150
. Both of these disadvantages result in performance degradation.
Therefore, it would be advantageous to provide a communication system and dynamic data traffic load distribution scheme which mitigates the above-identified disadvantages.
SUMMARY OF THE INVENTION
Briefly, in accordance with one embodiment, the communication system comprises a plurality of nodes in which at least one node is implemented with adaptive driver software to dynamically control the distribution of data traffic among multiple communication channels. Of these communication channels, when under dynamic control, only one channel is capable of receiving data requests and at least two of the channels are capable of transferring data packet(s) to the data requesting node.
REFERENCES:
patent: 5175733 (1992-12-01), Nugent
patent: 5339392 (1994-08-01), Risberg et al.
patent: 5345229 (1994-09-01), Olnowich et al.
patent: 5367523 (1994-11-01), Chang et al.
patent: 5374952 (1994-12-01), Flohr
patent: 5461611 (1995-10-01), Drake et al.
patent: 5511169 (1996-04-01), Suda
patent: 5513322 (1996-04-01), Hou
patent: 5533198 (1996-07-01), Thorson
patent: 5534911 (1996-07-01), Levitan
patent: 5539621 (1996-07-01), Kikinis
patent: 5548532 (1996-08-01), Menand et al.
patent: 5555441 (1996-09-01), Haddad
patent: 5557316 (1996-09-01), Hoarty et al.
patent: 5590328 (1996-12-01), Seno et al.
patent: 5596722 (1997-01-01), Rahnema
patent: 5675738 (1997-10-01), Suzuki et al.
patent: 56921
Blakely , Sokoloff, Taylor & Zafman LLP
Coulter Kenneth R.
Intel Corporation
Nguyen Hai V.
LandOfFree
System and method for dynamic distribution of data traffic... 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 and method for dynamic distribution of data traffic..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for dynamic distribution of data traffic... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3273873