Electrical computers and digital processing systems: multicomput – Distributed data processing
Reexamination Certificate
1996-06-11
2001-01-16
Rinehart, Mark H. (Department: 2756)
Electrical computers and digital processing systems: multicomput
Distributed data processing
C709S203000, C709S219000, C709S229000
Reexamination Certificate
active
06175854
ABSTRACT:
FIELD OF THE INVENTION
The present invention is directed to a computer system architecture and method for processing computer applications accessed by multiple users and, more particularly, to a computer system architecture and method for processing multiple-user, real-time applications using dedicated compute servers.
There exist various computer system architectures that provide multiple user access to computer applications. FIGS.
1
-
5
illustrate such known architectures and will be described in detail. The advantages and disadvantages of these known architectures will be described with reference to their running real-time computer applications and, more particularly, game applications. Of course, other real-time applications can be processed.
FIG. 1
illustrates a monolithic server, local access computer architecture
10
. This architecture includes one monolithic compute server
12
and a plurality of user stations
14
locally connected to the compute server
12
. Each user station
14
may be a terminal or a computer emulating a terminal. It may even be a simple game input device, for example, a joystick with no display at all. An example of a dedicated game device is the PLAYSTATION available from the Sony Corporation. The sole compute server
12
performs all of the processing needed to run the application. The user stations
14
may all display the same scene or each may display a scene tailored to the particular local user. A popular application of this system is for two users to connect their game input devices into a shared personal computer (PC) or dedicated game device.
Several advantages are associated with this architecture. Dedicated game devices are much cheaper than PCs so this is an inexpensive system to use for playing games. Because many users play games on such a system, there is a large supply and selection of these devices and of compatible games. Since each user station
14
has a dedicated connection, access line latency is both low and fixed. Latency in a communications link measures the time from when a data bit enters the link at the data source to when it leaves the link at the data sink. It is often measured in milliseconds. If the compute server is a dedicated game device, there are no problems with competing applications slowing down the speed of play. Since all processing is performed on one machine, event resolution is not an issue. If the users are all in the same room, they may be able to talk with each other.
There are several disadvantages, however, associated with the monolithic server, local access computer architecture. While users are located close to each other, they may still be far enough apart that they cannot talk to one another during play. In addition, the number of users is limited, often to two with dedicated game devices. If the compute server is not dedicated to one application, applications may compete for processing cycles, thereby affecting each other's performance. In addition, since one compute server
12
does everything, its performance sets limits on the kinds of applications that can be supported. In addition, monolithic compute servers are not flexible in size or through time, and they cannot adapt readily to accommodate changing demand. Thus, once purchased, they may become quickly obsolete.
FIG. 2
illustrates a modem-to-modem architecture. Two user stations
22
are connected by way of modems
24
. Each user station
22
has everything needed to support a single user in a stand-alone application in addition to the ability to drive the modem
24
. Usually the user station
22
is a personal computer or dedicated game device. Each user station
22
performs all local processing display functions. Usually one user station
22
(the “master”) is chosen to resolve event contentions. To accomplish this the master receives inputs from both the local user and from the remote user, resolves contention between these inputs and decides the temporal progression of the application, and then communicates the status of that progression. The master may tell each station only what it needs to know or may send the complete status to both stations and let each decide how much of that status is relevant to its local display. Alternatively, the event contention resolution function may be shared between the stations.
There are advantages to the modem-to-modem architecture. Although personal computers are expensive compared with dedicated game servers, their purchase is justified by the fact that they have many uses beyond the processing of game applications. Many homes already have a personal computer and a modem so there is little incremental cost involved beyond the application itself and the optional game input device. Also, having the enormous processing power of a personal computer dedicated to one user allows for graphics that are usually beyond the capability of a monolithic server that has to run everything for everybody. In addition, the bandwidth and latency of the dial-in line are fixed and unwavering. Bandwidth is a measure of the rate at which data can be sent over a communications link. It is usually measured in bits per second. In general, it is desirable to have high bandwidth and low latency. In the modem-to-modem architecture, however, bandwidth is fixed at a fairly low value compared with the local access architecture discussed with reference to FIG.
1
and latency is fixed at a fairly high value, again compared with the local access architecture. Because bandwidth and latency vary independently of each other, the applications developers can often effectively mask the fact that the bandwidth is low and latency is high.
In addition, each user controls his own station environment so there should never be contention from competing applications. Also, two people from anywhere in the world can play together and, once the multimode modem standards are complete, can talk with each other while they play. A multimode modem refers to a modem that can transmit voice at the same time and over the same link that it transmits data. If multimode modems are used in the modem-to-modem method, then the application communicates over the data channel while at the same time, the voice channel lets the users talk to each other.
There are, however, disadvantages associated with the conventional modem-to-modem architecture. The number of users is limited to two. The PCs and communications equipment are expensive when compared with dedicated game devices. If event contention resolution is not shared, the master must spend some of its processing power performing this function, possibly to the detriment of the local user. On the other hand, if event contention resolution is shared, latency increases.
FIG. 3
illustrates a local area network (LAN) architecture. Several user stations
32
are connected by LAN
34
. The number of users is limited by the specific application, usually from four to eight users. The user stations
32
are similar to those described in the modem-to-modem architecture shown in
FIG. 2
with the difference that LAN support replaces modem support. The split between local processing and remote processing seen in the modem-to-modem architecture also applies here. Usually, event resolution is accomplished the same way, however, now the master is resolving four to eight users instead of only two. Alternatively, this function can be handled by a separate server on the LAN or shared among user stations.
The LAN architecture has most of the advantages associated with the modem-to-modem architecture. In addition, the LAN architecture allows more than two users to run a single application. As long as the LAN is not overloaded with traffic from other applications, it offers the high bandwidth and low latency that applications developers desire. Also, the LAN's broadcast mode can be used to reduce overall traffic requirements. The processing burden placed upon a user station's central processing unit by the access line is reduced because LAN cards have their own communications processors.
There are, however,
Ameritech Services, Inc.
Brinks Hofer Gilson & Lione
Rinehart Mark H.
LandOfFree
Computer system architecture and method for multi-user,... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Computer system architecture and method for multi-user,..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Computer system architecture and method for multi-user,... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2508449