Electrical computers and digital processing systems: multicomput – Computer-to-computer protocol implementing – Computer-to-computer data transfer regulating
Reexamination Certificate
1998-03-11
2001-10-09
Rinehart, Mark H. (Department: 2756)
Electrical computers and digital processing systems: multicomput
Computer-to-computer protocol implementing
Computer-to-computer data transfer regulating
C370S229000
Reexamination Certificate
active
06301620
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a method of transmitting data from a server computer and the server computer, more particularly a method of transmitting data from a server computer incorporated in a data transmission system and the server computer.
BACKGROUND OF THE INVENTION
The prior art data transmission system using a server computer is outlined in a block diagram in FIG.
14
.
In the prior art system, a plurality of terminal computers
12
are connected to a server computer
11
via switching hub
13
with a packet switching function, and upon demand from a terminal computer
12
, the server computer
11
transmits the wanted data to the terminal computer that sent in the demand.
The prior art network data transmitter is illustrated in further detail in FIG.
15
. The server computer
11
is provided with k units of cards
130
. each comprising a local CPU
131
(k) and a local memory
132
(k) as well as a port correspoonding to a local disk
133
(k); k units of send-only ports Ps
1
(k), one for each card
130
(k); and one receive-only port Pr
1
. The suffix k in parentheses indicates the cardnumber and is an integer (in the case of
FIG. 15
, an integer of 0 to 7).
The switching hub
13
is equipped with k units of ports Ps
2
(k) as opposed to the aforesaid send-only ports Ps
1
(k) on the server computer and a port Pr
2
opposite to the above-mentioned receive-only port Pr
1
also on the server computer
11
as well as a plurality of ports p(a) to p(c) for the terminal computers
12
(in the case of
FIG. 15
,
3
terminal ports). To each of the terminal ports p(a) to p(c) can be connected one or more terminal computers
12
. In
FIG. 15
, three units of terminal computers
12
-
1
to
12
-
3
are connected to the terminal port p(a), and one terminal computer
12
-
4
to the terminal port p(b) and one
12
-
5
to the terminal port p(c) respectively.
The local disks
133
(k) on the aforesaid server computer combined forms a logical storage area. Physically, one data is fragmented in blocks of, say one kilobyte, and stored there.
FIG. 16
is a functional block diagram showing the configuration of the server computer in still more detail. The operating process of the prior art system will now be described further elaborately, referring to the figure.
If the terminal computer
12
-
1
sends in a data processing demand packet Pd
1
as shown in FIG.
17
(
a
), for example, the packet is transmitted to the server computer
11
(the receive only port Pr
1
) via the terminal port p(a) and the port Pr
2
on the switching hub
13
. By the way, the data processing demand packet Pd
1
contains the following information as shown in FIG.
17
(
a
): address (server computer address) (
1
), sender's address (address of the terminal computer making a data processing demand) (
2
), data reading out demand (
3
), file number (
4
), address of data to be read out (address on the logical storage area formed by the whole local disks) (
5
) and data size (
6
). The data processing demand packet Pd
0
which is received at a packet receiver
101
in the server computer is relayed to a packet processor
102
where the processing demand Pd
1
is deciphered and sent on to a data processor
103
.
The data processor
103
in turn instructs a data reader
105
to read out data at a specified address, the reader provided on the local CPU
131
(k) in each card
130
(k). According to the instructions, the data reader
105
reads out the specified data on to a local memory
132
(k) from the local disk
133
(k) provided in the card
130
(k).
Further, the data processor
103
hands over the “sending” demand packet Ps
1
(k) shown in FIG.
17
(
b
) to a sending data packet processor
104
in the card
130
(k), and instructs the processor to incorporate the read out data into a packet. A group of sending data packets thus formed in the card
130
(k) is sent out from the packet sender
106
via the port Ps
1
(k) provided for the card
130
(k). It is noted that the “sending” demand packet Ps
1
(k) carries the following information as shown in FIG.
17
(
b
): destination address (address of the terminal computer) (
1
), sender's address (address of the server computer
11
) (
2
), packet number (
7
), address of the data to be read out (address of the local disk for the data to read out) (
8
), and data size (
6
).
The sending data packet thus sent out is outputted to the switching hub
13
, where the packet is stored in the buffer of the port P(a) for the destination terminal computer
12
-
1
before being send out therefrom. The data packet outputted this way has the address written at the header, and can be taken in by the terminal computer
12
-
1
at that address. Receiving a group of sending data packets, the terminal computer
12
-
1
edits the packets of data in a proper order and uses them for the intended purpose.
The question is that there is a possibility that the size of a single sending data packet will exceed the maximum permissible volume of a packet that flows on the network. To cope with it, a plurality of data packets, say 8 sending data packets, are usually grouped into one unit. In the ethernet, the maximum permissible packet size is 1.5 kilobytes for single packets. But if a plurality of data packets are grouped into a unit, up to 8 kilobytes can be transmitted as one unit. In an example with 8 local disks, a unit of 8 kilobytes of data packets can be formed with one kilobyte allocated for each local disk
133
(k).
In the foregoing system configuration, however, there arises a possibility that a plurality of sending data packet units or groups may be transmitted almost simultaneously and successively to a specific terminal port on the switching hub to which the terminal computers
12
are connected.
When processing demands happen to come almost simultaneously from terminal computers
12
-
1
and
12
-
2
both connected to the port P(a) on the switching hub
13
, for example, two units of sending data packets to be received by the terminal computers
12
-
1
and
12
-
2
[(the number of packets in a send packet group)×2] will rush to the port P(a) on the switching hub
13
almost at the same time.
Also, when data processing demand packets Pd
0
are sent almost simultaneously from three computers
12
-
1
to
12
-
3
, for example, the switching hub
13
will have deal with the same number of sending data packets [(number of packets in a send packet group)×3] the same way.
While a buffer is provided in the terminal port p(a) on the switching hub
13
, a simultaneous rush of data could exceed the capacity of the buffer. If too large a size of data rushes in the buffer, the part of the data that fails to be stored in the buffer will be discarded. That is, the units of data sent out from the server computer
11
will come in with the data for the terminal computer
12
-
1
or
12
-
2
cut off in part. That presents such problems as interrupted moving images as in a moving image data system, for example, where real time processing is required.
The present invention is proposed in view of such problems that are encountered with the prior art system. It is an object of the present invention to provide a server computer so configured that if there are simultaneous processing demands from a plurality of terminal computers connected to the switching hub, send data would not be concentrated in one terminal port on the switching hub.
SUMMARY OF THE INVENTION
To attain the object of the present invention, a number of means are adopted on a server computer
11
and system in which data processing demand packets Pd
0
sent out from a plurality of terminal computers on the network are received via a specific terminal port on the switching hub, and a group of sending data packets is formed against the data processing demand packets Pd
0
and then sent, via the specific port, to the terminal computers
12
that send in the data processing demand packets Pd
0
. In other words, the prior art server computer and the system form a basis on which the
Okamoto Keiji
Yoshida Yasuhiro
Jaroenchonwanit B.
Matsushita Electric - Industrial Co., Ltd.
McDermott & Will & Emery
Rinehart Mark H.
LandOfFree
Method of sending data from server computer, storage medium,... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of sending data from server computer, storage medium,..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of sending data from server computer, storage medium,... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2617398