Electrical computers and digital processing systems: multicomput – Computer-to-computer data addressing
Reexamination Certificate
2000-05-31
2004-01-20
Geckil, Mehmet B. (Department: 2142)
Electrical computers and digital processing systems: multicomput
Computer-to-computer data addressing
C709S230000
Reexamination Certificate
active
06681258
ABSTRACT:
TECHNICAL FIELD
The present invention relates in general to network communications of processing systems. The invention relates to techniques for effecting communications between a network and multiple partitions of a data processing system employing a host-network interface. More particularly, the invention relates to a facility for retrieving data from a network adapter having an address resolution table shared by multiple communications stacks coupled to the network adapter. Specific parameters can be selectively retrieved and manipulated by an initiating communications stack.
BACKGROUND OF THE INVENTION
Mainframe class data processing systems have hardware and software facilities that enable partitioning thereof. Such processing systems may be subdivided into multiple partitions whereby a user of a partition, or software executing in a partition, has the impression that the processing system is exclusively used by that application. Each partition has the appearance of being a separate and distinct processing system and may even run its own multi-tasking and multi-user operating systems independent from each other partition. An IBM Enterprise Systems Architecture (“ESA”)/390 Mainframe Computer is an example of one such partitionable mainframe class data processing system. Partitioning thereof is described in, for example, various publications by International Business Machines Corporation, including “IBM ESA/390 Principles of Operation”, IBM Publication No. SA22-7201-02, December 1994, and in the “IBM Enterprise System/9000 Processor Resource/Systems Manager Planning Guide”, IBM Publication No. GA22-7123-11 (April 1994), which are both hereby incorporated herein by reference in their entirety.
Software executing in individual partitions within the mainframe class data processing system may require a network connection such as a local area network (“LAN”) connection or a wide area network (“WAN”) connection. This may be used to facilitate connectivity to users, or to application programs used in, for example, a client-server processing environment. Shown in
FIG. 1
is a conventional configuration used to connect individual partitions, including the software running therein, to a LAN. The configuration includes a processing system
11
that has partitions
13
,
15
,
17
,
19
,
20
and
21
.
Network connectivity for each partition of system
11
of
FIG. 1
is achieved using separate network interfaces for each partition. For example, partition
13
is conventionally connected through channel connection
29
to an IBM 3172 Interconnect Controller
23
(with 8232 Channel Interface Attachment) that has, for example, a token ring or Ethernet LAN port
32
attached to LAN
37
thereby providing LAN connection
31
. Network connectivity is accordingly directly provided between partition
13
and computers
43
and
45
on LAN
37
through the IBM 3172 23. However, according to conventional techniques, this configuration has no other direct logical or physical connections from any of the other partitions to LAN
37
. To further note, each application within partition
23
must communicate with a different network port on IBM 3172 23. The IBM 3172 (having internal 8232 Channel Interface Attachment), is described in a publication entitled “8232 LAN Channel Station”, Apr. 15, 1998, IBM Publication No. ZZ25-8577-0, that is incorporated herein by reference in its entirety.
The conventional software executing on IBM 3172s restricts direct logical connectivity to being between a single partition and its corresponding LAN. Thus, to facilitate direct connectivity from a computer
47
on a LAN
39
to both partition
17
and
21
, multiple IBM 3172s would traditionally be used. Partition
17
is coupled to LAN
35
via channel connection
29
′n, IBM 3172 25 and LAN port
34
thereby establishing LAN connection
33
. Similarly, partition
21
is coupled to LAN
39
via channel connection
29
″, IBM 3172 27, and LAN port
35
thereby establishing LAN connection
36
.
An enhanced network interface for a mainframe class data processing system having multiple partitions and a port to a network is described in commonly assigned U.S. Pat. No. 5,740,438, which is hereby incorporated herein by reference in its entirety. Briefly summarized, this patent describes establishing a table which defines communications paths between the port to the network and at least two partitions of the multiple partitions. More specifically, each partition has at least one application executing therewithin and the communications paths are defined thereto. Data frames are passed between the network and the applications within the partitions through the port to the network and along the communications paths defined in the table such that the network communications is effected.
One embodiment of the network interface approach of U.S. Pat. No. 5,740,438 is depicted in FIG.
2
. Shown is a partitionable mainframe class data processing system
11
(e.g., an IBM Enterprise System/9000) having an integral host-to-network interface (“HNI”)
51
that facilitates a LAN connection
55
from multiple partitions
13
,
15
,
17
,
19
,
20
and
21
to LAN
53
through LAN port
54
. Each application in each partition may directly communicate with computers
61
,
63
and
65
on LAN
53
through the single host-to-network interface
51
and single LAN port
54
. The LAN shown is a token ring LAN; however, the system is equally applicable to other types of LANs such as, for example, Ethernet and Fiber Distributed Data Interface (“FDDI”). Further, the host-to-network interface may support multiple network connections by way of multiple network ports. For example, a WAN connection
57
comprising, for example, a Peer-to-Peer Protocol (“PPP”) connection may be established to a computer
59
through WAN port
56
. Any mix of LAN and WAN connections among multiple ports of host-to-network interface
51
is possible.
The host-to-network connectivity techniques described above have certain limitations, particularly in an Ethernet environment where two different frame types are possible, i.e., Ethernet DIX and Ethernet 802.3. For client/server systems, Transmission Control Protocol (TCP)/Internet Protocol (IP) has become the leading protocol for network communications. Using Ethernet, when a client application running over TCP/IP wants to communicate with a server application, the application must specify one of the two existing Ethernet frame formats. The frame format must also be known by the server machine in order for the TCP/IP connection to be established and any data transfer to occur. Conventionally, in order to make sure that the client and server communicate using the same Ethernet frame formats, both the client and server must specify the specific frame format to be used in their appropriate configuration files. If the configuration files do not match, then the client/server application will not work properly.
The most common server TCP/IP environment today has the complete TCP/IP functionality on one platform. For example, reference “TCP/IP Tutorial and Technical Overview,” IBM Document No. GG24-3376-03 (December 1992). In this environment, one device driver exists in each partition for each LAN connection. Each device driver can specify a different Ethernet frame format, but will not support both frame formats. Thus, a different device driver is used for each of the two frame formats. In operation, a dedicated device driver of a partition of the host system takes care of providing both channel headers and media or LAN headers necessary for transmission of an IP packet across the LAN to a client coupled thereto.
With the above as background, the present invention is directed to enhancements to the state-of-the-art of network communications of multiple partitions employing a host-network interface, and in particular, to enhanced techniques for retrieving data from a network adapter having an address resolution table shared by multiple communications stacks of the multi-partition processing system.
DISCLOSURE OF THE
Ratcliff Bruce H.
Stagg Arthur J.
Valley Stephen R.
Geckil Mehmet B.
Gonzalez, Esq. Floyd A.
Heslin Rothenberg Farley & & Mesiti P.C.
Radigan, Esq. Kevin P.
LandOfFree
Facility for retrieving data from a network adapter having a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Facility for retrieving data from a network adapter having a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Facility for retrieving data from a network adapter having a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3255183