Connection manager capable of supporting both distributed computing sessions and non distributed computing sessions

• US 7,945,677 B2
• Filed: 09/06/2005
• Issued: 05/17/2011
• Est. Priority Date: 09/06/2005
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: Stanev, Georgi, Luik, Oliver, Fleischer, Christian
• Primary Examiner(s): Colin; Carl
• Assistant Examiner(s): THIAW, CATHERINE B
• Application Number:
  US11/221,243
• Publication Number:
  US 20070055781A1
• Time in Patent Office: 2,079 Days
• Field of Search: 709227-229, 709217-219, 709/203, 709212-215, 702/32, 702/35, 707 1- 9, 718/105, 718/100, 719/312
• US Class Current: 709/227
• CPC Class Codes: H04L 67/133 Protocols for remote procedure calls [RPC]
• CPC Class Codes: H04L 67/14 Session management for real-time applications in data packet communications networks H04L65/1066

Abstract

A method is described that involves establishing a connection over a shared memory between a connection manager and a worker node. The shared memory is accessible to multiple worker nodes. Then sending, from the connection manager to the worker node over the connection, a first request containing a method call to a remote object on the worker node. Also sending, from the connection manager to the worker node over the connection, a second request containing a second method call to a second remote object on the worker node.

Claims

1. A method comprising:

establishing a session between a cluster of application servers and an external client;

determining that said session is a distributed session, said distributed session involves bi-directional request and response flows;

establishing a connection, associated with said distributed session, over a shared memory between a connection manager and a worker node, said shared memory, connection manager, and worker node being located within a first application server within said cluster of application servers, said shared memory accessible to multiple worker nodes within said first application server;

receiving, at said connection manager, a first distributed computing request from a external client over a network;

sending, from said connection manager to said worker node over said connection, said first distributed computing request containing a first method call to a first worker object, said first worker object instantiated on a virtual machine within said worker node, said first worker object being a communication endpoint between said first worker object and said external client;

preparing a first response to said first distributed computing request on said virtual machine within said worker node;

sending said first response from said worker node to said connection manager through said connection; and

,

sending, from said connection manager, said first response to said first external client;

receiving, at said connection manager, a second distributed computing request from said external client over said network;

sending, from said connection manager to said worker node over said connection, said second distributed computing request containing a second method call to a second worker object, said second worker object instantiated on said virtual machine, said second worker object being a communication endpoint between said second worker object and said external client;

preparing a second response to said second distributed computing request on said virtual machine within said worker node;

sending said second response from said worker node to said connection manager through said connection;

sending, from said connection manager, said second response to said external client

establishing a second session between said cluster of application servers and a second external client;

determining that said second session is a non-distributed session;

establishing a second connection, associated with said non-distributed session, over said shared memory between said connection manager and said worker node;

sending, from said connection manager to said worker node through said second connection, a non distributed computing request containing a command for an application that is instantiated on said virtual machine of said worker node; and

,

tearing down said second connection in response to said worker node having responded to said non distributed computing request.

2. The method of claim 1 wherein said first and second distributed requests are formatted according to a protocol selected from the group consisting of:

an IIOP request;

an RMI request;

a CORBA request; and

,

an RMI/IIOP request.

3. The method of claim 1 wherein said non distributed computing request is formatted according to a protocol selected from the group, consisting of:

an HTTP request; and

,

an HTTPS request.

4. The method of claim 3 comprising:

said connection manager executing a session request/response communication protocol for said first distributed computing request; and

,

invoking a first worker object'"'"'s skeleton on said worker node.

5. The method of claim 1, wherein said second external client is a second application server of said cluster of application servers.

6. An article of manufacture including program code stored on a non-transitory machine readable medium which, when executed by a machine, causes the machine to perform a method, the method comprising:

establishing a session between a cluster of application servers and an external client;

determining that said session is a distributed session, said distributed session involves bidirectional request and response flows;

establishing a connection, associated with said distributed session, over a shared memory between a connection manager and a worker node, said shared memory, connection manager, and worker node being located within a first application server within said cluster of application servers, said shared memory accessible to multiple worker nodes within said first application server;

receiving, at said connection manager, a first distributed computing request from said external client over a network;

sending, from said connection manager to said worker node over said connection, said first distributed computing request containing a first method call to a first worker object, said first worker object instantiated on a virtual machine within said worker node, said first worker object being a communication endpoint between said first worker object and said external client;

preparing a first response to said first distributed computing request on said virtual machine within said worker node;

sending said first response from said worker node to said connection manager through said connection; and

,

sending, from said connection manager, said first response to said first external client;

receiving, at said connection manager, a second distributed computing request from said external client over said network;

sending, from said connection manager to said worker node over said connection, said second distributed computing request containing a second method call to a second worker object, said second worker object instantiated on said virtual machine, said second worker object being a communication endpoint between said second worker object and said external client;

preparing a second response to said second distributed computing request on said virtual machine within said worker node;

sending said second response from said worker node to said connection manager through said connection;

sending, from said connection manager, said second response to said external client

establishing a second session between said cluster of application servers and a second external client;

determining that said second session is a non-distributed session;

establishing a second connection, associated with said non-distributed session, over said shared memory between said connection manager and said worker node;

sending, from said connection manager to said worker node through said second connection, a non distributed computing request containing a command for an application that is instantiated on said virtual machine of said worker node; and

,

tearing down said second connection in response to said worker node having responded to said non distributed computing request.

7. The article of manufacture of claim 6 wherein said first and second distributed requests are formatted according to a protocol selected from the group consisting of:

an IIOP request;

an RMI request;

a CORBA request; and

,

an RMI/IIOP request.

8. The article of manufacture of claim 6 wherein said non distributed computing request is formatted according to a protocol selected from the group, consisting of:

an HTTP request; and

,

an HTTPS request.

9. The article of manufacture of claim 8 wherein said method further comprises:

said connection manager executing a session request/response communication protocol for said first distributed computing request; and

,

invoking a first worker object'"'"'s skeleton on said worker node.

10. The article of manufacture of claim 9 wherein said first and second distributed requests are formatted according to a protocol selected from the group consisting of:

an IIOP request;

an RMI request;

a CORBA request; and

,

an RMI/IIOP request.

11. A computing system comprising a machine, said computing system also comprising instructions disposed on a non-transitory computer readable storage medium, said instructions causing said machine to:

establish a session between a cluster of application servers and an external client;

determine that said session is a distributed session, said distributed session involves bidirectional request and response flows;

establish a connection, associated with said distributed session, over a shared memory between a connection manager and a worker node, said shared memory, connection manager, and worker node being located within a first application server within said cluster of application servers, said shared memory accessible to multiple worker nodes within said first application server;

receive, at said connection manager, a first distributed computing request from said external client over a network;

send, from said connection manager to said worker node over said connection, said first distributed computing request containing a first method call to a first worker object, said first worker object instantiated on a virtual machine within said worker node, said first worker object being a communication endpoint between said first worker object and said external client;

prepare a first response to said first distributed computing request on said virtual machine within said worker node;

send said first response from said worker node to said connection manager through said connection; and

,

send, from said connection manager, said first response to said external client;

receive, at said connection manager, a second distributed computing request from said external client over said network;

send, from said connection manager to said worker node over said connection, said second distributed computing request containing a second method call to a second worker object, said second worker object instantiated on said virtual machine, said second worker object being a communication endpoint between said second worker object and said external client;

prepare a second response to said second distributed computing request on said virtual machine within said worker node;

send said second response from said worker node to said connection manager through said connection;

send, from said connection manager, said second response to said external client

establish a second session between said cluster of application servers and a second client;

determine that said second session is a second non-distributed session;

establish a second connection, associated with said second non-distributed session, over said shared memory between said connection manager and said worker node;

send, from said connection manager to said worker node through said second connection, a non distributed computing request containing a command for an application that is instantiated on said virtual machine of said worker node; and

,

tear down said second connection in response to said worker node having responded to said non distributed computing request.

12. The computing system of claim 11 wherein said first and second distributed requests are formatted according to a protocol selected from the group consisting of:

an IIOP request;

an RMI request;

a CORBA request; and

,

an RMI/IIOP request.

13. The computing system of claim 11 wherein said non distributed computing request is formatted according to a protocol selected from the group, consisting of:

an HTTP request; and

,

an HTTPS request.

14. The computing system of claim 13 wherein said instructions disposed on a non-transitory computer readable storage medium further cause said computer to:

execute, within said connection manager, a session request/response communication protocol for said first distributed computing request; and

,

invoke a remote object'"'"'s skeleton on said worker node.

15. A method, comprising:

establishing a first session to manage computing requests received by a cluster of application servers;

establishing a second session to manage computing requests received by said cluster of application servers;

determining that said first session is a distributed session and said second session is a non-distributed session;

establishing a quasi-permanent connection, associated with said distributed session, over a shared memory between a connection manager and a first worker node operating on a common server, said shared memory accessible to multiple worker nodes, and wherein said quasi-permanent connection remains active through multiple request/response cycles;

establishing a second connection, associated with said non-distributed session, over said shared memory between said connection manager and a second worker node operating on said common server;

receiving at a dispatcher a first distributed computing request containing a first method call to a first worker object, said first worker node is running an instance of said first worker object;

sending, from said dispatcher to said connection manager, said first distributed computing request over a network connection;

sending, from said connection manager to said first worker node over said quasi-permanent connection, said first distributed computing request;

receiving at said dispatcher a second non-distributed computing request containing a second method call to a second worker object, said second worker node is running an instance of said second worker object;

sending, from said dispatcher to said connection manager, said second non-distributed computing request over a network connection;

sending, from said connection manager to said second worker node over said quasi-permanent connection, said second non-distributed computing request.

16. The method of claim 15, further comprising:

receiving at said connection manager a third distributed computing request containing a third method call to said first work object over a network connection, said third distributed computing request received from an application server within said cluster of application servers;

sending, from said connection manager to said first worker node over said quasi-permanent connection, said third distributed computing request.

17. A method, comprising:

establishing a first session to manage computing requests received by a cluster of application servers;

determining that said first session is a distributed session, said distributed session involves bi-directional request and response flows;

establishing a second session to manage non-distributed computing requests received by said cluster of application servers;

establishing a quasi-permanent connection, associated with said distributed session, between a connection manager and a virtual machine instance running on a common server within said cluster of application servers, said quasi-permanent connection established within a share memory on said common server;

establishing a second connection, associated with said second session, between said connection manager and said virtual machine instance within said shared memory;

sending, from said connection manager to said virtual machine instance over said quasi-permanent connection, a first distributed computing request;

receiving, from said virtual machine instance at said connection manager over said quasi-permanent connection, a response to said first distributed computing request;

sending, from said connection manager to said virtual machine instance over said second connection, a first non-distributed computing request;

receiving, from said virtual machine instance at said connection manager over said second connection, a response to said first non-distributed computing request;

closing said second connection associated with said second session;

sending, from said connection manager to said virtual machine instance over said quasi-permanent connection, a second distributed computing request;

receiving, from said virtual machine instance at said connection manager over said quasi-permanent connection, a response to said second distributed computing request; and

closing said first session;

tearing down said quasi-permanent connection between said connection manager and said virtual machine instance;

wherein quasi-permanent connections remain active through multiple request/response cycles.