Patent application title: SYSTEM AND METHOD FOR COLLABORATIVE PROCESSING OF DISTRIBUTED APPLICATIONS
Inventors:
Robert Betros (San Diego, CA, US)
Mansour Raad (Winchester, MA, US)
IPC8 Class: AG06F15173FI
USPC Class:
709223
Class name: Electrical computers and digital processing systems: multicomputer data transferring computer network managing
Publication date: 2011-09-08
Patent application number: 20110219107
Abstract:
A system and method for generating and executing machine-language filter
code. A filter generator receives a text-based filter string. The filter
string contains filter criteria for filtering messages communicated from
one computer program to another computer program or to the same computer
program. A machine language generator converts the filter string to a
machine-language filter code. A central processing unit of a host loads
and executes the machine-language filter code to process or discard
messages based on the filter criteria. The machine-language filter code
is maintained by the computer program for continual reuse eliminating the
interpretation and translation of the filter string from data to code for
each message.Claims:
1. A method of filtering messages, comprising: receiving a text-based
filter string representing filter criteria; automatically dynamically
converting the text-based filter string directly to machine-language
filter code once using a machine language generator for use by a code
section of a computer program, the machine language generator being
configured to dynamically convert the text based filter string directly
to machine language filter code; and using the code section, executing
the machine-language filter code to accept or discard multiple messages
received by or residing within one or more software components within the
computer program based on the filter criteria.
2. The method of claim 1, wherein executing the machine-language filter code is performed by a central processing unit running the computer program.
3. The method of claim 1, wherein the machine-language filter code is directly executable by the central processing unit.
4. The method of claim 1, wherein the text-based filter string is human-readable.
5. The method of claim 1, wherein the text-based filter string is formatted according to a conditional expression syntax.
11. A system of filtering information transmitted from one or more software components within a computer program to one or more other software components within another computer program or the same computer program, comprising: a filter generator configured to receive a text-based filter string representing filter criteria; a machine language generator coupled with the filter generator and being configured to automatically dynamically convert the text-based filter string directly to machine-language filter code; and a processor for executing the machine-language filter code for a software component of a computer program to accept or discard information received by the software component within the computer program based on the filter criteria.
12. The system of claim 11, wherein the processor is included with a central processing unit.
13. The system of claim 11, wherein the machine-language filter code is directly executable by the central processing unit.
14. The system of claim 11, wherein the filter string is human readable.
15. The system of claim 11, wherein the filter string is formatted according to a conditional expression syntax.
Description:
CROSS-REFERENCE TO RELATED PATENT APPLICATIONS
[0001] This application is a divisional of U.S. patent application Ser. No. 11/485,229, filed Jul. 11, 2006, which is a continuation of U.S. patent application Ser. No. 09/766,382, filed Jan. 19, 2001, now issued as U.S. Pat. No. 7,080,120, the contents of all of which are incorporated herein by reference.
BACKGROUND OF THE INVENTION
[0002] The present invention generally relates to computer systems, and more particularly to executing collaborative application software in a computer network.
[0003] Computer networks have become an important platform for storing and delivering application programs to users. In a model widely referred to as client/server, an application server is considered a central hub from which a number of clients can request and receive application software for performing a specific function. In the past, most applications were executed locally at a client, with application software that was locally stored in the client platform. Today, however, in one configuration, applications are stored in a central server, and delivered to a client in components for local processing. Another common configuration is for the client to send processing requests to the server, which executes the application that is specified in the request on the server and sends the results back to the client for local processing. In this configuration, the application is maintained in the server.
[0004] The tremendous growth of use of the Internet and World Wide Web (the "Web") in the past few years has resulted in increasingly more communication between clients and servers. Most of the communications in the short history of the Web relate to delivery of requested content data from a web server to a client. A web server is a specialized type of server that communicates according to a particular protocol, the hypertext transport protocol, or HTTP. Other hypermedia protocols are also used or are being developed.
[0005] Traditionally, Internet-based application processing in a client/server model occurs mostly through a web server. Network-based applications for the client/server model, such as those through the Web, are becoming increasingly complex and interactive. Problems with current software infrastructure for administering network-based applications include difficulty in the creation and delivery of application components, complexities of managing the interface between a client and a server, and processing delays and inefficiencies. The current and next generations of the Web, in its Internet, intranet and extranet embodiments, demand a platform-neutral architecture that can effectively handle complex Web-based distributed client/server applications for collaborative processing.
[0006] A host is a physical machine having operating system software that manages one or more other computer programs that reside on the physical machine. A computer program includes one or more software components that work together to perform the logic of the program. Each computer program can communicate information with other computer programs on the same host or on a different host. A unit of information to be communicated is referred to as a payload. Before transmission, the payload is supplemented with one or more properties. Properties include data for identifying information within and about the payload. A message represents a combination of the properties and the payload.
[0007] A computer program includes a code section and a data section. The code section contains machine-language code, which is a sequence of byte values that are arranged in such a way to be executed directly by the central processing unit (CPU) of the host, to perform a specific task. The data section is a sequence of byte values, that is used to store data structures required by the computer program. The data section byte values may or may not be arranged in machine-language code.
[0008] An important mechanism for efficient communication of messages from one computer program to another is a filter, by which messages of interest are communicated and all other messages are discarded. A filter string is a text-based expression of logic that is readable by a human user, and which is executed by a computer program for performing a filtering task. One example of a filter string is a Boolean logic expression that includes Boolean operators such as "and" or "or" and references to the message properties and or the payload. Computer programs can include one or more filters to execute the Boolean logic expression on each received message to determine if the message should be processed or disregarded by the computer program.
[0009] Typically, when a computer program performs filtering, a filter generator is used. The filter generator accepts a filter string and reorganizes the filter string into a data structure. The data structure represents the filter string, and is stored in the data section of the computer program employing the filter. Typically, the filter string is represented as a tree structure. Each time a new message is received by the computer program, the data structure representing the filter string is retrieved from the data section and is interpreted to execute machine-language code specific to the host machine in which the computer program resides. Since the machine-language code is specific to a particular type of host, it is an effective means of filtering messages.
[0010] However, the interpretation of the data structure created from the filter string to execute the machine-language code is required for every message and adds significant processing delay, especially when a large volume of messages are being received by an computer program.
SUMMARY OF THE INVENTION
[0011] The aforementioned problems are overcome with a message filtering system and method that converts the filter string directly into machine-language code once for reuse on multiple messages, thereby eliminating the intermediate step of interpreting the filter string representation to execute the appropriate machine-language code on each message. In accordance with one embodiment of the invention, a method of filtering information transmitted from one computer program to another computer program or to the same computer program includes receiving a textbased filter string representing filter criteria, and converting the filter string directly into machine-language code. The embodiment further includes loading the machine-language code into a computer program to accept or discard information based on the filter criteria.
[0012] In accordance with an alternative embodiment of the invention, a system of filtering information transmitted from one computer program to another computer program or to the same computer program includes a filter generator configured to receive a text-based filter string representing filter criteria, and a machine-language code generator coupled with the filter generator and being configured to convert the filter string to machine-language code. The system further includes a mechanism by which the machine-language code is loaded into a computer program. Once loaded into the computer program, the machine-language code is termed machine-language filter code and is continually executed to accept or discard information based on the filter criteria.
BRIEF DESCRIPTION OF THE DRAWINGS
[0013] FIG. 1 shows a system for collaborative processing with distributed applications, according to one embodiment of the invention.
[0014] FIG. 2 shows communication between a host and a remote host via the messaging bus.
[0015] FIG. 3 shows a message filtering system according to one embodiment of the invention.
[0016] FIG. 4 shows a method for filtering messages according to an embodiment of the invention.
DETAILED DESCRIPTION
[0017] This invention relates to executing and processing applications. An application is software that performs specific logic. As used herein, the term application refers to an entire application, or an application component that defines a portion of an application. A number of application components can work together to execute the logic of the application. In one embodiment, this invention provides a method for communication between client-side application components and server-side application components. Further, the invention provides for the execution and administration of applications residing within a CGI engine hosted on a server platform. The invention provides a structure for performing collaborative processing with distributed applications.
[0018] In a preferred embodiment of the invention, a system and method for collaborative processing of distributed applications is embodied as one or more computer programs configured to work together. Every computer program includes at least one code section and at least one data section. The code section is a sequence of bytes that can be executed directly by the central processing unit (CPU) of a host. The data section includes sequences of bytes that may be executed by the CPU of the host, but which are mostly used to store data structures for use by the code section.
[0019] FIG. 1 shows a system 100 for collaborative processing of distributed applications, according to one embodiment of the invention. The system 100 includes a host 102. In one embodiment, the host 102 is a physical machine having an operating system that manages one or more computer programs. A computer program can include one or more operations running within it that share the same data section of the computer program. The host 102 can also include a CPU that executes the code section of the computer programs. The computer programs can be resident on the host 102 in permanent or temporary storage, or can be communicated to the host 102 by external computer programs.
[0020] The host 102 includes a computer program 130, which includes a CGI engine 104, which in an embodiment is an operation configured to execute and administer one or more CGIs, and a messaging bus 120, which in an embodiment is another operation configured to enable multiple operations within the computer program to communicate with each other. In a specific exemplary embodiment, the CGIs administered by the CGI engine 104 are servlets. A servlet is a particular type of CGI that can be configured to execute a certain function, such as manage an application, for example. While a servlet is typically written in the Java programming language, a CGI can be written according to other programming languages or logic structures. According to the invention, the CGI engine 104 is configured to function as an application server, including, without being limited to, logging information related to the use and operation of applications, authentication of users of CGIs, and authorization of the usage of the CGIs.
[0021] The host can also include a web server 106. The web server 106 functions as a conventional web server that communicates with a client according to the hypertext transfer protocol (HTTP). As used herein, HTTP means any application-level protocol that is supported by a web server or a web browser, including but not limited to secure HTTP (HTTPS). In an embodiment of the invention, the web server 106 also communicates with the CGI engine 104, from which the web server 106 executes one or more CGIs running in the CGI engine 104. In a specific embodiment, the web server 106 invokes and executes one or more servlets, through which communications are performed between the web server 106 and an application 114. One particular type of CGI according to the invention is an application CGI 112. The application CGI 112 is configured to execute and manage an associated application 114, in response to requests received from the web server 106. The application CGI 112 can be a servlet. One specific type of servlet for the application CGI is called a sublet. A sublet is a servlet extension that is specifically configured for executing and managing an associated application.
[0022] The CGI engine 104 can be configured with at least two contexts for running CGIs. One context is an application context 108. The application context 108 is a self-sustained environment of resources within the CGI engine 104 that is configured to, but not limited to, enable multiple CGIs and other programs to commonly share resources. The application context 108 also provides a secure environment in which an application CGI 112 and an application 114, and its supporting resources, can operate shielded from corruption by other CGIs or applications running in the CGI engine 104. For each application 114, the CGI engine 104 is configured to create an application context 108 that includes an application CGI 112 and one or more administration CGIs 116 associated with the application CGI 112. The application context 108 enables CGIs, such as administration CGIs 116, and application CGIs 112 with features that include, but are not limited to, getting the actual location of a physical resource within the host 102 from a representation of that resource, and logging messages and exceptions being generated by the application. The application context 108 is further configured to enable the application 114 to get a reference, i.e. find where in a memory, to other applications within the CGI engine 104 that are registered to communicate with the application 114. The reference can be based on a name of the application 114.
[0023] A second context is a gateway context 110, an environment having resources for performing two-way asynchronous communication between applications on the same host and different hosts. One form of such communication is notification, in which two programs, or two operations within a program or within multiple programs, communicate data with each other without being polled for the data by the other.
[0024] Applications 114 within the CGI engine 104 communicate with each other asynchronously via a messaging bus 120. In a specific embodiment, the messaging bus 120 has multiple inputs and outputs and is configured to communicate data between associated inputs and outputs. A unit of information that enters and leaves the messaging bus 120 is a message. There are multiple pairs of entry and exit points on the messaging bus 120. Each pair represents a topic. Publishing is the act of sending a message to a topic on the messaging bus 120. Subscribing is the act of registering to a topic, and receiving the messages published on that topic. An application 114 can be both a publisher and subscriber to one or more topics. When a message is published to a particular topic, the message is communicated to all subscribers registered to that topic simultaneously. In this way, collaborative communication and processing between publishers and subscribers to a topic, through a particular application, is enabled.
[0025] For efficient communication between applications, an application 114 or client-side application 105 can employ a filter on the messaging bus to scan the content of messages on a certain topic, and determine whether the application employing the filter will either process the message or disregard it. Such a filtering mechanism is described in U.S. patent application Ser. No. 09/766,383 entitled "System and Method for Generating Machine-Language Code from Readable Text Code for Information Filtering," incorporated by reference herein for all purposes and reproduced in part below in reference to FIGS. 3 and 4.
[0026] The messaging bus 120 can logically extend across multiple hosts, enabling external applications to communicate with applications 114 that reside within the local CGI engine 108. Extension of the messaging bus 120 can occur whether or not a firewall 109 or proxy server resides in the communication path between the local CGI engine 104 and the external applications. One example of an external application is a client-side application 105 that resides in a client host 101 and runs within a client application server 103.
[0027] In one embodiment, the messaging bus 120 is extended through the use of a multicast protocol and configured to communicate with one or more other messaging busses, and wherein each other messaging bus is resident on a remote host 135. FIG. 2 illustrates a multicast communication path 140 between the host 102 and any remote host 135. Each host must support this protocol.
[0028] When the communication path does not support the multicast protocol, an external application can communicate asynchronously with the messaging bus 120 through a direct socket connection to the messaging bus extension 122. In one embodiment, the messaging bus extension 122 is created using the socket select method. The socket select( )method enables the efficient handling of large numbers of simultaneous socket connections to the messaging bus 120. The socket select( ) method, which can be executed by the operating system of the host 102, waits until information is available on any of the open socket connections to the host 102. Upon the arrival of the information, the socket select( )method returns a data structure that indicates which socket connections have information available to be processed. The messaging bus extension 122 uses the information within the returned data structure to read the available information from each specified socket connection. The information read from each socket is converted to a message and published onto the messaging bus 102 on a specific topic.
[0029] The messaging bus extension 122 multiplexes the input and output operations of the socket to publish and subscribe messages. The multiplexing can be directed by the operating system of the host 102.
[0030] When the host 101 is located behind a firewall 109 or proxy server, the extension of the messaging bus is implemented as described in U.S. patent application Ser. No. 09/766,439 entitled "System and Method for Maintaining Two-Way Asynchronous Notification Between a Client and a Web Server," incorporated by reference herein for all purposes. A gateway CGI 118 implements the two-way asynchronous communication through a firewall 109 or proxy server. In one embodiment, the gateway CGI 118 can be one or more servlets, administered and executed within the gateway context 110. When an external application that is not behind a firewall 109 or proxy server communicates messages to the messaging bus 120, a messaging bus extension 122 is used.
[0031] External applications can reside on another host, or on the same host 102 in another CGI engine. External applications can also reside on a client host 101, where they are executed as a client-side application 105. A client host 101 can include a client application server 103 in which the client-side application 105 is pre-installed, or to which the client-side application 105 is dynamically delivered from the web server 106 in the host 102.
[0032] The CGI engine 104, in communication with the web server 106, provides one form of access to the application 114. The web server 106 is configured to direct certain client requests from a client 101 to the relevant application CGI 112 in the CGI engine 104.
[0033] In an exemplary embodiment of the invention, the application CGI 112 is configured as a sublet and interacts with the application 114 according to an interface definition. In a specific embodiment, the interface definition is an extension of a servlet interface definition. This interface definition includes, without limitation, the following operations: [0034] a) an initialization operation that accepts as an argument the application CGI configuration, [0035] b) a service operation to return information about the associated application 114, [0036] c) a destroy operation to be invoked by the termination of the execution of the CGI engine 104 that enables the release of reserved resources and terminates the application CGI 112 and its associated application 114, [0037] d) operations to load and save the current state of the associated application 114, [0038] e) operations to set and get the internal attributes of the associated application 114, [0039] f) an operation to invoke an operation within the associated application 114.
[0040] The initialization operation of each application CGI 112 receives the application CGI configuration and applies it to the associated application 114 to prepare it for use. The application CGI configuration includes, without limitation, the following information:
a) a list of initialization parameter names, b) an initialization parameter value for a given a parameter name, c) a reference to the application context 108.
[0041] The service operation of the application CGI 112 gathers and returns information about the associated application 114. This information includes without limitation: [0042] a) a list of information about the attributes of the application 114 that includes the attribute name, description, type, and read/write accessibility, [0043] b) a list of information about the operations within the application 114 that includes each operation's name, description, type of information returned and information about the operation's parameters. Information about an operation parameter includes the parameter name, description and type.
[0044] The application CGI 112 can handle HTTP requests through its service operation. The implementation of the service operation in the application CGI 112 is used to administer the associated application 114. The administration includes, but is not limited to, the retrieval of information about the application 114, the retrieval and modification of its attribute values, the invocation of an operation within the application 114 and the saving of the current state of the application 114. The information and status returned from the administrative tasks is forwarded to the administration servlet 116 to format this information for presentation purposes.
[0045] Operation of the system 100 is as follows. In one embodiment, the CGI engine 104 administers the application 114 through the use of the web server 106. For instance, a client host 101 initiates the execution the application 114 by submitting an HTTP request to the web server 106 from the client-side application 105. The client-side application 105 can be, for example, an application running inside of a web browser program. The web server 106 redirects the request to the CGI engine 104. The CGI engine 104 invokes the application CGI 112 associated with the application 114 to process the request, which returns a response containing the result of the request to the web server 106 who forwards it on to the client host 101, for use by the client-side application 105.
[0046] When the CGI engine 104 starts, it loads the configuration files associated with each CGI. The content of a configuration file is called the CGI configuration. For each CGI in the CGI engine 104, the CGI engine 104 scans the CGI configuration to determine which of the CGIs needs to be instantiated at startup time. The CGI engine 104 invokes a standard initialization operation on each CGI that is instantiated with the CGI configuration as an argument. Upon the successful completion of the initialization operation, the CGI is registered to run within the CGI engine 104.
[0047] For the application CGI 112, the standard initialization operation composes a sublet configuration from the content of the scanned configuration file, and invokes the initialization operation of the associated application 114 with the application CGI configuration as an argument. If the initialization operation is successful, then the load operation of the application CFI 112 is invoked. The load operation is used to load the previously saved state of the application 114. When the CGI engine 104 is instructed to terminate the application CGI 112, the destroy operation in each registered CGI is invoked. This inherently invokes a sublet destroy operation to terminate the application 114.
[0048] Administrative actions that can be requested by the client host 101 from the client-side application 105 and executed by the CGI engine 104 include, without limitation, initiating and terminating the application 114, and modifying its logic flow.
[0049] When the messaging bus 120 is extended to a client-side application 105 running within the environment of the client application server 103, the client-side application 105 can publish, subscribe and process messages from the messaging bus 120. The client-side application 105 receives a message, processes the message, and composes a data structure that can be understood by the client application server 103. The client-side application 105 then delivers the data structure to the client application server 103 and instructs it to process the content of the delivered data structure.
[0050] In a specific exemplary embodiment of the invention, the administration CGI 116 is implemented using the Java Server Page (JSP) specification. The information forwarded to the administration CGI 116 is packaged as an attribute value within the HTTP request. The administration CGI 116 retrieves the attribute value and reformats it into a document that is readable by the requesting client application 103. In a preferred embodiment, the document format is in HTML. As used herein, HTML means any document format that is supported by a web browser, including but not limited to Dynamic HTML (DHTML).
[0051] In the specific exemplary embodiment, the client application server 103 is a web browser, the client-side application 105 is an HTML document-with a Java Applet. The web server 106 that is running in conjunction with the CGI engine 104 delivers the HTML document with the Java applet to the web browser 103. Upon the completion of the delivery, the Java Applet is initiated and establishes a connection to the messaging bus 120 as described above. The Java Applet now contains an extension of the messaging bus 120 and is a publisher and/or subscriber to one or more topics on the messaging bus 120.
[0052] The Java Applet is configured to call specific functions in the HTML document upon the receipt of a message on a particular topic or topics. The functions within the HTML document or template accept as arguments the property values of the message and convert these values into a HTML formatted text string that is executed by a component within the HTML document or template.
[0053] In another embodiment, the client-side application 105 is a Java Applet and the HTML document or template can be delivered as a property value within a message. The Java Applet reads the HTML document or template within the message and instructs the web browser 103 to render the HTML document or template.
[0054] In still another embodiment, the client-side application 105 is a Java Applet that dynamically generates the HTML document or template and the result is forwarded to a component in the web browser 103 for rendering.
[0055] FIG. 3 shows a system 300 for generating machine-language code for message filtering. System 300 preferably includes a host 302. The host 302 represents a physical machine having a central processing unit (CPU) 320 and an operating system software that manages one or more computer programs that reside on the physical machine and which are executed by the CPU 320.
[0056] A computer program 303 includes one or more software components 304. In accordance with an embodiment of the invention, one such software component is a filter generator 310. The filter generator 310 accepts filter strings from a software component 304. The filter string defines an acceptance criteria by which the software component 304 processes or discards messages.
[0057] In one embodiment, filter strings are expressed in a text-based format that is readable by a human user, and whose syntax is based on the Structured Query Language (SQL) conditional expression syntax. The text can include, without limitation, one or more expressions separated by a Boolean operator such as "and" or "or" for example. Each expression is made up of two elements separated by a Boolean conditional operator such "less-than" or "greater-than." One element of the expression can be a numerical or functional value, or can be a reference to an element in the message.
[0058] The filter generator 310 includes a machine-language code generator 312. The filter generator 310 reorganizes each submitted filter string into a format that is acceptable to the machine-language code generator 312. The machine-language code generator 312 accepts the newly formatted filter string as an input, and generates machine language filter code 317. The machine language filter code 317 is then loaded into the computer program 303 to produces a filter software component 322 in machine-language code, which can be directly executed by the CPU 320 of the host 302.
[0059] According to an exemplary embodiment, the machine-language code generator 312 represents a Java® compiler, the computer program 303 represents a Java virtual machine (JVM), and the software component 304 is a Java class. The filter generator 310 accepts a filter string as an input, from which it generates a Java® text file that defines a Java® class, which implements a predefined filter interface definition.
[0060] The filter interface definition calls for the implementation of a function that returns a Boolean value and expects as an argument an implementation of a predefined message interface definition. The message interface definition is empty to enable the generalization of the filter to any specific message implementation. The returned Boolean value indicates to the software component 304 if the received message should be processed or disregarded.
[0061] The Java® text file is compiled by the Java® compiler 312 to produce a Java® class 317. The Java® class 317 is loaded into the computer program 303 to produce the filter software component 322 for message filtering. Accordingly, each time a message 330 is received by the computer program 303, the filter software component 322 evaluates the received message 330 by applying the filter criteria to the message and instructs the software component 304 to either process the message 335, or reject the message 330.
[0062] FIG. 4 is a flowchart illustrating a method 400 according to an embodiment of the invention. The method 400 can be executed at other functional levels of a computer system. At block 405, a filter string is received. The filter string may be received from a separate computer program residing on another host, or the same host, or from within a single computer program. The filter string is in a text-based format that is readable by a human user and whose syntax is based on the SQL conditional expression syntax.
[0063] At block 410, a determination is made whether the filter string requires conversion into machine-language filter code. The conversion into machine-language filter code is required if there is no recollection of a recently-performed translation of a similar filter string. If conversion is required, at block 415, the text-based filter string is prepared for the machine-language code generator. At block 420, the prepared filter string is converted into machine-language code by the machine-language code generator. If the conversion is not required, at block 425, a determination is made whether the machine-language filter code is already loaded into the computer program.
[0064] If the machine-language filter code is not loaded, then at block 430 the machine-language code is loaded into the computer program. At block 435, the computer program with the filter software component is executed on multiple incoming messages. Accordingly, the multiple incoming messages are evaluated against the filter criteria represented in the filter software component, to process or discard messages in the specified manner.
[0065] Other embodiments, combinations and modifications of this invention will occur readily to those of ordinary skill in the art in view of these teachings. Therefore, this invention is to be limited only by the following claims, which include all such embodiments and modifications when viewed in conjunction with the above specification and accompanying drawings.
User Contributions:
Comment about this patent or add new information about this topic: