Method for synchronous support of fault-tolerant and adaptive communication

Abstract

An invention is provided for fault-tolerant and adaptive communication in a distributed computing environment. The invention includes detecting an interruption in a digital connection on a network between two endpoints. Upon detection, a connection state value indicating a number of data elements processed by the application is obtained. The connection state value is then exchanged with the opposite endpoint over the network. In this manner, remaining data elements of the transaction can be received sequentially over the network from the opposite endpoint. In particular, the remaining data elements begin with a data element at a point indicated by the connection state value plus one.

Description

BACKGROUND OF THE INVENTION

[0002] 1. Field of the Invention

[0003] This invention relates generally to computer networking, and more particularly to application and protocol independent synchronous support of fault tolerant and adaptive communication.

[0004] 2. Description of the Related Art

[0005] With the increased use of mobile and remote computing, distributed processing has become a central element in many computer processing systems. Distributed processing has many different forms depending on the nature of the data and the objectives of the application. For example, one emerging form of distributed processing is mobile computing, such as used in telematics.

[0006] Telematics refers to systems used for communications, instrumentation, control, and information technology in the field of transportation. Over the years, manufacturers of on-road vehicles, such as automobiles, vans, trucks, buses, and so on, have utilized computer technology to enhance the operations of existing features and functions in the vehicles as well as to provide new features and functions. For example, programmed controllers, custom-designed processors, embedded systems, and/or computer modules have been developed that support or even control various kinds of mechanical equipment in vehicles. For example, programmed controllers or computer modules have been developed that control or support various engine functions, such as fuel injection, timing, and so on. Programmed controllers or computer modules have been developed to enhance or support operation of transmission systems, suspension systems, braking systems, and so on. The sophistication of these enhancements has advanced as the processing power available for these purposes has increased. It is expected that in the future more aspects of the mechanical equipment in vehicles will be controlled or supported by processors or controllers in order to enhance performance, reliability, safety, and to reduce emissions.

[0007] As can be appreciated, the distributed nature of telematic functions generally requires a digital distributed communication structure such as that provided in distributed computing environments. However, as with most communication, digital communication is subject to interruption or failure. As such, an ability to restart communication after interruption is important to most distributed applications.

[0008] FIG. 1 is a block diagram showing a typical prior art distributed environment 100. The distributed environment 100 includes a client application 102 executing on a client device, which is in communication with a server application 104 executing on a server device. Generally, the client application 102 and server application 104 communicate using a logical connection 112, which is a logical entity used by applications to exchange data between two endpoints, such as the client application 102 and the server application 104. Although the application programs 102 and 104 function as though the logical connection 112 were a physical entity, the logical connection 112 requires a communication channel to actually transmit data.

[0009] Software known as the communication stack 106a and 106b is used to map logical connections to communication channels, which includes the actual communication hardware 108a and 108b. The communication stacks 106a and 106b handle data routing, flow control, buffering, error correction, and other computing issues encountered in real-world communication. Entities, such as the client application 102 and the server application 104, use one or more logical connections 112 to communicate with other entities by sending and receiving data in a sequential fashion over a period of time. Due to the sequential nature of the communication, data communication typically is stateful, meaning that the completeness and order of the data transmitted should be preserved.

[0010] However, the flow of data over a connection can be interrupted for many reasons, such as the failure of the underlying communication hardware 108a and 108b or because the connection is rerouted to a more advantageous communication channel. During such interruptions data may be lost, corrupted, or reordered. In order to continue communication once the channel has been reestablished, a method should be in place that restarts data transfer. Unfortunately, many prior art systems restart lost connections from the beginning of the data transaction, thus re-transmitting the entire transaction.

[0011] For example, in FIG. 1, the server application 104 begins sending data elements 1114a, 2114b, 3114c, and 4114d to the client application 102. However, during the transmission the logical connection 112 between the server application 104 and the client application 102 is interrupted. As a result, the client application 102, in this example, received only two data elements, 1114a and 2114b, of the four data elements transmitted. Once the connection is reestablished, the prior art distributed environment 100 will restart the entire transmission beginning again with data element 1114a, thus re-transmitting the entire transaction.

[0012] If the connection state of the transaction was maintained when interruption occurred, the transaction can be restated at the point of the interruption. For example, in FIG. 1, the server application 104 can begin re-transmitting from data element 3114c and continue the remainder of the transaction, instead of re-transmitting from the beginning of the transaction with data element 1114a. However, prior art solutions to obtaining and maintaining connection state information for restarting after connection interruptions have been overly burdensome and inconsistent.

[0013] For example, some prior art systems have attempted to make fault tolerant logical connections transparent to the applications by implementing all the fault tolerant functionality in the communication stack 106a and 106b. However, these systems are overly complex, burdensome, and inconsistent because the communication stack 106a and 106b is required to perform buffering, keep track of how much data was sent and received from each endpoint, and keep track of what data was lost. Since the communication stack 106a and 106b cannot determine what data is actually important to a particular application, the communication stack 106a and 106b must have the functionality to track all data, and store all network information regarding the state of the connection.

[0014] To avoid such complexity, some prior art distributed systems shift the burden of fault tolerant communication entirely to the applications. Unfortunately, different applications can implement fault tolerant schemes in different ways, causing inconsistency and requiring greater care in developing distributed software for existing distributed applications. Moreover, different applications will require similar functions because they must be able to handle the same communication scenario. This causes waste, in terms of duplication effort, and additional implementation inconsistencies.

[0015] In view of the foregoing, there is a need for a method for supporting fault tolerant and adaptive communication that promotes consistency and reduced complexity. The method should allow reestablished connections to restart from the point where interruption occurred, and should be independent of the communication protocol, the format of the data transmitted, and any application policies.

SUMMARY OF THE INVENTION

[0016] Broadly speaking, the present invention fills these needs by providing synchronous support of fault-tolerant and adaptive communication. Broadly speaking, embodiments of the present invention utilize control layers, which provide fault tolerant connectivity to applications. The control layers at each endpoint exchange connection state information with each other that allows connections to restart from the point where an interruption occurred. In one embodiment, a method is disclosed for providing fault-tolerant and adaptive communication in a distributed computing environment. The method includes detecting an interruption in a digital connection on a network between two endpoints. The digital connection is utilized to transmit data elements of a transaction sequentially from an opposite endpoint to an application. Upon detection, a connection state value indicating a number of data elements processed by the application is obtained. The connection state value is then exchanged with the opposite endpoint over the network. In this manner, remaining data elements of the transaction can be received sequentially over the network from the opposite endpoint. In particular, the remaining data elements begin with a data element at a point indicated by the connection state value plus one. The control layer communication can be in both directions. For example, the opposite connection state value indicating a number of data elements processed using the opposite endpoint can be received so that remaining data elements can be transmitted sequentially over the network to the opposite endpoint. As above, the remaining data elements can begin with a data element at a point indicated by the opposite connection state value plus one.

[0017] A computer program embodied on a computer readable medium for providing fault-tolerant and adaptive communication in a distributed computing environment is disclosed in an additional embodiment of the present invention. The computer program includes a connection monitor module that detects interruptions in digital connections on a network between two endpoints. As above, the digital connection is utilized to transmit data elements of a transaction sequentially from an opposite endpoint to an application. The computer program also includes an application programming interface that obtains a connection state value indicating a number of data elements processed by the application. Further included is a reconnect protocol module that exchanges the connection state value with the opposite endpoint over the network. In this manner, the computer program allows remaining data elements of the transaction to be received sequentially over the network from the opposite endpoint such that the remaining data elements begin with a data element at a point indicated by the connection state value plus one. The computer program can also be utilized to facilitate changing communication channels during a digital transaction. For example, the connection monitor module can detect a 2nd communication channel that is currently more efficient than a current 1st communication channel. In this case, the connection monitor module can interrupt the digital connection on the current 1st communication channel and re-map a logical connection to the 2nd communication channel.

[0018] In a further embodiment, a distributed environment having fault-tolerant and adaptive communication is disclosed. The distributed environment includes a server application in communication with a server control layer. In addition, a client application is included that is in communication with the server application over a logical connection. The logical connection is utilized to transmit data elements of a transaction sequentially from the server application to the client application. Also included is a client control layer having a control connection with the server control layer. The client control layer exchanges a client connection state value with the server control layer using the control connection when the logical connection is interrupted. As above, the client connection state value indicates a number of data elements processed by the client application. In this manner, the server application can transmit remaining data elements of the transaction sequentially over the network beginning with a data element at a point indicated by the client connection state value plus one. Other aspects and advantages of the invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, illustrating by way of example the principles of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The invention, together with further advantages thereof, may best be understood by reference to the following description taken in conjunction with the accompanying drawings in which:

[0020] FIG. 1 is a block diagram showing a typical prior art distributed environment;

[0021] FIG. 2 is a block diagram showing a distributed computing environment having fault-tolerance support for a one-way transaction, in accordance with an embodiment of the present invention;

[0022] FIG. 3 is a block diagram showing a distributed computing environment having fault-tolerance support for a two-way transaction, in accordance with an embodiment of the present invention;

[0023] FIG. 4 is a block diagram showing a distributed computing environment having fault-tolerance support and communication channel switching capability, in accordance with an embodiment of the present invention;

[0024] FIG. 5 is a block diagram illustrating the interfaces of a control layer, in accordance with an embodiment of the present invention;

[0025] FIG. 6 is a flowchart showing a method for providing fault-tolerant and adaptive communication in a distributed computing environment, in accordance with an embodiment of the present invention;

[0026] FIG. 7 is a flowchart showing a method for providing fault-tolerant and adaptive communication when switching communication channels in a distributed computing environment, in accordance with an embodiment of the present invention; and

[0027] FIG. 8 is a flowchart showing method for reestablishing a connection using a control layer, in accordance with an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] An invention is disclosed for synchronous support of fault-tolerant and adaptive communication. Embodiments of the present invention provide protocol and data format independent fault-tolerant communication to applications with minimal support from applications. Broadly speaking, embodiments of the present invention insert control layers in communication with the communication stack, which provide fault tolerant connectivity to applications. The control layers at each endpoint exchange the connection state information with each other that allows connections to restart from the point where an interruption occurred. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced without some or all of these specific details. In other instances, well known process steps have not been described in detail in order not to unnecessarily obscure the present invention.

[0029] FIG. 1 was described in terms of the prior art. FIG. 2 is a block diagram showing a distributed computing environment 200 having fault-tolerance support for a one-way transaction, in accordance with an embodiment of the present invention. The distributed environment 200 includes a client application 202 executing on a client device, which is in communication with a server application 204 executing on a server device. As discussed previously, the client application 202 and server application 204 generally communicate using a logical connection 112, which is a logical entity used by applications to exchange data between two endpoints, such as the client application 202 and the server application 204. Although the application programs 202 and 204 function as though the logical connection 112 were a physical entity, the logical connection 112 requires a communication channel to actually transmit data.

[0030] The communication stacks 106a and 106b map logical connections to communication channels, which include the actual communication hardware 108a and 108b. As disclosed above, the communication stacks 106a and 106b handle data routing, flow control, buffering, error correction, and other computing issues encountered in real-world communication. Entities, such as the client application 202 and the server application 204, use one or more logical connections 112 to communicate with other entities by sending and receiving data in a sequential fashion over a period of time.

[0031] Embodiments of the present invention provide synchronous support of fault-tolerant and adaptive communication using control layers 206 and 208, which are in communication with the applications 202 and 204 and the communication stacks 106a and 106b. In one embodiment, a client control layer 206 is placed in communication with the client application 202 and the client side communication stack 106a, and a server control layer 208 is placed in communication with the server application 204 and the server side communication stack 106b. In operation, the control layers 206 and 208 facilitate reconnection and transaction restarting from the point of interruption.

[0032] Broadly speaking, the control layers 206 and 208 obtain connection state values from the respective applications 202/204 upon detecting an interruption in a digital connection. The connection state values indicate the number of data elements processed by each application prior to the connection interruption. The data elements can be any discrete grouping of data, such as a byte of data, a group of data bytes, an object, a group of objects, or any other discrete data grouping utilized by an application. The control layers 206 and 208 then exchange the connection state values with each other. In this manner, remaining data elements of the transaction can be sent sequentially over the network beginning with the data element at the point indicated by the connection state values plus one.

[0033] More particularly, during normal operation the client application 202 and server application 204 communicate with each other by sending data elements 114a-114d to each other using the logical connection 112. During this time the control layers 206 and 208 monitor the state of the connection. Upon detecting an interruption in the connection, the control layers 206 and 208 notify the applications 202 and 204 that an interruption has occurred. Once notified of the interruption, the applications can make appropriate responses. In one embodiment, the client application 202 sends a request to the control layer 206 requesting reconnection. In addition, the client application 202 provides the control layer 206 with the number of data elements the client application 202 has processed up to the point of interruption.

[0034] For example, in FIG. 2, the server application 204 is transmitting data elements 1114a, 2114b, 3114c, and 4114d to the client application 202 using the logical connection 112. During the transmission an interruption occurs in the connection such that data can no longer be communicated over the logical connection 112. Upon detection of the connection interruption, the control layer 206 notifies the client application 202 of the interruption. In response, the client application 202 transmits the connection state value to the control layer 206 and requests a reconnection.

[0035] At this point, the client control layer 206 and the server control layer 208 establish a control connection 210 that allows the control layers 206 and 208 to communicate with each other. In one embodiment, the control layers 206 and 208 utilize the same communication stacks 106a-106b and communication hardware 108a-108b for digital communication as utilized by the applications 202 and 204. In additional embodiments, the control connection 210 is established over a different physical channel that can, in one embodiment, be specifically utilized for control communication only. Once the control connection 210 is established, the control layers exchange connection state values. For example, in FIG. 2, the client control layer 206 transmits the connection state value for the client application 202 to the server control layer 208.

[0036] In the example of FIG. 2, the client application 202 processed data elements 1114a and 2114b prior to the connection interruption. Hence, the client application 202 did not receive data elements 3114c and 4114d of the current transaction. As a result, upon being notified of the connection interruption, the client application 202 transmits a connection state value of two, indicating two data elements were processed prior to interruption, to the client control layer 206. Upon establishing the control connection 210, the client control layer 206 transmits the connection state value, in this example having a value of two, to the server control layer 208. The server control layer 208 then provides the connection state value information to the server application 204. At this point, the client control layer 206 passes control back to the client application 202 and the server control layer 208 passes control back to the server application 204.

[0037] Thereafter, the server application 204 can begin transmitting the remaining data elements of the transaction to the client application 202. In particular, the server application 204 begins transmitting from the data element at a point indicated by the connection state value plus one. In the example of FIG. 2, the server application 204 begins retransmitting data starting from the third data element in the transaction (connection state value of two plus one), which is data element 3114c, and continues with the remaining data element 4114d. Although the example of FIG. 2 illustrates a one-way transaction, embodiments of the present invention also support two-way transactions were data is being transmitted to both endpoints, as illustrated next with reference to FIG. 3.

[0038] FIG. 3 is a block diagram showing a distributed computing environment 250 having fault-tolerance support for a two-way transaction, in accordance with an embodiment of the present invention. The distributed environment 250 includes a client application 202 executing on a client device, which is in communication with a server application 204 executing on a server device. As discussed previously, the client application 202 and server application 204 generally communicate using the logical connection 112.

[0039] During normal operation the client application 202 and server application 204 communicate with each other by sending data elements 114a-114d and 116a-116d to each other using the logical connection 112. As discussed above, the control layers 206 and 208 monitor the state of the connection during this communication. Upon detecting an interruption in the connection, the control layers 206 and 208 notify the applications 202 and 204 that an interruption has occurred. In one embodiment, the client application 202 sends a request to the control layer 206 requesting reconnection in response to being notified of the interruption. The client application 202 also provides the control layer 206 with the number of data elements the client application 202 has processed up to the point of interruption. In addition, the server control layer 208 is provided with the number of data elements the server application 204 has processed up to the point of interruption. It should be noted that the server control layer 208 can obtain the number of data elements processed in either a synchronous manner as discussed above, or in an asynchronous manner, wherein the server application periodically updates the server control layer 208.

[0040] For example, in FIG. 3, the server application 204 is transmitting data elements 1114a, 2114b, 3114c, and 4114d to the client application 202 using the logical connection 112. In addition, the client application 202 is transmitting data elements A 116a, B 116b, and C 116c to the server application 204 also using the logical connection 112. During the transmission an interruption occurs in the connection such that data can no longer be communicated over the logical connection 112. Upon detection of the connection interruption, the control layers 206 and 208 detect the connection interruption and notify the client application 202 and server application 204. In response, the client application 202 transmits the connection state value to the control layer 206 and requests a reconnection.

[0041] At this point, the client control layer 206 and the server control layer 208 establish a control connection 210 that allows the control layers 206 and 208 to communicate with each other. The control layers 206 and 208 utilize the same communication stacks 106a-106b and communication hardware 108a-108b for digital communication as utilized by the applications 202 and 204. Once the control connection 210 is established, the control layers exchange connection state values. For example, in FIG. 3, the client control layer 206 transmits the connection state value for the client application 202 to the server control layer 208. As discussed above, the server control layer 208 can obtain the connection state value from the server application 204 in either a synchronous manner or asynchronous manner wherein the server application periodically updates the server control layer 208.

[0042] In the example of FIG. 3, the client application 202 processed data elements 1114a and 2114b prior to the connection interruption. Hence, the client application 202 did not receive data elements 3114c and 4114d of the current transaction. As a result, upon being notified of the connection interruption, the client application 202 transmits a connection state value of two, indicating two data elements were processed prior to interruption, to the client control layer 206. In addition, the server application 204 only processed data element A 116a prior to the connection interruption. Hence, the server application 204 did not receive data elements B 116b and C 116c of the current transaction. As a result, the server control layer 208 obtains a connection state value of one, either synchronously or asynchronously, from the server application 204 indicating one data element was processed prior to interruption.

[0043] Upon establishing the control connection 210, the client control layer 206 transmits its connection state value, in this example having a value of two, to the server control layer 208. Also, the server control layer 208 transmits its connection state value, in this example having a value of one, to the client control layer 206. The client control layer 206 then provides the connection state value information received from the server control layer 208 to the client application 202. Similarly, the server control layer 208 provides the connection state value information received from the client control layer 206 to the server application 204. At this point, the client control layer 206 passes control back to the client application 202 and the server control layer 208 passes control back to the server application 204.

[0044] Thereafter, the server application 204 can begin transmitting the remaining data elements of the transaction to the client application 202. In particular, the server application 204 begins transmitting from the data element at a point indicated by the connection state value received from the client control layer 206 plus one. In the example of FIG. 3, the server application 204 begins re-transmitting data starting from the third data element in the transaction (connection state value of two plus one), which is data element 3114c, and continues with the remaining data element 4114d. The client application 202 also begins transmitting the remaining data elements of the transaction to the server application 204. The client application 202 begins transmitting from the data element at a point indicated by the connection state value received from the server control layer 208 plus one. In the example of FIG. 3, the client application 202 begins re-transmitting data starting from the second data element in the transaction (connection state value of one plus one), which is data element B 116b, and continues with the remaining data element C 116c . In addition to handling unintentional connection interruptions, embodiments of the present invention can handle intentional connection interruptions, such as when a communication channel is changed.

[0045] FIG. 4 is a block diagram showing a distributed computing environment 400 having fault-tolerance support and communication channel switching capability, in accordance with an embodiment of the present invention. The distributed environment 400 includes a client application 202 executing on a client device, which is in communication with a server application 204 executing on a server device. As discussed previously, the client application 202 and server application 204 communicate using a logical connection 112.

[0046] As mentioned previously, the communication stacks 106a and 106b map logical connections to communication channels 110a and 110b, which include the actual communication hardware 108a and 108b. The communication stacks 106a and 106b handle data routing, flow control, buffering, error correction, and other computing issues encountered in real-world communication. Entities, such as the client application 202 and the server application 204, use one or more logical connections 112 to communicate with other entities by sending and receiving data in a sequential fashion over a period of time.

[0047] However, different communication channels 110a and 110b can perform differently. For example, although a cell phone based wireless communication channel is widely available, a cell phone based wireless communication channel generally performs much slower than a local 802.11b wireless connection. As such, network software can be capable of switching to faster communication channels when they become available. However, switching to a new communication channel interrupts any currently established connection using the previous communication channel. Advantageously, embodiments of the present invention can facilitate communication channel switching, allowing data transactions to continue from the point of interruption before the communication channels were changed.

[0048] For example, in FIG. 4 the client application 202 and server application 204 initiate a data transaction using a first communication 110a. That is, the communication stacks 106a and 106b map the logical connection 112 to communication channel 110a, which can be, for example, a cell phone based wireless network. During communication, the client device, for example, can enter an area having a better communication channel available. For example, if the client device were a telematics device incorporated into an automobile, the first communication channel 110a could be a cell phone based wireless network. Subsequently, the automobile can enter, for example, a home garage offering 802.11b wireless network connectivity using a second communication channel 110b. In this case, the communication stacks can remap the connection to the second communication channel 110b.

[0049] As mentioned above, the remapping of the connection causes an interruption in the logical connection 112. However, the control layers 206 and 208 monitor the state of the connection. Thus, upon detecting an interruption in the connection, the control layers 206 and 208 notify the applications 202 and 204 that an interruption has occurred and the applications can make appropriate responses. That is, the client application 202 sends a request to the control layer 206 requesting reconnection and provides the control layer 206 with the number of data elements the client application 202 has processed up to the point of interruption.

[0050] For example, in FIG. 4, the server application 204 is transmitting data elements 1114a, 2114b, 3114c, and 4114d to the client application 202 using the logical connection 112. Then, the connection is interrupted to change from communication channel 110a to communication channel 110b. Upon detection of the connection interruption, the control layer 206 notifies the client application 202, and the client application 202 transmits the connection state value to the control layer 206 and requests a reconnection.

[0051] When the communication channel change is complete the client control layer 206 and the server control layer 208 establish a control connection 210 that allows the control layers 206 and 208 to communicate with each other. At this point, the control layers exchange connection state values, for example in FIG. 4, the client application 202 processed data elements 1114a and 2114b prior to the connection interruption caused by the communication channel change. Hence, the client application 202 did not receive data elements 3114c and 4114d of the current transaction. As a result, upon being notified of the connection interruption, the client application 202 transmits a connection state value of two, indicating two data elements were processed prior to interruption, to the client control layer 206. Upon establishing the control connection 210, the client control layer 206 transmits the connection state value, in this example having a value of two, to the server control layer 208. The server control layer 208 then provides the connection state value information to the server application 204. At this point, the client control layer 206 passes control back to the client application 202 and the server control layer 208 passes control back to the server application 204.

[0052] As above, the server application 204 can begin transmitting the remaining data elements of the transaction to the client application 202 beginning from the data element at a point indicated by the connection state value plus one. In the example of FIG. 4, the server application 204 begins re-transmitting data starting from the third data element in the transaction (connection state value of two plus one), which is data element 3114c, and continues with the remaining data element 4114d.

[0053] FIG. 5 is a block diagram illustrating the interfaces of a control layer, in accordance with an embodiment of the present invention. As shown in FIG. 5, the control layer 206 communicates with the application 202 using an application programming interface (API) 502. The control layer 206 communicates with the communication stack using a connection monitor 504 and a reconnect protocol 506.

[0054] The connection monitor 504 monitors the status of the connection and detects any interruption of communication on the connection. As mentioned above, the connection can be interrupted for many different reasons, including signal drops or fades, and communication channel changing. When the connection monitor 504 detects an interruption in the connection, the connection monitor 504 notifies the application 202 of the interruption using the API 502. If the application 202 was transmitting data to another endpoint, the application 202 stops transmitting data. Also, if the application was receiving data from another endpoint, it stops receiving data and stores its connection state value, which is the number of data elements of the current data transaction processed to this point. The application 202 then indicates it is ready to reestablish communication over the connection and provides the connection state value to the control layer 206 using the API 502. The application 202 then waits until the API 502 indicates the connection has been reestablished.

[0055] The connection monitor 504 then monitors the connection to detect when the connection has been reestablished. Once the connection is reestablished, the reconnect protocol 506 transmits information to the opposite control layer indicating that a ‘reconnect’ is about to occur. In addition, the reconnect protocol 506 transmits the connection state value to the opposite control layer. The opposite control layer can respond with a connection state value of an application at the opposite endpoint.

[0056] The API 502 then notifies the application 202 that the connection has been reestablished. In addition, if the opposite control layer responded with a connection state value of the endpoint, API 502 provides opposite endpoint's connection state value to the application 202. At this point, the application 202 can continue to perform its work. That is, if the application 202 was sending data over the connection, the application can continue sending data elements to the opposite endpoint starting from the point indicated by the connection state value received from the opposite endpoint. If the application 202 was receiving data over the connection, the application 202 will begin receiving data from the opposite endpoint starting with the data element immediately following the last data that was processed before the connection was interrupted.

[0057] FIG. 6 is a flowchart showing a method 600 for providing fault-tolerant and adaptive communication in a distributed computing environment, in accordance with an embodiment of the present invention. In an initial operation 602, preprocess operations are performed. Preprocess operations can include mapping a physical communication channel to a logical connection, executing a client application and a server application, and other preprocess operations that will be apparent to those skilled in the art after a careful reading of the present disclosure.

[0058] In operation 604, a logical connection is established between applications on opposite endpoints of a network. A typical distributed environment can include, for example, a client application executing on a client device, which is in communication with a server application executing on a server device. The client application and server application generally communicate using a logical connection, which is a logical entity used by applications to exchange data between two endpoints, such as the client application and the server application. Although the application programs function as though the logical connection were a physical entity, the logical connection requires a communication channel to actually transmit data, as discussed in greater detail previously.

[0059] The applications then begin transmitting data elements to each other, in operation 606. Applications generally communicate over a network by sending a sequential stream of data elements to the other application. As mentioned above, the data elements can be any discrete grouping of data, such as a byte of data, a group of data bytes, an object, a group of objects, or any other discrete data grouping utilized by an application.

[0060] A decision is then made as to whether the logical connection has been interrupted, in operation 608. A connection monitor monitors the status of the connection and detects any interruption of communication on the connection. As mentioned above, the connection can be interrupted for many different reasons, including signal drops or fades, and communication channel changing. If the logical connection has been interrupted, the method branches to operation 610. Otherwise, the method 600 continues with another transmit data elements operation 606.

[0061] In operation 610, the connection is reestablished using the control layer. Broadly speaking, the control layers obtain connection state values, indicating the number of data elements processed each application, from the respective applications upon detecting an interruption in a digital connection. The control layers then exchange the connection state values with each other. In this manner, remaining data elements of the transaction can be sent sequentially over the network from the opposite endpoint beginning with the data element at the point indicated by the connection state values plus one. Connection reestablishment will be described in greater detail with reference to FIG. 8 below.

[0062] Advantageously, using the embodiments of the present invention, applications at both endpoints of a connection are able to reestablish communication at the precise point where the interruption occurred. In addition, the applications only need to resend the minimal amount of data required, regardless of the underlying communication mechanism and protocols.

[0063] FIG. 7 is a flowchart showing a method 700 for providing fault-tolerant and adaptive communication when switching communication channels in a distributed computing environment, in accordance with an embodiment of the present invention. In an initial operation 702, preprocess operations are performed. Preprocess operations can include mapping a physical communication channel to a first logical connection, executing a client application and a server application, and other preprocess operations that will be apparent to those skilled in the art after a careful reading of the present disclosure.

[0064] In operation 704, a logical connection is established between applications on opposite endpoints of a network using a first communication channel. As mentioned previously, the communication stacks map logical connections to communication channels. The communication stacks handle data routing, flow control, buffering, error correction, and other computing issues encountered in real-world communication. Entities, such as the client application and the server application, use one or more logical connections to communicate with other entities by sending and receiving data in a sequential fashion over a period of time.

[0065] The applications then begin transmitting data elements to each other, in operation 706. Applications generally communicate over a network by sending a sequential stream of data elements to the other application. As mentioned above, the data elements can be any discrete grouping of data, such as a byte of data, a group of data bytes, an object, a group of objects, or any other discrete data grouping utilized by an application.

[0066] A decision is then made as to whether a better communication channel is available, in operation 708. Different communication channels can perform differently. For example, although cell phone based wireless communication channel are widely available, a cell phone based wireless communication channel generally performs much slower than a local 802.11b wireless connection. As such, network software can be capable of switching to faster communication channels when they become available. As will be discussed subsequently, from the application's point of view the logical connection appears to be interrupted when switching to a new communication channel even though the connection or underlying channel remain open and uninterrupted from the point of view of the control layer. Advantageously, embodiments of the present invention can facilitate communication channel switching, allowing data transactions to continue from the point of interruption before the communication channels were changed. If a better, second communication channel is available, the method 700 branches to operation 710. Otherwise, the method 700 continues to transmit data elements in operation 706.

[0067] In operation 710, the logical connection on the first communication channel is interrupted. During communication, the client device can enter an area having a better communication channel available. For example, if the client device is a telematics device incorporated into an automobile, the first communication channel could be a cell phone based wireless network. Subsequently, the automobile can enter, for example, a home garage offering 802.11b wireless network connectivity using a second communication channel. In this case, the logical connection is interrupted on the first communication channel to allow the connection to be changed to the second communication channel, as described next.

[0068] The logical connection is re-mapped to the better, second communication channel, in operation 712, and the connection is reestablished using the control layer in operation 610. As mentioned above, the control layers obtain connection state values, indicating the number of data elements processed each application, from the respective applications. The control layers then exchange the connection state values with each other once the connection is re-mapped to the second communication channel. In this manner, remaining data elements of the transaction can be sent sequentially over the network from the opposite endpoint beginning with the data element at the point indicated by the connection state values plus one.

[0069] FIG. 8 is a flowchart showing method 610 for reestablishing a connection using a control layer, in accordance with an embodiment of the present invention. In an initial operation 800, preprocess operations are performed. Preprocess operations can include detecting a connection interruption, re-mapping a logical connection to another communication channel, and other preprocess operations that will be apparent to those skilled in the art after a careful reading of the present disclosure.

[0070] In operation 802, the application is notified that connectivity is interrupted. Embodiments of the present invention utilize a connection monitor to monitor the status of the connection and detects any interruption of communication on the connection. When the connection monitor detects an interruption in the connection, the connection monitor notifies the application of the interruption using the API. If the application was transmitting data to another endpoint, the application stops transmitting data. Also, if the application was receiving data from another endpoint, it stops receiving data and stores its connection state value, as described in operation 804.

[0071] A connection state value and a request to reconnect are received, in operation 804. The connection state value indicates the number of data elements processed by the application at the time the interruption occurred in the digital connection. As mentioned previously, the data elements can be any discrete grouping of data, such as a byte of data, a group of data bytes, an object, a group of objects, or any other discrete data grouping utilized by an application.

[0072] In operation 806, connectivity is reestablished with the opposite control layer. In particular, the two client control layers establish a control connection that allows the control layers to communicate with each other. The control layers utilize the same communication stacks and communication hardware for digital communication as utilized by the applications.

[0073] Once the control connection is established, the control layers exchange connection state values, in operation 808. As mentioned above, connection state values are values indicating the number of data elements processed by each application up to the point of interruption in a digital connection. The data elements, as previously discussed, can be any discrete grouping of data, such as a byte of data, a group of data bytes, an object, a group of objects, or any other discrete data grouping utilized by an application.

[0074] Thereafter, the opposite control layer can begin transmitting data elements, beginning at the received connection state value plus one, in operation 810. Also, control is passed back to the application layers, in operation 812, and post process operations are performed in operation 814. Post process operations can include, for example, transmitting additional data elements, changing communication channels, and other post process operations that will be apparent to those skilled in the art after a careful reading of the present disclosure. Thus, embodiments of the present invention advantageously allow applications at both endpoints of a connection to reestablish communication at the precise point where the interruption occurred, regardless of the underlying communication mechanism and protocols. In addition, the applications only need to resend the minimal amount of data required.

[0075] Although the foregoing invention has been described in some detail for purposes of clarity of understanding, it will be apparent that certain changes and modifications may be practiced within the scope of the appended claims. Accordingly, the present embodiments are to be considered as illustrative and not restrictive, and the invention is not to be limited to the details given herein, but may be modified within the scope and equivalents of the appended claims.

Claims

1. A method for providing fault-tolerant and adaptive communication in a distributed computing environment, comprising: detecting an interruption in a digital connection on a network between two endpoints, the digital connection being utilized to transmit data elements of a transaction sequentially from an opposite endpoint to an application; obtaining a connection state value indicating a number of data elements processed by the application; exchanging the connection state value with the opposite endpoint over the network; and receiving remaining data elements of the transaction sequentially over the network from the opposite endpoint, wherein the remaining data elements begin with a data element at a point indicated by the connection state value plus one.

2. A method as recited in claim 1, further comprising the operations of: receiving an opposite connection state value indicating a number of data elements processed using the opposite endpoint; and transmitting remaining data elements sequentially over the network to the opposite endpoint, wherein the remaining data elements begin with a data element at a point indicated by the opposite connection state value plus one.

3. A method as recited in claim 1, further comprising the operation of notifying the application that an interruption has occurred.

4. A method as recited in claim 1, further comprising the operation of receiving a request to reconnect from the application, wherein the application provides the connection state value.

5. A method as recited in claim 4, further comprising the operation of passing control back to the application after exchanging state information with the opposite endpoint.

6. A method as recited in claim 1, wherein an endpoint is a client application and the opposite endpoint is a server application.

7. A method as recited in claim 1, further comprising the operation of detecting a 2nd communication channel that is currently more efficient than a current 1st communication channel.

8. A method as recited in claim 7, further comprising the operation of re-mapping a logical connection to the 2nd communication channel.

9. A computer program embodied on a computer readable medium for providing fault-tolerant and adaptive communication in a distributed computing environment, comprising: a connection monitor module that detects an interruption in a digital connection on a network between two endpoints, the digital connection being utilized to transmit data elements of a transaction sequentially from an opposite endpoint to an application; an application programming interface that obtains a connection state value indicating a number of data elements processed by the application; and a reconnect protocol module that exchanges the connection state value with the opposite endpoint over the network, wherein remaining data elements of the transaction are received sequentially over the network from the opposite endpoint, and wherein the remaining data elements begin with a data element at a point indicated by the connection state value plus one.

10. A computer program as recited in claim 9, wherein the reconnect protocol module further receives an opposite connection state value indicating a number of data elements processed using the opposite endpoint.

11. A computer program as recited in claim 10, wherein the application transmits remaining data elements sequentially over the network to the opposite endpoint, and wherein the remaining data elements begin with a data element at a point indicated by the opposite connection state value plus one.

12. A computer program as recited in claim 9, wherein the application programming interface notifies the application that an interruption has occurred.

13. A computer program as recited in claim 9, wherein the application programming receives a request to reconnect from the application, and wherein the application provides the connection state value.

14. A computer program as recited in claim 9, wherein an endpoint is a client application and the opposite endpoint is a server application.

15. A computer program as recited in claim 9, wherein the connection monitor module detects a 2nd communication channel that is currently more efficient than a current 1st communication channel.

16. A computer program as recited in claim 15, wherein the connection monitor module interrupts the digital connection on the current 1st communication channel and re-maps a logical connection to the 2nd communication channel.

17. A distributed environment having fault-tolerant and adaptive communication, comprising: a server application in communication with a server control layer; a client application having logical connection with the server application, the logical connection being utilized to transmit data elements of a transaction sequentially from the server application to the client application; and a client control layer having a control connection with the server control layer, wherein the client control layer exchanges a client connection state value with the server control layer using the control connection when the logical connection is interrupted, the client connection state value indicating a number of data elements processed by the client application.

18. A distributed environment as recited in claim 17, wherein the server application transmits remaining data elements of the transaction sequentially over the network beginning with a data element at a point indicated by the client connection state value plus one.

19. A distributed environment as recited in claim 18, wherein the server control layer exchanges a server connection state value with the client control layer indicating a number of data elements processed using the server application.

20. A distributed environment as recited in claim 19, wherein the client application transmits remaining data elements sequentially over the network to the server application beginning with a data element at a point indicated by the server connection state value plus one.