Call Management Method and Apparatus

Abstract

A system and method for handling calls at a call facility wherein each call is handled by a distinct call container instance, the system comprising a platform server programmed to perform the steps of maintaining an idle container inventory for handling new incoming calls, upon receiving a new call, assigning the new call to one of the idle containers in the inventory, upon completion of a call within one of the containers, causing the container that completed the call to be killed, monitoring the idle container inventory and stage instantiation of additional idle containers to replenish the idle container inventory as containers complete calls and are killed off.

Claims

1. A system for handling calls at a call facility wherein each call is handled by a distinct call container instance, the system comprising: a platform server programmed to perform the steps of: maintaining an idle container inventory for handling new incoming calls; upon receiving a new call, assigning the new call to one of the idle containers in the inventory; upon completion of a call within one of the containers, causing the container that completed the call to be killed; monitoring the idle container inventory; and stage instantiation of additional idle containers to replenish the idle container inventory as containers complete calls and are killed off.

2. The system of claim 1 wherein the platform server is further programmed to operate a state monitoring container where the state monitoring container monitors the idle container inventory and stages instantiation of additional idle containers to replenish the inventory.

3. The system of claim 2 wherein the state monitoring container also manages killing of containers that complete calls.

4. The system of claim 3 wherein the state monitoring container is further programmed to maintain a minimum number of idle containers in the idle container inventory.

5. The system of claim 4 wherein the minimum number of idle containers is a set number of idle containers.

6. The system of claim 4 wherein the state monitoring container tracks the number of active containers handling ongoing calls and wherein the minimum number of idle containers is dynamic and is based at least in part on the number of active containers.

7. The system of claim 6 wherein the number of idle containers maintained is a percentage of active containers.

8. The system of claim 4 wherein the state monitoring container is further programmed to kill off containers that complete calls.

9. The system of claim 4 wherein the container that handles a call is programmed to report out to the state monitoring container when the container becomes active to handle a call.

10. The system of claim 4 wherein each container that handles a call performs an application program to handle the call and wherein the application program includes a container kill code segment to automatically kill the container at the end of the application program.

11. The system of claim 4 wherein the minimum number of idle containers maintained in the idle container inventory depends at least in part on historical call volume at different times of the day.

12. The system of claim 2 wherein the state monitoring container is a first state monitoring container, the system further including a second state monitoring container that performs identical redundant functions simultaneously with the first state monitoring container.

13. The system of claim 1 wherein the step of assigning a new call to one of the idle containers in inventory includes, upon receiving a new call, using the called number to identify a client call routing table, using the call routing table to identify an IT application associated with the called number, activating one of the idle containers and downloading the identified IT application to the activated container.

14. The system of claim 1 wherein the state monitoring container instantiates additional idle containers in groups.

15. The system of claim 2 wherein the state monitoring container monitors active containers to identify containers that have ceased working and performs a process to kill those containers.

16. The system of claim 15 wherein the process to kill a container that has ceased working includes obtaining state information from the container that has ceased working, activating a new container from the idle container inventory, downloading an IT application and the state information to the newly active container and killing the container that ceased working.

17. A system for handling calls at a call facility wherein each call is handled by a distinct call container instance, the system comprising: a platform server programmed to perform the steps of: instantiating a state monitoring container that performs the steps of: (i) maintaining an idle container inventory for handling new incoming calls; (ii) upon receiving a new call: (a) using the called number to identify an application program to be used to handle the new call; (b) activating one of the idle containers; (c) downloading the identified application program to the activated container; (d) monitoring the container for completion of the call; (e) upon completion of the call, causing the container that handled the call to be killed; (iii) monitoring the idle container inventory; and (iv) staging instantiation of additional idle containers to replenish the idle container inventory as containers complete calls and are killed off.

18. The system of claim 17 wherein each of the application programs includes code for indicating completion of a call and wherein the step of monitoring containers for call completion includes receiving indications of call completion from containers.

19. A method for handling calls at a call facility wherein each call is handled by a distinct call container instance, the method comprising the steps of: using a platform server to perform the steps of: maintaining an idle container inventory for handling new incoming calls; upon receiving a new call, assigning the new call to one of the idle containers in the inventory; upon completion of a call within one of the containers, causing the container that completed the call to be killed; monitoring the idle container inventory; and stage instantiation of additional idle containers to replenish the idle container inventory as containers complete calls and are killed off.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0021] FIG. 1 is a schematic diagram illustrating exemplary components in a telephone system used to implement various methods and processes that are consistent with at least some aspects of the present disclosure;

[0022] FIG. 2 is a schematic diagram illustrating an exemplary client call routing table;

[0023] FIG. 3 is a schematic diagram illustrating an exemplary client application table;

[0024] FIG. 4 is a schematic diagram illustrating functions and processes performed by the platform server shown in FIG. 1;

[0025] FIG. 5 is a flow chart illustrating a process for developing execution code corresponding to an interactive telephony application that may be facilitated by the platform server shown in FIG. 1;

[0026] FIG. 6 is a simple flow chart illustrating a process to be implemented via an interactive telephony application;

[0027] FIG. 7 is a schematic showing a greeting screen shot that may be presented to an application developer upon logging on to an application development web site provided by the platform server of FIG. 1;

[0028] FIG. 8 is a schematic diagram illustrating another screen shot, albeit showing various tools to be used when developing an application;

[0029] FIG. 9 is similar to FIG. 8, albeit showing a development process at a later stage;

[0030] FIG. 10 is similar to FIG. 8, albeit showing a development process at a later stage;

[0031] FIG. 11 shows executable code that may be generated by a platform server of FIG. 1 when the visual flow shown in FIG. 10 is published;

[0032] FIG. 12 is a flow chart showing a process for handling a call from a customer to a telephony client that maybe performed by the platform server of FIG. 1 that is consistent with at least some aspects of the present disclosure;

[0033] FIG. 13 is similar to FIG. 8, albeit showing a different visual flow being developed using the application development tool;

[0034] FIGS. 14a-14e are similar to FIG. 11, albeit showing executable code for a far more complex interactive telephony application;

[0035] FIG. 15 is a screen shot showing a scripting interface that includes a run application icon for testing an application during development;

[0036] FIG. 16 is similar to FIG. 15, albeit showing a screen shot including a sub-window for entering information required when testing an application during development;

[0037] FIG. 17 is a flow chart illustrating a method of maintaining an idle container inventory that is consistent with at least some aspects of the present disclosure;

[0038] FIG. 18 is a schematic of a system that is consistent with at least some aspects of the present disclosure;

[0039] FIG. 19 is a schematic illustrating an exemplary call routing table that may be used to implement some aspects of the present disclosure;

[0040] FIG. 20 is a schematic illustrating an exemplary container database that is consistent with at least some aspects of the present disclosure; and

[0041] FIG. 21 is a sub-process that may be substituted for a portion of the process illustrated in FIG. 5.

DETAILED DESCRIPTION OF THE DISCLOSURE

[0042] The various aspects of the subject disclosure are now described with reference to the drawings, wherein like reference numerals correspond to similar elements throughout the several views. It should be understood, however, that the drawings and detailed description hereafter relating thereto are not intended to limit the claimed subject matter to the particular form disclosed. Rather, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the claimed subject matter.

[0043] As used herein, the terms “component,” “system” and the like are intended to refer to a computer-related entity, either hardware, a combination of hardware and software, software, or software in execution. For example, a component may be, but is not limited to being, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, and/or a computer. By way of illustration, both an application running on a computer and the computer can be a component. One or more components may reside within a process and/or thread of execution and a component may be localized on one computer and/or distributed between two or more computers or processors.

[0044] The word “exemplary” is used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other aspects or designs.

[0045] Furthermore, the disclosed subject matter may be implemented as a system, method, apparatus, or article of manufacture using standard programming and/or engineering techniques to produce software, firmware, hardware, or any combination thereof to control a computer or processor based device to implement aspects detailed herein. The term “article of manufacture” (or alternatively, “computer program product”) as used herein is intended to encompass a computer program accessible from any computer-readable device, carrier, or media. For example, computer readable media can include but are not limited to magnetic storage devices (e.g., hard disk, floppy disk, magnetic strips . . . ), optical disks (e.g., compact disk (CD), digital versatile disk (DVD) . . . ), smart cards, and flash memory devices (e.g., card, stick). Additionally it should be appreciated that a carrier wave can be employed to carry computer-readable electronic data such as those used in transmitting and receiving electronic mail or in accessing a network such as the Internet or a local area network (LAN). Of course, those skilled in the art will recognize many modifications may be made to this configuration without departing from the scope or spirit of the claimed subject matter.

[0046] Referring now to the drawings wherein like reference numerals correspond to similar elements throughout the several views and, more specifically, referring to FIG. 1, the present disclosure will be described in the context of an exemplary telephony system 10. In general, system 10 includes components that are controlled by three different entities including a telephony service provider, a service provider client or telephony client (e.g., a business that uses the services of the service provider) and client customers (e.g., persons that call in to a telephony system to receive some type of service from the telephony client). The general idea is that when a client customer calls a telephony client for some reason, the call is intercepted by the service provider and is handled by an interactive telephony (IT) system operated by the service provider. The call may ultimately be routed to a telephony client employee after interaction from the customer. Here, the service provider maintains applications customized for specific telephony clients and runs one of the applications when a client customer calls the client. The system enables each telephony client to develop and edit their own customized applications with intuitive interfaces and/or using a standard and well known scripting language and a small set of extensions to the standard language that are easy to learn and understand. The LUA programming language is an exemplary scripting language that is particularly useful.

[0047] Referring still to FIG. 1, telephony system 10 includes a platform server 12, a client server 14, a platform database 18, a client database 20, customer communication devices 16, at least one client interface device 15 and several internet telephone service providers (ITSPs) or carriers 50a through 50c. Customer communication device 16 may be a standard telephone linked to a PSTN line, a wireless telephone, a cellular telephone, a computer capable of VOIP communication, etc. The exemplary device 16 includes a handset including a microphone for receiving voice input from a customer and a speaker for providing audible output to the customer. Device 16 also includes a key pad for entering numbers into the system. Where device 16 is linked to a PSTN line so that device 16 communicates via a common carrier communication protocol, the common carrier calls may be terminated by one of the carriers 50a through 50c upstream of platform server 12 so that the calls are handed over to server 12 using the session initiation protocol (SIP). Hereinafter, unless indicated otherwise, it will be assumed that all calls initiated by a customer device 16 and received at server 12 use SIP.

[0048] Client server 14 is operated by a telephony client (e.g., a service provider's customer) and may perform various processes and run various software programs required for the client to run its business. For instance, in the case of a full service automotive dealer, the software programs may manage automobile inventory, a service department, a parts department, a sales department, an accounting department, a procurement department, etc. As another instance, in the case of a bank, the software programs may manage customer saving accounts, bank loan accounts, real estate assets, legal matters, security, etc. Server 14 generates data while running programs and that data is stored in client database 20. The data in database 20 is accessed by server 14 for various purposes.

[0049] Client interface 15 can be any type of computer including a desktop, a laptop, a tablet type computing device, a smart phone, etc. The exemplary device 15 includes a keyboard for entering information and also a display screen 22. Device 15 is linked to client server 14 and can be used to run various applications required to manage the client's business. In addition to running business management applications, interface 15 may also be used by a client's employee or a third party contractor (hereinafter “a client developer”) working for the client to develop an interactive telephony application. To this end, a client developer may use interface 15 to link to platform server 12 and access a browser based application development interface via screen 22 that can be used to specify a complete IT application as will be described in detail hereafter.

[0050] Referring still to FIG. 1, while server 12 is shown as a single server, the functions performed by server 12 may be performed by a set of cooperating servers that are provided and maintained by the service provider independent of the telephony clients. For instance, server 12 may include a database server, a web server, a telephony control server, an application development server, etc., that all cooperate the provide the functions of the platform server 12 described here. Server 12 receives SIP calls on line 24. Server 12 is linked via an internet or other network link to tenant server 14.

[0051] Platform server 12 is also linked to platform database 18 which stores development programs for developing tenant IT applications, applications specified using the development programs and data generated using the development programs and data generated when running the IT applications as well as other software used to manage overall server operation. To this end, database 18 stores application development programs 30 used by tenant developers to develop customized IT applications. The application development programs include functional modules collectively identified by numeral 31 that can be used by a developer to specify pre-canned functions for an application. For instance, if a developer wants an IT application to annunciate a greeting and initial instructions when a call is received, a “say” functional module 31 may be instantiated as part of an application where text entered into the functional module after selection is converted into a voice message when that functional module occurs in the flow of the application. As another instance, if a developer wants to collect a number entered into a keypad on a customer's device 16, a “gather digit” functional module 31 may be instantiated to obtain an entered digit and provide the digit to the application program for further processing. The functional modules are designed to be intuitive and will, in at least some cases, be limited in number so that learning how to use the modules is a trivial task for most client employees.

[0052] Referring still to FIG. 1, platform database 18 also includes a routing table database 29 that includes client call routing tables 32. Routing table database 29 includes a separate call routing table for each of the service provider telephony client. Referring also to FIG. 2, an exemplary client call routing table 32 for a first client Ace Automotive is illustrated. The exemplary table 32 includes three columns of associated information including a number column 36, an application column 40 and a data column 42. The number column 36 includes a list of telephone numbers or extensions that are associated with the client. For instance, the last two numbers in column 36 correspond to direct dial telephone numbers while the top three numbers correspond to extensions. As seen, several numbers may be associated with a single client where each of the numbers corresponds to a different application. For instance, in the case of Ace Automotive, the client may have one general number to field general calls, a second number to contact a service department, a third number to contact a sales department, etc. Here, each of the different numbers may be associated with a different IT application.

[0053] Application column 40 lists a separate application for each of the numbers in column 36. For instance, an “IVR Application 1” corresponds to number “1111” in column 36, an “IVR Application 2” corresponds to number “1222” in column 36 and so on. Each of the applications is associated with a different interactive telephony process.

[0054] Data column 42 includes data that has been entered by a tenant application developer for each of the applications in column 40 where the data is useable to modify application functionality without altering the application itself. For instance, the data in column 42 can be used to specify a language (e.g., English, Spanish, etc.) to be used by the application, to set parameters to be used when the application is run, etc. For example, for the first instance of IVR Application 2 in column 40, the data in column 42 specifies the English language, for the first instance of Advanced App 1 in column 40, data in column 42 specifies two variable values Foo and Bar and for the second instance of Advanced App 1 in column 40 data in column 42 specifies that a male voice should be used when voice messages are broadcast. As shown in FIG. 2, In at least some cases a single application may be associated with two numbers in the same call routing table (see two instances of IVR Application 2 and two instance of Advanced App 1 in column 40.

[0055] Referring again to FIG. 1, platform database 18 also includes client application tables 34. Referring also to FIG. 3, an exemplary client application table 34 is illustrated that includes an application ID column 80 and an application column 82. The application column 82 includes a list of all of the IT applications that have been developed by an associated tenant (e.g., there is a separate application table for each service provider tenant). Exemplary applications in column 82 include IVR Application 1, IVR Application 2, Advanced App 1 and a simple IVR application. The ID column 80 points to a location in a database 33 (see again FIG. 1) that holds an execution code file for the associated application in column 82. Each application ID is unique and independent of which client developed the application. In at least some cases the application IDs are ‘universally unique identifiers’ so there is a guarantee that every application has a globally unique ID.

[0056] By storing the IT application code files independent of identity of the developing client, IT applications may be used by more than one client. For instance, a service provider may want to enable a second client to use a first application developed by a first client. In this case, the service provider may simply enter the unique application ID for the first application (e.g., the application developed by the first client) into an application table associated with the second client. This action makes the first application available for the second client to use but, in at least some cases, will disable the second client from being above to edit the first application.

[0057] Referring yet again to FIG. 1, platform database 18 further includes IT application database 33. Database 33 stores all of the IT applications (e.g., the executable code) 35 developed by the telephony clients and maintained by the service provider. Again, each of the developed IT applications will have a unique identifier that is independent of the client that generated the application. In addition, where a GUI was used to develop a visual flow of an application during application development, the visual flow representation may also be stored for subsequent access so that a client developer has the ability to use the simple GUI to edit an application if desired.

[0058] While FIG. 1 shows all of the software, routing and client application tables and applications 33 stored on a single platform database, other configurations are contemplated where the software, tables and applications are stored within a distributed database. For instance, one storage device may store the software programs run by server 12 during application development, a second device may store all of the route point tables and a third database device may store all of the applications developed by clients. In other cases subsets of the software, tables and applications may be stored on different distributed storage devices.

[0059] Referring again to FIG. 1, server or servers 12 include processors, memory and other hardware that cooperate to perform various functions. To this end, FIG. 4 includes a schematic of an exemplary server configuration 12 including internet protocol telephony switches 52, a message queue 54 and a plurality of channel coordination services (CCSs) 56a through 56d. The telephony switches 52 are linked to the carriers 50a through 50c and receive tenant customer calls there from and link those calls to specific CCSs through the message queue 54. As an IT application progresses and generates information to be conveyed to a client customer or as a client customer provides information back to the server 12, messages are placed into the scalable message queue 54 to be processed. Message queue 54 feeds the CCSs 56a through 56d.

[0060] Referring still to FIG. 4, each of the CCSs 56a through 56d is similarly provided and operates in a similar fashion and therefore, in the interest of simplifying this explanation, only CCS 56a will be described in any detail. CCS 56a is provided by a server processor that runs software to facilitate CCS functions. Each CCS is capable of handling a plurality of simultaneous calls from one or more of the carriers 50a through 50c by simply opening up a separate CCS thread 60 or “container” for each of the calls. A control supervisor process 61 is run by server 12 to monitor execution of each of the CCS containers. In FIG. 4 the exemplary container 60 includes un-trusted user code 62, a sandbox API 63, a voice control API 64, a utilities API 65 and a web services API 71.

[0061] Referring still to FIG. 4, when a call is received and assigned to one of the CCSs, after a container is opened for the call, the CCS accesses the source code file 62 for the application associated with the called number and sets up the secure sandbox API 63 to execute the user code. Server 12 creates or instantiates a logging function 66, a debugging function 67, a statistics function 68, and other functions or services to track execution of the code. Finally the execution code for the application is processed starting at the top of the code file.

[0062] Any executable application code file will include commands that are part of the LUA or other programming language as well as other commands that were specified by a client developer or that result from actions by a client's customer linked to a call. Any code commands that were specified by a client developer or result from actions by a client's customer are considered un-trusted. All un-trusted code is passed through the sandbox API 63 where the un-trusted code is run in a secure manner that prohibits actions that are not known and supported by the system. To this end, the sandbox API performs security checks on the un-trusted code and makes sure the code is using API functions correctly. Where un-trusted code cannot be deciphered, the code is ignored and an error message may be generated. Where un-trusted code is understood, the code is either passed on to one of the voice control, the utilities or the web services APIs 64, 65 or 71, respectively, to be processed, or the function specified by the code is mapped into correctly formatted API calls that the system can use and the correctly formatted calls are passed on to API's 64, 65 or 71. For example, if executable code includes the command “channel.play (‘filename.mp3’)”, a LUA interpreter running the code would pass the play command on to the sandbox API. In the sandbox API, the play command is performed which takes in the argument ‘filename.mp3’ and makes sure that the file specified by the argument exists. If the file specified does not exist, the command is ignored. If the file specified exists, the sandbox API maps the command to the voice control API that actually plays the file.

[0063] Utilities API 65 is used to perform functions that have been defined by other sources such as a third party. For instance, a date parser functional module may be provided by a third party to be called via a utility API. A program to convert HTML documents to another form may be provided by another third party to be called via an API. Other utilities APIs are contemplated.

[0064] Web services API 71 is used to control web service calls to external services. For instance, if an application requires retrieval of data managed by a client server 14, a web service call may be generated using API 71 to link to server 14 to access the required data. As another instance, if a client customer wants to transfer money from one account to another, API 71 may be used to update or alter at least some data in a client's database 20 to reflect the transfer. The web services API may also be used to obtain data or change data maintained by a third party.

[0065] Referring now to FIG. 5, a process or method 100 for developing an IT application that is consistent with at least some aspects of the present disclosure is illustrated. The process 100 will be described in the context of a process for developing an IT application to perform the simple IVR process shown in FIG. 6. To this end, FIG. 6 includes a follow chart showing an extremely simple exemplary interactive telephony process 84. In FIG. 6, when a customer call is received, at block 86 the phrase “Welcome to simple.” is to be voiced as a message to the customer. At block 88 the phrase “Please enter a digit.” is to be voiced as a message. At block 90 the process is to gather the digit entered by the customer. Here, in some cases the digit may be gathered as a digit selected via a customer device keypad or may be gathered as a voiced digit to be recognized by server software via voice recognition. At block 92 the process is to respond by confirming the digit entered by voicing the message “You entered ______”, where the underline is replaced by the entered digit. At block 94 the process is to be completed when the system automatically hangs up on the call.

[0066] Referring again to FIG. 5, at block 102 the platform server 12 monitors for a client device 115 to link to the server via a browser internet link. Once a tenant device links to the server 12, in at least some embodiments, a browser screening will present the client developer with an option to select one of two different ways to develop the LUA execution code needed to perform the IT process (e.g., the simple process shown in FIG. 6). To this end, in at least some embodiments of the present disclosure, a developer will be able to develop an application by either using a graphical user interface (GUI) or by directly scripting LUA code. The GUI is designed to enable the developer to build a visual flow of the IT process that is similar to the flow chart shown in FIG. 6 and therefore is very intuitive and often can be used after only minimal training. After the visual flow is developed, the visual flow is compiled or published causing the platform server to convert the flow to LUA execution code. If a developer works in direct LUA code, the code is entered directly by the developer and no compiling is required.

[0067] FIG. 7 shows an exemplary browser screen shot 140 for greeting a tenant when the tenant initially logs onto the development system. As illustrated, the exemplary screen 140 includes a greeting and some simple instructions for the developer to get started. Here, two on-screen selectable icons 144 and 146 are provided that allow the developer to select one of the two ways to specify an IT application program. Icon 144 is a “GUI” icon which, when selected, provides the graphical user interface for graphically developing a visual flow. Icon 146 is a “LUA Programming Language” icon which, when selected, enables the developer to script LUA program code directly. An on screen pointer icon or cursor 200 is moveable about screen 22 and used to select either of icons 144 or 146. In this example it will be assumed that the developer selects the GUI icon 140 to access the graphical user interface and therefore, in FIG. 5, control passes to block 104 where a GUI is provided.

[0068] Referring now to FIG. 8, an exemplary simplified GUI screen shot 140 is illustrated that includes a visual flow space 164, a programming functions tool bar 160, a connector tool bar 162, a name field 169, and publish and quit icons 190 and 192, respectively. Name field 169 is provided to receive a name for an IT application to be developed by a user. The application name can simply be entered into field 169. Often times the name entered into the field will be descriptive and have a clear meaning to the developer.

[0069] Visual flow space 164 is a space in which the developer can create the visual flow of an IT application by adding programming functions to the space, connecting the functions and adding properties to the functions. A set of functions are represented as separate blocks in tool bar 160 including a “Play” block 166, a “Say” block 168, a “Gather Digit” block 170, a “Gather Voice” block 172, a “Transfer” block 174, a “Web Service” block 176 and a “Hang-up” block 178. Other function blocks are contemplated and, where a full set of function blocks cannot be presented in the tool bar 160, a vertical scrolling tool 180 may be provided to scroll through the blocks.

[0070] Each of the blocks 166 through 178 can be selected to instantiate an instance of the associated function within visual flow space 164 to add the associated function to the application being developed. Block selection may include selecting a block with cursor 200 and dragging the block into space 164, double clicking on a block to create an instance of the block in space 164, etc. Once a block instance has been created, the block can be moved about in space 164, can be linked to other blocks, can be deleted, etc. At least some of the blocks can be customized by setting different characteristics as will be explained in greater detail below.

[0071] Referring still to FIG. 8, play block 166 can be used to cause server 12 to play an audio clip stored at a database address to a customer. Thus, for instance, if the phrase “Welcome to simple” is stored at a first database address, a play block specifying the first database address can be instantiated to play the phrase when the block occurs in an IT process. Say block 168 can be used to cause server 12 to convert text into a simulated voice message. For instance, a say block including the phrase “Welcome to simple.” could be instantiated to have the server generate the phrase as a simulated voice message. The gather digit block 170 causes server 12 to receive a digit entered by a client customer. The transfer block 172 causes server 12 to transfer a customer call to another phone number or extension. For instance, where a customer selects the number “2” for service, a call may be transferred to a service extension.

[0072] The web service block 176 can be used by server 12 to access other functions performed by other servers. For instance, referring again to FIG. 1, while performing an IT application, platform server 12 may require some data from a client database 20 in order to complete a portion of the process. For instance, where a bank customer uses an IT application to requests an account balance from a client, the platform server 12 would have to obtain the balance information from the client's database. A request to a bank server 14 to access an account balance would be one web service. Other exemplary data retrieving web services may include obtaining weather data, obtaining information related to a service call, obtaining a stock quite from a brokerage firm, obtaining pricing information from a retailer, obtaining flight status information from an airline, etc.

[0073] In addition to being useful for retrieving data, a web service call may be used to modify, delete or add to data maintained by a third party. For instance, a customer may use an IT application to change a delivery data for furniture being delivered to the customer's home. Here, a web service call may be used to alter data in a delivery database. Other web services are contemplated.

[0074] Hang-up block 178 causes platform server 12 to hang up a customer call.

[0075] Connector tool toolbar 162 includes connector tools that can be used to link function block instances within space 164 together to form a visual flow. Only one exemplary connector tool 182 is shown but other tools are contemplated. Here, once a tool 182 is selected, cursor 200 takes on additional functionality and may be used to draw lines between blocks to specify process flow.

[0076] Publish icon 190 is selectable to submit the visual flow within space 164 to server 12 to be compiled into LUA execution code. Quit icon 192 can be selected to exit the developer's software program.

[0077] Referring again to FIG. 5, at block 106 the developer uses the GUI to build up the IT application or process. To this end, see again FIG. 8 that shows initial step in specifying the IT application where a developer drags an instance 196 of the say function block into space 164 as shown at 194. Once block instance 196 is placed in space 164, in at least some cases, server 12 provides additional tools to guide the developer in specifying properties of the say function. In this regard see FIG. 9 where block 196 has been expanded (see 196a) upon placement to indicate other properties of the function that may be set. In the case of a say block, the most important additional information required is the text to be converted to the voice message. Thus, expanded block 196a includes a field for entering the text to be annunciated. Expanded block 196a also includes a field in which a developer can select different voice types including male 1, male 2, female 1, etc. Other property setting fields are contemplated.

[0078] In other embodiments, instead of expanding a placed cell to indicate additional properties that may be set, other interface tools may be used. For instance, in some cases when some additional information is required for a function like text to be voiced, upon placing the block a sub-window may be opened requesting entry of the text to be voiced or other required information. In other cases a sub-window may be opened upon clicking on a block instance or upon hovering cursor 200 over the instance to access tools for setting properties.

[0079] Referring again to FIG. 9, in addition to providing tools for setting other function properties, the GUI may also prod the developer by providing other graphical guidance. For instance, see that once block 196a is expanded, a next block 210 may be presented in phantom to prod the developer to add another block at the illustrated location where the next block is already linked via a line to the expanded block. Here, the developer would have the option to move a next block into the phantom space or to move a next block to any other location in space 164 and to use line tool 182 to manually link blocks.

[0080] Referring to FIG. 10, see that a developer has specified text “Welcome to simple.” and a male voice for the first say block 196b, has added additional function blocks to the visual flow space, has specified properties for several of the function block instances and has linked the blocks together to form a single simple flow. The other blocks include two additional say blocks 220 and 224, a gather digit block 222 and a hang-up block 226. Once a block is added to a flow, a developer can modify the flow by deleting a block, adding new blocks, amending existing block properties, changing the flow between blocks, etc. Comparing the visual flow in FIG. 10 to the intuitive flow chart shown in FIG. 6, it should be appreciated that the visual flow in FIG. 10 substantially mirrors the intuitive flow chart in FIG. 6. Thus, the GUI should be very intuitive to use regardless of the amount of programming training a developer has received.

[0081] Referring again to FIG. 5, after the visual flow has been completely specified at block 106, at block 108 the developer publishes the flow (e.g., selects icon 190 in FIG. 10) causing platform server 12 to compile the flow into executable LUA code. The executable code is stored along with its name in a database in an undeployed state so that the application is not available for use.

[0082] To deploy application code for use, a developer has to push the executable code to the platform server for storage. When a developer attempts to push new code to the platform server, the server 12 requires the developer to securely log into the system and add the application and an associated number to the tenant's call routing table (see again FIG. 2). Secure access is controlled by a username and password log on process and a user may be required to enter an RSA or other public key to identify the user.

[0083] In FIG. 5, at block 112, the server 12 determines if a user is authorized to manage a tenant's IT application routing. If the user is not authorized, control passes to block 116 where an error message is sent to the user and the user is not allowed to change the routing table. If the user is authorized, control passes to block 114 where the user can access the routing table and add the application name and an associated phone or extension number to the tenant's call routing table. The client may also add data to column 42 (see again FIG. 2) to be used by an associated application when the associated number is called. Once the application is added to the routing table the application has been deployed.

[0084] Referring again to FIG. 5 and also to FIG. 7, if a client developer selects the LUA Programming Language icon 146, control passes from block 102 to block 120 in FIG. 5 where LUA programming tools are provided as part of a scripting interface (not illustrated). At block 122 the developer enters the LUA logic or code and at block 110 the completed LUA code is pushed to the server 12 for storage after which blocks 112 and 114 occur as described above.

[0085] Exemplary LUA execution code that may be published from a visual flow or may be directly entered via the LUA scripting interface that corresponds to the simple process shown in FIG. 6 is shown in FIG. 11. The code includes various API calls that cause application sub-functions to be performed such as “say” functions 230a, 230b, 230c, a “get digit” function 232 and a “hang-up” function 234. Again, the FIG. 11 code may be generated either by compiling a visual flow developed using a GUI interface or directly via scripting. Where a GUI interface is used to generate code, the visual flow may be stored along with the code or in some other database to be accessed subsequently if the application has to be edited in some fashion. After publishing a flow, the LUA code can be opened in a scripting interface to be modified.

[0086] Referring now to FIG. 12, a process 300 facilitate an IT application when a customer calls a service provider telephony client is illustrated. At block 302, the platform server 12 monitors for an incoming call. Once a call is received, control passes to block 304 where the called number (or a client identifier provided by one of the ITSP carriers) is used to identify a client call routing table associated with the called client. At block 306 the called number is used to identify a route point for the call and at block 308 the application associated with the route point is identified in the routing table. At block 310 a channel process is instantiated by one of the CCSs 56a, 56b, etc., opening up a new CCS container to handle the call. At block 312 the tenant code is read from the application identified at block 308 and the code is executed at block 314. Execution entails running untrusted user code in the CCS sandbox API 63 and performing API calls via APIs 64 and 65 (see again FIG. 4). The code executes until the call is completed at block 316 after which control passes back up to block 302 where the process continues to loop.

[0087] While a simple process is described above in relation to FIG. 5, it should be appreciated that far more complex processes and GUI and other tools for specifying those processes are contemplated. Referring now to FIG. 13, another screen shot similar to the screen shot shown in FIG. 10 is illustrated, albeit showing a different visual flow for a different portion of a IT application. Here the flow includes a play block 250, a gather digit block 252 and a say block 254. In the play block 250, other properties that have been set include a number of times to repeat the audio clip associated with the block and the volume level at which to play the clip. In this case the process branches out from the gather digit block in one of three directions depending on which digit is selected by the customer. For instance, a “1” selection leads to block 254 while other selections will lead to other blocks yet to be specified. Thus, the GUI can be used to specify relatively complex branching functions.

[0088] Referring to FIGS. 14a through 14e, another executable code file is illustrated which provides a bilingual application with support for dynamic translation. Additionally web service calls are used to obtain an provide temperature data from a third party service, to use utility functions to parse an XML response and to use math functions to convert Fahrenheit to Celsius.

[0089] In at least some embodiments it is contemplated that as a client developer uses one of the interfaces to specify a telephony application, the developer may be able to run the application in intermediate form to experience what a customer would experience when using the system. To run an application, an application code file has to be generated. Thus, when the GUI interface is used, the visual flow of an application has to be compiled prior to testing. In the case of the scripting interface, the code file already exists and therefore no compilation is required.

[0090] Referring to FIG. 15, an exemplary screen shot showing a scripting interface is illustrated where a portion of the simple IVR application code is shown at 400 within a scripting workspace. The interface includes a left edge tool bar that includes at least a “run application” icon 402 for testing the existing code in the workspace. Referring to FIG. 16, when icon 402 is selected, a query window 404 opens up and requests additional information required to run the application. For instance, in the case of the simple IVR, the additional information may only include a telephone number (e.g., a number to be called to initiate the application during the test activity) to be associated with the simple IVR application. In FIG. 16 a number field 406 is provided to enter a number. In other cases other information required to perform an application may be requested such as the time of day at which the call is to be simulated (e.g., some applications operate differently based on the time of day), the geographic location from which a call is to be initiated (e.g., some applications operate differently based on where a customer calls from), whether or not the call is initiated on a holiday, the date on which a call is initiated, etc.

[0091] Referring again to FIG. 16, window 404 includes a second run icon 408 that can be selected to run the application after required information has been entered. When the application is run, the developer uses a phone to call the number entered into field 406 and the platform server runs the application based on the other information that was entered using window 404 (e.g., date, time, location, etc.) so that the developer experiences the application.

[0092] Referring again to FIG. 10, in the case of the GUI interface, a run application icon 420 may be provided for testing. When icon 420 is selected, a window like window 404 in FIG. 16 may be opened to enter other required information. Once the second run icon 408 is selected, the platform server 12 compiles the visual flow into code and runs the immediately runs the code for the developer in a fashion similar to that described above. In this case, while the flow may be compiled to generate code to execute, the compiling and running of the code can be done behind the scenes so that the developer still sees the visual flow. In at leas some cases the interface may highlight function blocks on the developer's visual flow as the functional blocks are performed during the test simulation. To this end see that block 222 is highlighted to indicate that block 222 is currently being performed during a test.

[0093] A new interactive telephony or interactive voice recognition telephony system has been described where client developers can use either a GUI interface or a scripting language interface to develop LUA code or other familiar language code to specify customized and complex interactive telephony applications. Pre-developed functional modules are provided by a service provider that facilitate routinely useful functions which can be called via API calls from within applications. The telephony application programs are not run by clients themselves. Rather, after deployment, the application programs are stored and run by a platform server maintained by a telephony service provider. The service provider receives calls to the telephony client and facilitates client specified applications so the client need not dedicate substantial resources to facilitate interactive telephony services. The service provider uses web services to access client stored data and/or to change client stored data when necessary.

[0094] While the invention may be susceptible to various modifications and alternative forms, specific embodiments have been shown by way of example in the drawings and have been described in detail herein. However, it should be understood that the invention is not intended to be limited to the particular forms disclosed. For example, while the system is described above as one where a telephony client can develop an interactive telephony application using a browser based interface provided by a platform server operated by a service provider independent of the client, in other embodiments it is contemplated that a client may use a local LUA scripting interface that is independent of the service provider to develop LUA code. Here, the client developer can should be familiar with a set of API calls used to access code for performing functional sub-processes that are supported by the system so that those calls can be used within the LUA code. After a complete LUA code file is specified using the local interface, the client developer can push the code file to the service provider's server for storage, association with numbers in the client's routing table and ultimate deployment to be used to field calls from customers.

[0095] As another example, as briefly discussed above, the processes and functions performed by the platform server 12 may be distributed and performed by a set of cooperating servers. For instance, application development may be handled by one server, data storage may be handled by another server, running application code may be facilitate by a third server and so on. One important aspect here is that none of the platform server operations are performed by the telephony client.

Additional Features

[0096] Referring again to FIG. 4, as explained above, in at least some embodiments, a separate instance of a CCS thread or call container 60 is used to process each call handled by platform server 12. Container isolation is important in many cases so that actions or processes in one container can be completely isolated from all other containers. For instance, where processing of one call in one isolated container becomes bogged down for some reason (e.g., gets stuck in an endless processing loop), the bogged down state does not affect other call processes handled by other containers which operate completely independently of the one container.

[0097] In at least some embodiments it is contemplated that a software tool called Docker may be used to provide call containers 60 (see FIG. 18 that is similar to FIG. 1, albeit showing a different platform database arrangement). FIG. 19 shows an exemplary call routing table associated with the FIG. 18 database arrangement and FIG. 20 shows an exemplary container database 60 that is consistent with at least some aspects of the following disclosure. The Docker tool is well known in the software programming arts and therefore will not be described here in detail. Here, it should suffice to say that Docker containers wrap up a piece of software in a complete file system that contains everything the container process or program needs to run including code, system tools, system libraries, etc. More specifically, Docker is an open-source project that automates the deployment of applications inside software containers, by providing an additional layer of abstraction and automation of operating-system-level virtualization on Linux. Docker uses resource isolation features of the Linux kernel (e.g., cgroups and kernel namespaces) to allow independent containers to run within a single instance of the Linux operating system. While Docker containers are particularly suitable for use in the context of the present disclosure, other container types or virtual machine types are contemplated.

[0098] In general, Docker containers work well for companies that run a relatively small number of programs (e.g., 5-10) where program startup time need not be extremely fast (e.g., a few seconds start up time is acceptable) and where the programs need to be isolated and run for long periods of time. For instance, a company may open 10 containers and persistently run ten application programs, one in each of the 10 containers. Here, if an eleventh container is needed for some reason, an eleventh container can be instantiated for use, typically within a short period. Container instantiation includes a container staging process (e.g., a staging state) at the end of which a container is in an idle (i.e., ready) state and available for activation to run an application program.

[0099] In the case of a telephony service provider operating a platform server as described above, one problem is that call volume fluctuates wildly at times and can be unpredictable. For instance, during the course of a day, call volume may fluctuate from high to low several times and, at times, may fluctuate from high to low and back within a matter of just a few seconds. Here, if containers are instantiated just in time when required for newly received calls, telephony client customers can experience an objectionable delay in IVR performance when a large number of containers need to be staged at substantially the same time.

[0100] One solution to the staging associated delay is to maintain large number of surplus idle containers ready for immediate activation when new calls are received so that there are always available idle containers to handle call volume spikes. For instance, a telephony service provider may have a rule of thumb that the provider wants to maintain at least 20% more idle containers available for use over their highest number of previously simultaneously handled calls. For example, where a maximum number of calls simultaneously handled by platform server 12 in the past has been 10,000, the provider may maintain at least 12,000 containers active or idle all the time. Here, after a container completes a call application (e.g., after a client customer ends a call process), the container may be returned to the idle state inventory for subsequent use in handling another received call.

[0101] Unfortunately, maintaining idle containers requires hardware overhead and therefore there is an appreciable cost associated with systems that maintain excessively large idle container inventories. For instance, while 10,000 or more call containers may be needed during high volume periods, less than 1,000 containers may be needed during low volume periods and therefore, where 12,000 containers are maintained idle or active at all times, during low volume periods 11,000 containers may be idle.

[0102] One other problem with container based systems is that container reuse can cause one call process to contaminate and cause a malfunction in a subsequent call process that is handled by the same container. For instance, in at least some cases, at least a sub-set of data from a first call may persist within an idle container or an application state may persist in an idle container after a call application has been completed and when a subsequent call application is performed by the same container. Not surprisingly, this data or application state persistence can causes errors in handling of subsequent calls.

[0103] One solution for handling the data/state persistence problem associated with reuse of a container that is consistent with at least some embodiments of the present disclosure is to “kill” the container after a single call application has been completed (e.g., after the application enters a “finished” state). Here, instead of reusing a previously “used” container (e.g., a container that has already run a call application for a single call) for a subsequent call, each call would be handled by an entirely new and unused container so that data and application state persistence would be impossible.

[0104] One solution for handling the problems associated with an excessively large idle container inventory is to monitor container states and dynamically control the number of idle containers in the idle container inventory so that there is always an idle container ready to handle a newly received call but where the total number of idle containers in inventory does not exceed the number required to handle an anticipated large spike in simultaneous calls.

[0105] The process for dynamically controlling the idle container inventory size can take many different forms. For instance, instead of following a rule of thumb to maintain 20% more idle containers available for use over a highest number of previously simultaneously handled calls, a different rule of thumb may be to maintain the greater of 300 idle containers and 20% more idle containers available for use over a current number of ongoing calls. Here, at least 300 idle containers would be maintained at all times but the number of idle containers would be increased if the total number of currently on going calls exceeds 1500 (e.g., 20% of 1501 is greater than 300). As another instance a different inventory requirement may require that the total number of idle containers in an inventory be tied temporally to historical call volume requirements. For instance, if it is known that a call spike occurs every morning from 8 AM to 10 AM, the inventory size may automatically be ramped up just prior to 8 AM to a level required to handle the anticipated spike. As one other instance, another inventory requirement may tie the number of idle containers in inventory to the maximum rate of call volume increase in the past, during a specific time period, a specific period during a day, during a specific sub-period during a week, to a specific sub-period of a year, etc. For instance, one inventory rule may be that the inventory during a 4 AM to 8 AM period be tied to the maximum rate of call volume increase in the past during the 4 AM to 8 AM time slot while the inventory during the 8 AM to 10 AM period is tied to the maximum rate of call volume increase in the past during the 8 AM to 10 AM time slot. In some cases the inventory requirement may change over time as a system server learns from real time control how an inventory number should change and be controlled to meet the ultimate goal of minimizing inventory size while still ensuring that idle containers exist in the inventory for every new call received. As one other instance, a very simple inventory requirement may be that at least 400 idle containers should be maintained in inventory at all times based on a monitored idle inventory count.

[0106] In still other cases some other condition associated with container states may be used as a trigger to increase idle container inventory. For instance, in some cases the state monitoring container may only increase idle container inventory whenever a new call requires staging of a new container to handle the call. Thus, for example, where an initial inventory of 300 idle containers is eventually used up or active to handle existing calls, when a 301st call is received and a new container has to be staged to handle the call, the state monitoring container may be programmed to stage another 300 containers to replenish the inventory. Here, while some calls may experience some delays, most calls would not as the inventory would only require staging of a new container for an already received call some of the time.

[0107] Another inventory rule may simply require that the state monitoring container count the number of activated containers and always maintain a running idle container count that exceeds the activated container count by a set number of containers (e.g., 300) without tracking the number of killed containers. Here, the inventory would always be 300. Many other inventory requirements or target rules are contemplated.

[0108] In some cases, the number of idle containers and the number of total open containers (e.g., staged, idle, active and finishing (e.g., at the end of an application)) may be tracked and compared to determine when additional idle containers need to be opened and when existing unused idle containers should be closed. For instance, each time one idle container is activated to run an application to handle a received call, the numbers of idle and open containers may be compared and the number of idle containers adjusted to meet a dynamic idle container inventory target. As another instance, each time one active container completes a call application and is killed, the numbers of idle and open containers may be compared to determine how to adjust (e.g., stage or kill) the number of idle containers in inventory.

[0109] To regulate the idle container inventory size, in at least some embodiments, at least one persistent state monitoring container (hereinafter “the state monitoring container”) may be instantiated to run a special call management application that tracks container states and manages staging new containers and killing unnecessary idle containers pursuant to a dynamic idle container number target, rule or requirement. Here, each call container may be programmed to track its own state and to report any or at least a subset of state changes to the state monitoring container immediately upon its occurrence. For instance, each application may include code causing a container to report out application initiation to the state monitoring container at the beginning of the application code so that the monitoring container knows when a container moves from the idle state to an active state. As another instance, each application may include code causing an associated container to report out when an application ends and that the container has again entered an idle state and should be killed off. Here, where the number of idle containers exceeds a dynamic goal, the state monitoring container may request that at least a sub-set of idle containers be killed (e.g., closed). Where the number of idle containers is less than a dynamic number target, the state monitoring container may cause the platform processor 12 to stage instantiation of additional idle containers for subsequent use.

[0110] When idle containers are to be killed, the state monitoring container can kill the containers in any of several different ways. For instance, the monitoring container may transmit a specific “kill” message to each separate idle container that is to be killed. As another instance, the state monitoring container may transmit a group message to a subset of containers (e.g., 300) indicating that if a receiving container is in the idle state, the container should be killed off. Here, where an initial group message achieves an idle container target number, the process of monitoring the goal would continue. Where an initial group message does not achieve a container goal within some pre-defined time period (e.g., some of the containers that receive the initial message are not idle and therefore do not act on the kill command), the monitoring container may transmit another group message to a different subset of containers requesting that idle containers be killed off within that subset. This triage process would continue until an instantaneous dynamic idle container inventory target is achieved.

[0111] In particularly advantageous embodiments, as described above, each container is only used once to perform a single call application after which the container is immediately killed. Here, where a “used” container is killed off, idle container number reduction can be accomplished by simply not staging new containers as applications are completed. In this case, the state monitoring container would still track the idle container number and compare that number to an idle container goal so that if the idle container number drops below an instantaneous goal, new containers could be staged rapidly to continually replenish the idle container inventory and meet the inventory number target. In some cases used containers would be killed pursuant to a command from the state monitoring container. In other cases it is contemplated that each application program could include a container killing code segment to automatically kill off the container at the end of an application program.

[0112] One advantage associated with a system where used containers are killed off automatically after an associated application is completed is that the problem of data persistence among consecutive applications run within the same container is entirely avoided. Thus, here, because applications are only performed in new and previously unused containers, there is no possibility of data, application state or code from a prior application contaminating a new application.

[0113] Referring once again to FIG. 4, one embodiment of the present disclosure is illustrated that maintains an inventory of new and unused idle containers 500 for performing call applications for received calls. In FIG. 4, active containers associated with CCS 56a (including active container 60) are collectively labeled 502. Control supervisor process 61 operates within CCS 56a as a state monitoring container (also labeled 61 hereafter) which monitors states of each of the active containers 502 and, when a container completes an application, kills off the container to avoid reuse that could lead to unintended data and/or code persistence in subsequent applications. State monitoring container 61 also tracks the total number of open containers (e.g., staging, active, finishing and idle containers) as well as the number of idle containers collectively identified by numeral 500 and stages new containers to replenish the idle inventory 500 to a number consistent with a telephony service provider's dynamic number target. Hereinafter, in the interest of simplifying this explanation, unless indicated otherwise, it will be assumed that the idle container inventory target requires an idle container inventory within a range between a minimum number of idle containers and 120% of the minimum number of idle containers where the minimum number is the greater of 300 idle containers and 20% of the total instantaneously “open” containers. Here, the term “open” is used to refer to currently staging (e.g., containers currently being spun up into an idle state) containers as well as active containers. Thus, for instance, the idle container inventory would be maintained, at a minimum, within a range between 300 and 360 (e.g., 120% of 300) and within a maximum range of 20% of the number M of instantaneously open containers and 24% (e.g., 120% of 20%) of the number M of instantaneously open containers. Here, for instance, if the number M of open containers is 10,000, the number of idle containers in inventory would be maintained between 12,000 and 12,400.

[0114] While some systems may have a single state monitoring container 61 that monitors container states and numbers for all CCS's (e.g., 56a through 56d in FIG. 4), in other cases each CCS may maintain its own state monitoring container and the state monitoring containers may communicate via a network connection to maintain idle container goals together. In addition, in some cases a single CCS may maintain two or more state monitoring containers that cooperate to maintain target idle goals or for redundancy purposes.

[0115] Referring now to FIG. 17, a process 520 that may be performed by the state monitoring container 61 and the platform server 12 generally shown in FIG. 4 is illustrated wherein the exemplary idle container goal described above is implemented. At process block 522, upon initial startup of the system, monitoring container 61 stages 300 new containers to build up an idle container inventory of 300. At decision block 524, the platform server 12 monitors for a new received call from a telephone client customer. If a new call is not received, control passes down to block 531 which is described below.

[0116] Once a new call is received at block 524, control passes to block 526 where the called number is used to identify a client call routing table (see again FIG. 2). At block 528 an IT application associated with the called number is identified. Referring still to FIGS. 4 and 17, at block 530 one of the idle containers in inventory 500 is activated and the IT application identified at block 528 is downloaded and initiated within the newly activated container. At block 531 the number M of open containers (e.g. active and staging) is determined by state monitoring container 61.

[0117] At decision block 532 the number of idle containers in inventory 500 is compared to the low end of the idle container inventory target range (e.g., to the value 300 and 20% of the total number M of open containers). When the number of idle containers in inventory is greater than or equal to this low end range value, control passes to block 536. When the number of idle containers in inventory is less than the low end range value, control passes to block 534 where state monitoring container 61 stages a new idle container which is added to inventory 500. As shown, the low end of the inventory range is the greater of 300 and 20% of the total open (e.g., active and staging) containers. While only one container is staged at process block 534, in other embodiments it is contemplated that a plurality (e.g., 50 or some percentage (e.g., 15%) of the instantaneous number of open containers) of containers may be staged whenever the inventory target is not achieved at decision block 532.

[0118] Referring still to FIGS. 4 and 17, at process block 536 the number of idle containers in inventory 500 is compared to the high end of the idle container inventory target range (e.g., to the value 360 and 24% (e.g., 120% of 20%) of the total number M of open containers). When the number of idle containers in inventory is less than or equal to this high end range value, control passes to block 540. When the number of idle containers in inventory is greater than the high end range value, control passes to block 538 where state monitoring container 61 kills off some of the idle containers in inventory. The container killing process at block 538 may include killing off one container or a plurality (e.g., 50 or some percentage (e.g., 15%) of the instantaneous number of open containers) of containers whenever the inventory target is not achieved at decision block 536.

[0119] Referring still to FIGS. 4 and 17, at block 540 states of applications running or completed in the open containers are determined. At decision block 542, when an application in a container has been completed, control passes to block 544 where the container associated with the completed application is killed. Thus, the 540 through 544 steps in the illustrated process cause natural attrition of the inventory number to occur as used containers are killed or eliminated.

[0120] In at least some cases a container running a call application will get hung up and run in an endless cycle for some reason. Here, in at least some embodiments it is contemplated that when an application stops progressing for some reason, the state monitoring container may recognize the hung up condition and perform some process to deal with the condition. For instance, in some cases, once some threshold level of certainty is reached that an application has ceased working is achieved, the state monitoring container may send a kill message to the container associated with the application causing the application to be killed or closed out. When a container handling a call is killed in this fashion, the call is simply ended and the client customer has to call back into the system and start the application process over in a new container.

[0121] In other cases it is contemplated that the state monitoring container may cause any hung up call to be rerouted to a different container in which a call cancel notification application is run. For instance, when a call gets hung up in one container, the calling customer may be linked to a new container in which an application simply plays the message “An error has occurred, please call back for service. Thank you.” Once the call is redirected to the new container, the container associated with the application that got hung up could be killed. Once the cancel notice message has been broadcast and the related application has ended, the container associated with the cancel notification can be killed.

[0122] In still other cases it is contemplated that each call application may track and store state information during a call which could be used to inform another instance of the application about application state when an application becomes hung up and that information could be used to drive the new instance of the application so that an associated call does not have to be reinitiated. For instance, an application may track the current step in the application process to which a customer progressed as well as data associated with customer's interactions with the application process up to the point at which the application became hung up. Here, when a first instance of the application gets hung up, upon the state monitoring container recognizing the hung up state, the monitoring container may open a new or second instance of the hung up application in a new container and then use the state and data from the first instance of the application to move to a point in the second instance of the application at which the hang up occurred so that the customer could proceed from that point instead of requiring the customer to call back into the system and restart the process anew. In at least some cases the application state information stored may include state information just prior to the time at which a hang up occurs so that if the hang up was due to the most recently developed state information in the first instance of the application, that information would not be ported to the second instance of the application.

[0123] In still other cases where an application gets hung up for some reason prior to completion of a call, the state monitoring server may simply start a second instance of the hung up application in a new second container and then kill the first container, thereby requiring the customer to restart the application process without having to recall into the system. Here, in some cases an intermediate container (e.g., one that operates after the initial container and prior to the second container) may be activated and the call routed thereto to provide some guiding instruction like “An error occurred, this process will start again from the beginning”, prior to routing the call to the second container to restart the application process.

[0124] FIG. 21 is a flow chart showing a sub-process that may be added to the process shown in FIG. 5 for handling a container crash that is consistent with at least some aspects of the present disclosure.

[0125] Thus, the invention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the following appended claims. For instance, while the system is described above as one where a state monitoring container monitors all container states and regulates idle container inventory to a target number or range, in other cases it is contemplated that some other system operating characteristic may be monitored and factored into the decision of when to stage new containers. For instance, in some cases, the state monitoring container may simply monitor the number of calls received by a system and may tie new container staging to that operating characteristic as opposed to an actual inventory count. Thus, where 500 new calls are received in a few second period, the monitoring container may stage 500 or more new containers to be added to the idle inventory. As another instance, the state monitoring container may only stage new containers once an inventory has been depleted without consideration of the states of any other open containers. Thus, here, the monitoring container would only track the current inventory number and not all container states and replenishment would be based on inventory condition alone as opposed to states of all containers.

[0126] While the systems described above include a monitoring container that essentially always monitors for an event or condition that warrants a change (e.g., staging or killing) in the number of idle containers, in at least some embodiments it is contemplated that the monitoring container may only operate periodically to assess inventory condition and rejigger the number of idle containers in the inventory. For instance, the monitoring container may only check inventory level every 10 minutes. In some cases the frequency with which inventory is checked may change as a function of temporally proximate system operating characteristics. For instance, if high volume fluctuations are occurring during a period, the monitoring container may reduce the interval between inventory assessment and re-jiggering activities.

[0127] To apprise the public of the scope of this invention, the following claims are made: