UMS: Software to Automate Operators’ Actions

2.1 Problems with OpS interface

Operations vary over time, but it is often difficult to continue using the same OpS interface. Moreover, modifying an OpS interface is not easy because it takes time and money. While it is possible to implement a new operation by using the original OpSs, the operator’s workload will be much greater because the original interface was not designed to support the new operation. This leads to high operating costs.

Another problem is that databases slowly become corrupted owing to changes in the managed entities over time. The solution is to occasionally indentify the actual situation and alter the corresponding database entries. However, we cannot directly alter the databases because the data must be checked by the logic written into the OpSs, and usually only the OpSs can access and modify databases. Thus, we need a tool that eliminates the manual work needed to feed data items through the appropriate OpSs to check for consistency and perform any necessary updates.

Consider, for example, the case in which a new high-speed plan on the access network is started. The OpS that handles customer records does not to be modified, but the new plan uses new equipment, so an operator in the field should confirm the equipment with the equipment-management OpS. Thus, the operators’ workload will be high.

2.2 Solutions

A prime objective of UMS is to reduce operating costs. To achieve this, UMS eliminates operator actions that can be automated (Fig. 1). Thus, UMS can reduce the operator’s workload and improve the efficiency of operations.

Fig. 1. Approach of UMS.

3.1 No modification of target OpS

The target OpS does not need to be modified because UMS manipulates the target OpS in the same way as the operator. Actually, OpS manipulation means manipulating the client-side application of the OpS by actions such as clicking a button and inputting a string into a textbox on a personal computer or other terminal. UMS can record and replay these actions without any OpS modification. Thus, UMS can automate most operations at low cost and in a short time.

This approach is similar to the macro function well known to software programmers. However, creating a macro function would require alterations to the OpS. Moreover, a macro function would not be able to access or control other OpSs. UMS automates operations that involve multiple OpSs without any of them being modified.

3.2 Automation by field operators

UMS is designed to support field operators since they are the intended users of OpSs. While field operators are specialists in operations, they are not necessarily computer specialists, and most operators have difficulty in creating advanced programs. To overcome this problem, UMS provides a visual editor and automatic recording via operator example. As a result, it can be used by any operator, even computer novices.

4.1 Automation of repeated and stylized operations

UMS can automate repetitive and stylized operations. An example of a repeated and stylized operation is database correction. In this operation, operators input a lot of correct equipment data into an OpS at a time. The operation proceeds as a specific action flow consisting of, in this example, initiating a search, entering the search target, selecting a search result (target equipment), opening an window containing information about the target equipment, updating the target equipment’s data items in the window, and closing the windows.

UMS lets the operator easily record the actions needed for one cycle, and it then executes the actions iteratively. Therefore, UMS can easily automate repeated and stylized operations.

4.2 Prevention of data mismatch

UMS is effective at preventing data mismatches arising from human error. When the operator must enter data items into several OpSs, he or she must accurately complete all the entries by hand. The operations involved are troublesome and errors such as entry omission and erroneous data entry may occur. As a result, the operations are likely to produce data mismatches.

With UMS, the operator can automate the operations needed for populating multiple OpSs with the appropriate data items, while, at the same time, automatically checking the correctness of data items against the logic of each OpS. Therefore, UMS can prevent data mismatches caused by human error.

5.1 Scenario auto-generation

A scenario is information that specifies actions and the interface manipulation flow for automation. The operator can automatically generate a scenario by performing desired actions manually. UMS detects the operator’s OpS manipulation events and infers his/her actions. It adds the actions to a scenario according to the manipulation flow.

5.2 Visual editor

UMS uses flowcharts for intuitive viewing and editing of scenarios (Fig. 4). An operator can understand a scenario as a visual presentation (flowchart) and can edit the scenario in an intuitive manner without coding.

Fig. 4. Visual editor of UMS.

5.3 Data handling at runtime

The same operation will use different data in each cycle, so UMS provides a data management function in addition to the scenario management function. When UMS reruns a scenario, it identifies the variables in the scenario and accesses the appropriate data list to discover the values of the variables to be entered, record by record (Fig. 5).

Fig. 5. Data handling at runtime.

After recording the scenario, UMS creates an empty Excel or CSV (comma-separated values) file (the data list) whose column headers reflect the variables entered by the operator and those returned from the front OpS. The operator can also create data lists manually. Moreover, the operator can also use data lists that were extracted from an OpS. While executing a scenario, UMS can copy data output by an OpS and enter it into the appropriate data list record for later reuse.

5.4 Plug-ins

Considering future scalability, we designed UMS’s architecture so that its functions can be easily extended. UMS contains program code that depends on the operating system, browser, and so on because it must detect the operator actions and replay them. Accordingly, we separated this code from the main UMS program as plug-ins. In the future, when we adapt UMS to other environments, such as a new operating system or new browser, we will only have replace the plug-ins. Moreover, we can add new functions to UMS by means of new plug-ins.

5.5 System requirements

The current version of UMS runs on Microsoft Windows and can automate actions related to typical graphical-user-interface-based applications and Microsoft Internet Explorer. In addition, low-level UMS functions include mouse and keyboard control. These functions can automate applications on other platforms (e.g., Java and Flash) and applications running in a server-based computing environment (e.g., Citrix XenApp).