Logo GISdevelopment.net

GISdevelopment > Proceedings > GITA > 1997


GITA 2002 | GITA 2001 | GITA 2000 | GITA 1999 | GITA 1998 | GITA 1997
Sessions

Advanced Technical Topics

Building & Supporting Applications

Business Evolution & Platform Migration

Expanding the User Base -- Non-Traditional Applications

From the office to the Field

Fundamental & Economic Issues of AM/FM/GIS

Lessons Learned

Major Technology Trends and their Impacts

Project Planning, Implementation and Management

Re-Engineering and Integration Issues

Scada and Real-Time Systems

User Project Presentations

Best of the Rest

Invited Presentation


GITA 1997


Lessons Learned


Telecommunications outside plant management throughout brasil


In spite of requiring from the GIS the support for imaging we have not being able to derive the advertised benefits from its technology. However, imaging is provided in SAGRE. We are fortunate that the GIS we selected remains competitive in this area. Some of our users implement a two-step strategy with regard to imaging, where imaging (AM) comes first and then, ?incrementally, the data is raised to the FM level. SAGRE allows the placement of images in any map or schematic as a background drawing and also to store an image linked to any of its objects, even if they do not have a graphical representation. Also, these images can be checked out and in for integration with other imaging systems.

Architecture
TELEB~S has developed an integrated operations support system plan and architecture. The plan defines the main sub-systems and databases to fulfill the functional needs of our operating companies. SAGRE follows that plan and architecture.

SAGRE architecture is open and built on three layers: user interface, application rules, and GIS database. The database layer is the core of SAGRE, and has been designed using object oriented techniques. The GIS adds greater flexibility and uniformity to this layer because it uses a commercial relational data base management system to store both spatial and nonspatial data, Both interface and application layers are based on object oriented standards. All layers are driven by metadata to encapsulate interface, data base and application objects. This has made possible the development of dynamic user interfaces and operations, whose components are specified at run time (Oliveira 95).

The metadata capability accelerated the development by reusing not only code but also design characteristics. The architecture also enabled the derivation of our conversion methodology that collects converted data in an object oriented fashion much like the on-line operations. A major challenge was to reconcile two entirely different user work modes -- operations and design/planning. Most of Operation transactions originate from customer phone calls regarding service repairs and changes. The Design mode comprises long duration activities as opposed to short operational transactions. The engineers interact with the database in order to pose spatial queries that analyze the existing facilities against projected demand for a given region. Design work is conditional and should not affect existing data unti 1the work is constructed.

SAGRE must thus be able to support such activity profiles, and the database must consider concurrent access by both maintenance and design/operations teams. To allow integration of these processes, the project had to use check-in/check-out and versioning mechanisms. Besides supporting alternative designs, this allows cooperative work among design and operational teams (Dias 95).

Another architectural feature in SAGRE is its ability to integrate a CAD tool into the design process. With this feature, we can start deploying SAGRE with the help of a CAD tool while conversion is taking place.

Implementation
Although it is very well known that systems like SAGRE take a long time to be implemented, the end users cannot wait for the completion of the whole system. We decided to build a first version after completion of data modeling but with limited functionality. This version had several objectives:

(a) to come out with some functionality to satisfy the end users pressure;
(b) to test the modeling with real data;
(c) to get feedback on the user interface;
(d) to start data conversion.

This decision proved to be very good. Note that this first version was not a prototype. It had limited functionality but implemented the complete data base structure. Usually, a prototype is constructed and a pilot is made. Sometimes, conversions are made without knowing the final data model causing many problems when trying to save conversion work. This approach also allowed us to modifi the paradigm of data conversion that will be discussed later in this paper. SAGRE first version was the seed for two of its modules, the SAGRE/Records (Gonqalves, 94) and SAGRE/Conversion modules. SAGRE/Records maintains the existing plant database and SAGRE/Conversion implements our conversion methodology and tools.

While maintaining the first modules, we started designing and implementing the complete functionality. The data model proved to be very stable but the user interface, as expected, had many problems. We learned that we should have taken more time in the construction of the interfaces. A new interface model was built and the complete system architecture was implemented. A new SAGRE/Records and SAGRE/Conversion, the SAGRE/Administration, SAGRE/Design, and SAGRE/Operations comprise the functionality required for SAGRE (Prezzoto, 95).

The high investment we did in modeling the database is paying off gradually. The database is steadily supporting the architectural enhancements being implemented as well as the increase in functionality required by the end users. Also, the databases being constructed with the conversions are coming up with sizes unexpectedly small. We believe this is also due to the GIS data structure.

Deployment
We learned from the beginning that conversion was the biggest problem to be considered in an AM/FM/GIS system. With the release of SAGRE/Records version 1.0 we started a bid for a pilot conversion. With the help of consultants we knew that doing in-house conversion could be a bad decision. Conversion requires specialized assembly line type of work. The bid called for the traditional conversion approach -- the conversion company would handle the data in our GIS interchange (low level and proprietary) format. We were not satisfied with the situation since the 390?Brazilian conversion market had very little experience with GIS conversion, and in particular, with the GIS we were using. Before the closing date for the bid, we decided to change the format (and the paradigm) of the conversion. We came out with an (land base and outside plant) object oriented format to be handled by SAGRE/Conversion module (Magalhiles, 94b).

The methodology and tools we adopted are based on the idea of manipulating the systems objects in the same level of the user interface. For example, to lay an aerial cable segment the user selects the poles, draws the cable path, enters the attribute data, etc., then the conversion data should resemble this mode of operation. The conversion company receives a detailed technical specification on how to read the maps and forms information and translate them into our standard format. Conversion data is input in batches. Each batch is submitted to SAGRE/Conversion module that reads the conversion data, makes the same consistency check performed by SAGRE/Records, resolves the elements continuity problems due to maps boundaries, reports all errors, calculates quality assurance figures, and separates batch into two others -- the set of objects that have been entered correctly and the set that did not enter due to errors. The data allowed into the system may have only layout problems, which is also minimized due to graphical consistency checks. For example, duplicate geometry is not allowed. This pilot conversion was set to provide us with: costs and timing for the Brazilian context; validation of the conversion technical specification; experience with contracting and accepting conversion work; a broader knowledge of the process; levels of quality assurance; qualification of conversion vendors; a real life and complete database representative of our reality for testing and integration of SAGRE modules. The pilot was a success and its pioneer data has helped TELEBRAS to cut significantly the conversion cost and complexity.

SAGRE conversion methodology and tools are evolving continuously. Each time a new conversion bid is set improvements are made in order to allow the operating companies to move into SAGRE faster without reducing quality level.

Conclusions
SAGRE is a system that has pushed AM/FM/GIS into the mainstream of TELEBRAS technology support. The key decision that allowed us to arrive earlier in this status was to develop a system to cover the entire outside plant information life cycle. Also, we had the courage and support to try new technologies and methods that worked well. In this process we took advantage of the cooperative spirit of the AM/FM community by sharing experiences with many users and vendors.

Acknowledgments
SAGRE is a cooperative effort among TELEB~S R&D Center and the Operating Companies. Our Operating Companies contributed in the developments in many aspects. In particular, they contributed with dedication in the specifications, system testing, prototyping and pilot implementations. In the R&D Center, a group of engineers and system analysts work on SAGRE design and implementation. SAGRE success is owed to all these people.

References
Magalhiies, G C, 1994, The development open systems for engineering applications: Proceedings of the XVII AM/FM International Conference, AM/FM International, Denver/CO, march 14-17, 1994.

Oliveira, J., Cunha, C. and Magalhiies, G. Object model for dynamic visual interfaces construction: Proceedings of the IX Brazilian Symposium on Software Engineering, SBC, Recife/PE, October 3-6, 1995. (in Portuguese language)

Dias, E., Granado, S., and Magalh~es, G. The use of versions to enforce consistency in hybrid operations and design environments. Proceedings of the IX Brazilian Symposium on Database, SBC, Recife/PE, October 2-4, 1995. (in Portuguese language)

Gonqalves, J. Erhardt, M., Obata, F. and Granado, S. Outside plant records automation: Proceedings of the I Meeting of Telecommunications Networks and Terminal Equipment Quality, Be16m/PA, September 13-15, 1994. (in Portuguese language)

Prezzoto, A., Teijero, D. and Obata, F. Telecommunications outside plant design automation: Proceedings of the II International Congress on Information Engineering, University of Buenos Aires, Argentina, November 14-15, 1995.

Magalhiies, G., Giglliotti, A., Santos, C., Teijero, D., and Argondizio, E. Data conversion technical specification - TELEB~S proposal - SAGRE Project: Proceedings of GIS Brasil 94, Curitiba/PR, October 17-21, 1994. (in Portuguese language)

Page 2 of 2
| Previous |

Applications | Technology | Policy | History | News | Tenders | Events | Interviews | Career | Companies | Country Pages | Books | Publications | Education | Glossary | Tutorials | Downloads | Site Map | Subscribe | GIS@development Magazine | Updates | Guest Book