ONE Logo

Technical Information

 

ONE Home Page Information Desk Page Courses: K-12 thru University-level Degree Search Page Information for K-12 and Distance Online Courses Student Services Page Career Information Page Faculty and Staff Information Page Course Catalog menu bar

Invisible spacer image

Questions and Answers about Technical and Operational Aspects of the ONE Project


The following questions and answers on technical and operational issues in creating and running the Oregon Network for Education (OregonONE.org) website were prepared by ONE staff with OUS Information Technology Services. This document is in two parts: Part One: On Building the Database (Questions 1 - 9) was completed March 2000 after completion of the database, but prior to its public unveiling. Part Two: After a Year of Public Operation (Questions 10 - 22) was completed March 2001 after one year of operation.

    PART ONE: On Building the Database

  1. What determinations went into the decision to select Oracle for the ONE database? Why was Oracle selected from among other choices? What were our other choices?

    Several issues were considered in deciding what software and hardware platform(s) to use to support the ONE database and web applications, key among them: cost, performance, and expertise.

    Cost: the Oracle products were cheaper for us than other alternatives we considered (e.g., Cold Fusion with either an Oracle DB or a Microsoft DB on a Windows NT platform; Microsoft's IIS with Active Server Pages on a Windows NT platform) because the Oregon University System (OUS) already had a full license for them in support of other projects, and we already had the necessary hardware platform (Sun E250 server running Unix).

    Performance/Expertise: the Oracle DB and Oracle Application Server scale much larger than does Cold Fusion on an NT platform, and Cold Fusion on a Solaris (Unix) platform was untested at that time. And at OUS's Information Technology Services (ITS), we have been running the Oracle DB and Oracle Application Server for quite awhile, so the investment in training and expertise had already been made, which would allow us to begin application development immediately instead of having to ramp up on a new platform or technology.
  2. Describe Oracle: Who owns it? What is it? What does it do/permit us to do? How does it stay current? What requirements does it take technically to operate?

    Oracle Corporation is the world's second-largest independent software company. Oracle's products include Oracle database, development tools, and Internet-enabled business applications for a wide range of customer needs.

    ONE is using two main products from the Oracle Corporation in support of our database and web applications: Oracle Relational Database Management Software (RDBMS), and the Oracle Application Server (OAS) software:
    • Oracle RDBMS is database software, similar to Microsoft Access, only much bigger and better for this purpose. (Oracle RDBMS is designed for enterprise computing instead of PC computing.)
    • The Oracle Application Server software is a set of tools that allows a business to develop and serve up applications that use data from an Oracle database over the web.

    Oracle provides updates and upgrades for both products as needed to licensed customers; this is the case for OUS.

    Technical requirements are a hardware platform to run the software (we're using a Sun Unix machine), plus the following areas of expertise:
    • Oracle RDBMS Database Analyst (DBA) skills to build and manage the database
    • programming skills in PL/SQL and PERL (the programming languages we're using to build the applications)
    • web server management skills
    • proficiency in configuring the OAS software.
  3. What types of staff expertise are needed to develop and implement a common course/program database?

    Experience in projects that require sharing data between/among diverse sources is necessary to plan and implement data-sharing from the participating campuses, as are good communication skills for cooperating with the campuses. Project planning skills are necessary to drive the project and keep it on track. Web programming (CGI) expertise and expertise in whatever programming language(s) are going to be used for application design are necessary. Some database design skills are necessary. Web page and web site design is necessary and, depending on how many graphics are being used, graphic design skills are needed.
  4. How long does it take to develop a database realistically? Which specific tasks within building the database take the most staff time?

    Since the database refers to the underlying software that stores the actual data, the question perhaps should be: how long does it take to develop the database, develop the feed applications that supply data to the database, and develop the web applications that display the data from the database?

    Once the data requirements are designed, the actual database design takes only a few hours. Defining the data requirements is the process where an analyst has to come to an understanding of what types of data elements are needed, how big they are, and relationships between/among them. This has to happen before anything else and can take weeks of discussion with the users and data providers. In ONE, this was the primary role of pilot institutions, to provide this type of information.

    After the data elements have been defined and the underlying database has been designed, then the processes of creating data feeds and building data display applications can move forward. This involves figuring out how data is going to get into the database and designing the instructions and programs to facilitate this. The most time- and labor-intensive part of the process is the web application design and programming. This is what one ITS staff person has been spending so many hours doing: designing and programming applications that read the data stored in the database and formatting it into sets of web pages for display, according to how we've decided the web pages should look and with all the functionality we want (e.g., searches and sorting).
  5. What does Oracle cost and what other costs are associated with building this database?

    The largest expense with ONE is personnel. Next, is the cost of the Oracle software (except OUS already owns this). Oracle products are expensive. For example, the license for the RDBMS (Enterprise edition) cost $4,500 several years ago, plus costs us $3,000 to $4,000 per year for support; the OAS license cost around $10,000, plus $2,000 to $3,000 per year for support. If we didn't already have a hardware platform for the ONE project, that would have cost $5,000 to $10,000, depending on how much capacity was needed (for this project alone, this probably would fall at the low end of the estimate). There are also costs associated with the support infrastructure but those are difficult to estimate (e.g., Internet connectivity costs, server support costs).
  6. What have we learned about building a database with entries from multiple campuses (and campuses from different sectors); for example, problems with Banner, need for multiple methods of entering data?

    This has reinforced what ITS staff already knew about the difficulty of coordinating between different informational and political systems. Campuses have different data sets, different personnel resources and different administrative, political, and social constraints on how they operate. Trying to interact between these organizations has been a challenge.
  7. Is there anything we would do differently knowing what we do at this point in the project?

    Bearing in mind that we're not done yet (we're only one year into the ONE project), the biggest lesson is we should have planned for two data-entry methods from the beginning. For campuses with a lot of courses and programming expertise, we needed to provide a data-feed method so that they can develop an automatic feed to our database, which was our initial approach. For small campuses with few courses and few resources. we needed to plan an easier way for them to enter the data without having to resort to (often non-existent) programming staff. The Microsoft Access database we handed out addresses this group, but isn't a very good long-term solution. But, designing a full-blown web data-entry application (a la Western Governors University approach) would have been a burden on our development timeline, which is the primary reason we insisted on a single entry method. If we'd better understood the difficulty that small campuses would have in meeting our data needs we might have been able to engineer a better solution from the beginning.
  8. Is it a good idea to have prototype campuses to go with, in first developing the database; i.e., is it best to start with some, and have others follow later?

    It was absolutely essential to start with some pilot campuses. The communications difficulties would have been overwhelming if we were trying to design the database and course feed in conjunction with all the campuses at once.
  9. What other approaches to building searchable databases are used nationally that you know about? Is our approach similar to others?

    In terms of similarities or differences, one of the key steps we used in designing the web application "look and feel" was to visit other sites that were doing a similar thing (distance education classes on the web) and analyze which aspects of their sites we liked or disliked. This helped shape how our site is shaped and how it functions. One model we copied quite a bit from was the Southern Regional Electronic Campus. But there were elements of several others that we used. And of course we're always paying attention to searchable sites, no matter what kind, to learn new things from how they do things and we incorporate those as appropriate.

  10. PART TWO: After a Year of Public Operation

  11. After nearly a year of operation, what is your evaluation of our choice of an Oracle database (cost, performance, expertise)? Have there been any problems that would have been better handled by another choice?

    Everyone has a bias for certain technology based on his/her experience with or ownership of the product. Given that the Oregon University System (OUS) has license to the Oracle software and is using it extensively for Banner and other applications, using Oracle for the database is still the best choice we could have made. There may be other systems that use a more mainstream programming language that is more web-suited, but we've had no problems with database performance. Oracle tools and PL/SQL are very stable.
  12. We have campuses submitting course data drawn from a back-end database, campuses who are using a Microsoft Access template to submit courses, and a few campuses who are hand-entering limited numbers of courses using our Course Edit tool. What are the issues/problems (for both the campuses and ONE staff) with each of these methods?

    Database-based submission (exporting from a centralized pool of data): In order to develop a common searchable database to merge data from multiple campuses, ONE staff and our pilot institutions defined the standard of the file feed. It then becomes the responsibility of individual campuses to match that standard. This generally requires campus Information Technology staff to write a program to extract the data matching the ONE specifications. Once this program is operational, very little time is involved in any future data feeds. This is the most automated process of the three and should theoretically be the most trouble-free. However, we encountered difficulties because some campuses employ customized coding symbols that require stripping out to get clean data, and slight dissimilarities in data field definition caused inaccurate search results for a few campuses. Fixing these problems generally requires a one-time revision of the campus data extraction program.
    Microsoft Access submissions: For campuses without a centralized database, we developed a Microsoft Access template to allow them to use a PC to get the data into a format from which they could export course data to the ONE database. Using the Microsoft Access template requires hand-entering (or cutting and pasting) data, so we expected that campuses would use it only if they had few courses to enter. However, we discovered that a few campuses with a centralized database and substantial numbers of courses used the Microsoft Access template in order to enable distance education staff rather than overworked technical staff to handle the data submission.
    Course Edit Tool submissions: The Course Edit Tool was developed at campus request in Fall 2000 to enable campuses to individually correct inaccuracies in their course data without having to submit an entirely new data feed. Campuses with few courses to submit can also use the tool to submit course data into the database. Its drawback is that all data must be hand-entered. Since it is a web-based tool, however, it is easier for non-technical staff to use than the Microsoft Access template.
  13. Are there any other general data-handling issues?

    One issue is the technical expertise of the person(s) submitting the data. The ONE technical staff have never experienced problems when dealing directly with campus-based Information Technology personnel who understand data file protocol. Technical expertise among distance education staff, however, varies greatly among the campuses; for example, some had great difficulty using Microsoft Access and exporting files. When it is necessary for our technical staff to walk someone through the data submission process over the telephone, having personalized desktop set-ups or different versions of the Microsoft Access software complicates the process. It is much easier with the web-based Course Edit tool.
    A minor difficulty is also caused when a few campuses submit their data completely in upper case (most use title and lower case). Since upper case fonts take more space, this results in course titles sometimes being truncated, which lowers readability. We recommend that campuses submit data in lower case or title case, but do not require it.
  14. Now that the ONE database and website are operating, what will it take to maintain them? (server capacity, programming, automating certain routines, personnel) What future problems or concerns about the database/website do you foresee?

    Hardware/Software. The advantage of using hardware and software that you are already maintaining for other applications is that adding one more application generally adds very little to the time or price of maintenance. Our server has sufficient capacity that a substantial increase in visitor traffic and courses in the database should not cause any problems in the future.
    There are always systems administration concerns when you share server resources among multiple applications. Upgrade needs of another application may force you to go through the same data migration (with the resultant need to performance tune programs again) simply because you are sharing the same machine. We do have a development machine to test out new programs/routines, but that doesn't mean that when a program is migrated to the production server that it will be trouble-free, since it now has to operate in a different environment because of the other applications on the server.
    Technical Staffing. Developing and testing the database and the Course Edit Tool, adding user statistics compilation, assisting campuses in the data submission procedure, and optimizing the database required approximately 1,200 hours of technical staff time over the two-year period. Additional programming or changes to the database will require substantial technical staff time, partly because any major changes to the database, generally requires re-beta-testing the entire site since a major change will affect many subroutines. A major change for 2001 will be the addition of key word search capability. We will also be adding a program to track user exits to be able to report ONE-generated traffic to our participating institutions. As campus program staff become more familiar with the Microsoft Access and Course Edit tools, they will require less assistance, which will lessen our technical staff responsibilities. We expect to spend approximately 350 hours of technical staff time this year to develop a key word search and an exit-tracking tool, and to provide routine technical support for ONE.
    Related Staffing. ONE program staff and the ONE Steering Committee with input from Information Technology Services determined data elements, search categories, timelines, and conducted an evaluation of the prototype site. ONE program staff also developed the content of support pages and created more than 80 informational static web pages.
  15. What routines and procedures have you created to assure smooth operation of the ONE database and website?

    • We are using two servers: one for the database and one for the static web pages, so there are subroutines that we had to write to integrate the two for user results. Generating usage statistics or checking log files also requires information in the database to be merged and then presented on the other server, so we had to write routines for this.
    • We asked campuses to submit their course data to us at the same time as they must submit it for their campus schedule in order to ensure that courses for upcoming term(s) would be available.
    • We also made a decision to "age" course data on four dates throughout the year so that no courses older than the previous term will be actively displayed (although the data remains until supplanted so that a campus can update a course using the Course Edit tool).
    • We use linkbot software monthly to search for broken links, which are then repaired or removed.
    • ONE staff keep track of information that needs periodic updating (e.g., campus calendars) and ask campuses to provide this and any updates of campus or program information on the ONE website. Every six months we also examine all static pages for out-of-date information.

  16. What potential improvements to the ONE database would make it more efficient or effective?

    Our database analysts examined the database parameters and made some changes that nearly halved search time. It is good practice to have a database analyst examine the database approximately twice a year to keep it running as fast as possible (size of the data packets, how the data are growing, where you're putting the data, estimating/forecasting growth, etc.).
    A planned key word search tool will offer the user the option of searching by course designator or course number, which is currently unavailable. It will also offer an alternative to the format of the five course searches ONE is now using.
  17. How is down-time for the server that supports ONE handled? How do potential users know when the server is not accessible?

    Oracle databases are always active, so we must back up the information nightly in order to have an accurate record for restoration purposes. The database that OregonONE shares with many other OUS projects is backed-up from 2:00 a.m. to 4:00 a.m. six nights a week. While this happens, users trying to access the database will receive a message informing them that the server is being backed-up and when it will be online again. This only affects the dynamic database; the server with the static web pages is still accessible.
  18. What has been the feedback from users about ONE website design, information contained, and effectiveness of searches/search categories?

    One aspect of our beta-testing of the prototype ONE site was to have faculty, staff, and students from participating campuses fill out a survey form evaluating the website for ease and effectiveness of course and program searches. We also asked for input on attractiveness, navigability, ease of use, information included, etc. We used this feedback to redesign various details in the database and website prior to and immediately after going public on March 27, 2000. Since that time, we have received numerous positive comments from users. The only negative feedback resulted from inaccurate subject searches caused by a few campuses using non-standard data field definitions.
    Two improvements have been requested by campuses: a Course Edit tool (instituted December 2000), and a Key Word Search (planned for implementation in 2001).
  19. Can the ONE database support research, e.g., user statistics, gap analyses, other?

    A database is an ideal tool for generating data to support research, though you may need a program specifying which data you want and how you want it displayed. ONE is already tracking user statistics at http://OregonONE.org/stats.html. In 2001, we will add a program to track exits so that we can report ONE-generated traffic to participating institutions.
    OregonONE supports searches by institution, subject area, level of study, delivery method, and term for courses and subject, institution, and degree level for degrees. To do a gap analysis of courses, you would have to choose the desired subject area (+ other delimiters if desired), and then examine the output to see which required/desired courses are unavailable. So, you can use the database for the search, but the individual must make the analysis.
  20. With the benefit of hindsight and the knowledge gained from the entire experience of planning for, creating, and operating the ONE database, what might you have done differently?

    The Microsoft Access template is not very user-friendly, and now that we have the Course Edit tool, we probably don't need it. However, developing procedures for data feeds (both database and Microsoft Access) benefited the development of the Course Edit tool, so it was a necessary but currently redundant step. Eventually we will probably drop the Microsoft Access template because it is complicated and support must be updated with each version of Microsoft Access.
  21. If we add a new component to ONE, e.g., a faculty "help" desk, how will this affect the database?

    Adding informational content about online and distance learning, distance education pedagogy, instructional design tips, articles about lessons learned, "how-to" handbooks, etc., will not affect the database at all since these are static pages. Reworking the homepage by a graphic designer, however, may be necessary.
    If ONE decides to host faculty discussion forums, set up a listserv, or provide an interface to display exemplary courses for other faculty to examine, that would require necessary software, but it still would not affect the database. If ONE wanted to catalog the articles, threaded discussions, etc., and make them key word searchable, this would add a major piece to the database and would require extensive programming.
  22. In the future, we may wish to partner with other states to create a regional clearinghouse of courses (similar to the Southern Regional Electronic Campus). What problems, if any, do you foresee in the creation of such a database?

    Any time you want to merge data from two or more databases, you need prior agreement on specific data standards and parameters. Unless a regional consortia decided to adopt ONE data and field standards for the database, we would have to modify our data for a feed into a regional database. It would be a major project.
  23. Are there any ongoing "domain" issues?

    We had hoped to register OregonONE as ".edu" because ONE represents the courses and programs of many colleges/universities. Since ONE does not actually issue degrees, however, we were denied the use of ".edu" and consequently chose ".org".
    Once we decided on the domain name OregonONE.org, we needed to affirm that no one else was using that name and then apply to one of many domain registrars (we used Network Solutions) to register the name. Fees for registration vary depending on the registrar; Network Solutions charges $70 for every two years to keep the registration current.