Submission

Submission to the Review of the Use of Information Technology in Commonwealth Agencies

Preamble

The National Library of Australia is a statutory authority which provides a range of library and information services to organisations and individuals throughout Australia. It has provided computer-based services through communication networks since the mid-1970's, and currently operates the Australian Bibliographic Network and Ozline services to more than 1,300 organisations, including all Australian universities, State and territory libraries and a large range of organisations in both the public and private sectors.

Introduction

To provide its range of computer-based services, the Library has operated a small-to-medium mainframe-based operation since 1977 with an IT organisation of between 30-50 staff during that time. The main systems operational on the mainframe installation have been the Australian Bibliographic Network (ABN), Ozline and a financial management system. In recent years, along with many other organisations the Library has moved towards the use of open systems on smaller computers when developing new applications. This trend will be continued in the Library's current major exercise, the NDIS Project, a joint development with the National Library of New Zealand to replace the present ABN and Ozline (and their New Zealand counterparts NZBN and Kiwinet). The system being developed in the project by CSC Australia as the Systems Integrator will be based not on a mainframe running proprietary operating system software, but on a powerful smaller computer utilising massively linked microprocessors and running a Unix operating system.

Other software which presently uses the mainframe system will be migrated to other platforms, either microcomputers used by individual staff or minicomputers supporting multiple users connected by the Library's general purpose local area network. The other major development in recent years with which the Library has been involved has been the introduction and growth of the Internet to the general community of government, business and individual users beyond its initial use in academic institutions. The Library began to use the Internet as one of several telecommunications options for supporting its remote users, but has within the past two years been an active advocate of its potential for more widespread dissemination of information by government and other organisations. This advocacy has been demonstrated by a number of submissions to reviews such as the Broadband Services Expert Group (BSEG), the ASTEC inquiry into Research Data Networks, and the Telecommunications Policy Review.

Distribution of government information

The benefits to the Government of the use of such a wide area network will occur in the areas of health; education; environmental data; library, information and cultural services: and individual transactions with Government. The mechanisms for government action in this area are based on the premise that effective access to government information and other electronic services must be viewed from the potential client perspective rather than that of particular government agencies. In order to make access by the public as easy as possible, the delivery mechanisms must be coordinated across levels of government as well as across different Commonwealth agencies. This requires that the Commonwealth Government takes the lead in influencing such developments, and the Library applauds the recent establishment of the Commonwealth Internet Reference Group, whose initial tasks comprise the collation of information on Internet initiatives in Commonwealth agencies, and the preparation of guidelines to assist agencies in further activities. In a more directed manner, there is a need to establish a central point for access to the distributed electronic resources of the Government. The National Library of Australia has already made a start in this direction with its Information Server, based on WWW technology, where pointers to available Government information will be stored. More needs to be done though, and the Library proposes that some intitiative (such as a joint new Policy proposal) be developed for the establishing in 1995/96 of a "open government pilot" project providing access to distributed electronic resources throughout the Commonwealth Government. Such a project could involve the Department of Finance, CommTel, the National Library and the Australian Government Publishing Service. The intent of this project would be that the Government would demonstrate its leadership role in relation to electronic information and services delivery.

The Library therefore recommends that:

an initiative be developed for the establishing in 1995/96 of a "open government pilot" project providing access to distributed electronic resources throughout the Commonwealth Government

At the same time, the Government must work with the State governments to develop a strategy for more coordinated delivery of government networked services to the community. This will be a prime responsibility of the Commonwealth/State Internet Working Party, set up after the recent Government Telecommunications and Technology Conference. However, there are other developments in this area, such as a current initiative to use public libraries throughout Australia for network access to the services of a number of Government agencies.

Directories and Network Support

In the emerging networked environment, Government attention needs to be given to government-wide (both State and Commonwealth) electronic directories and addressing. Some of the issues involved will include:

  • electronic records management matters such as file naming and indexing, and standards such as Universal Resource Locators (URLs)
  • initiatives such as the U.S. Government Information Locator Service (GILS)
  • the role of the Government in developing X.500 directory services

There is also the issue of providing support for the large number of Australians who will want to use the network, but will be baffled by the wealth of information available, in which the desired items will be hard to find. This may be addressed in part by the Australian Vice-Chancellors Committee's initiative for an AARNet Information Service (AIS), but it is clear that something larger, even if it is decentralised will be needed and this is also where the Commonwealth Government is well placed to take a lead.

IT Policy Framework

The Library has always strictly observed Government Purchasing Policies, despite some uncertainty from time to time as to their applicability to statutory authorities such as this organisation. Value for money, and open and effective competition have been important criteria in our systems acquisitions. Other policies such as GOSIP have also been adhered to - the Library having installed an OSI-compliant LAN in 1989 to support access by its staff to its internal computer systems. The Library is though concerned about the inherent conflicts between value for money and policies for the support of the indigenous IT industry. This has been obvious in the NDIS Project. Starting with a feasibility study in the late 80's, the Library acknowledged the need to redevelop the software which supported its ABN and OZLINE systems. This software was developed in the US during the 1970's and is a mixture of specialised applications based on the data base management systems of that period, and proprietary software supplied by a mainframe systems company. ABN and OZLINE are run as commercial undertakings by the library with current annual revenue of over $6m, and are under pressure from their clients to continually improve their services. These improvements cannot be delivered adequately using the current systems, so the redevelopment decision was taken. The National Library of New Zealand uses the same software for its services and was facing the same difficulties in providing the services required by its clients, so discussions were initiated between the two organisations to determine whether a joint development was feasible. This resulted in detailed planning commencing in 1992. It was soon evident that although the two libraries were convinced of the benefits of the joint development, the different purchasing policies of the Australian and New Zealand governments would provide some difficulties. This was mostly due to the Australian government's policy on the use of the Systems Integration Panel for acquisitions of more than $1m. Whereas the Australian government purchasing policy is aimed at "support for Australian and NZ industry" the interpretation of the policy by the then Department of Industry Technology and Commerce militated against the involvement of NZ companies or those proposed by the National Library of New Zealand. Although the matter was eventually resolved by discussions at Departmental Secretary and Director-General level, it did cause a delay of some months in the whole process. Clearly, as the agencies of the two governments proceed to work more closely together in the context of Closer Economic Relations (CER), there should be more attention paid to this sort of cooperation to ensure that the interpretation of policy between different departments does not impede such developments. Having resolved this problem, the libraries were then able to proceed to tender to a agreed short list of companies on or associated with the SI Panel. In line with Government policy, it was necessary to circulate the draft Request For Tender (RFT) for industry comment and then to take any comments into account when preparing the final tender. The draft was circulated in August 1993, the final RFT was issued in October 1993, the evaluation was completed and Ministerial approval was obtained in September 1994, and work commenced on the first stage of the project in November 1994. The project is expected to be completed early in 1997.

This process is unnecessarily lengthy, and could be improved (with benefits coming to Government agencies and the IT industry) by the adoption of a system which is being used on a trial basis in the Canadian Government. In this system, large projects such as NDIS are broken into phases, and companies are asked to tender for one phase at a time. Thus, once the basic requirements of the system have been determined (as had been done by the libraries in mid 1993), tenders would be called for an Analysis and Design Phase, with very strict limits on the time taken to evaluate tenders (forty days is a typical period). After this phase was completed, the specification would be used in a tender for the Development Phase, to be followed by the Installation Phase and so on. While it would be difficult to say exactly how much time such a procedure could have saved the NDIS Project, a reasonable estimate is that the resolution of the problem of New Zealand industry involvement and the time taken by having to call and evaluate tenders for the whole of the project has probably cost eighteen months. This represents lost opportunities for the Australian and New Zealand IT industry, the two libraries, and the clients who will benefit from the improved systems.

The Library therefore recommends that:

attention be paid to the implications of both Australian and New Zealand industry participation in Australian Government IT Purchasing; and a more efficient acquisition process be developed for the development of large projects, with particular attention to the Canadian system.

An alternative procedure is that for large projects, a partner for development should be identified, rather than requiring a "best price" approach against a detailed specification.

IT Operations Environment

General matters concerning Outsourcing are considered below, so only the issue of rationalisation is treated here. There is no doubt that there could be rationalisation of computing resources between different agencies, but all of the topics raised in the Issues paper need to be considered. One topic that is perhaps most significant is that the sort of regime that operated in mainframe environments is unlikely to provide the same benefits in rationalisation in organisations which have moved to open systems (this also applies to outsourcing in general). Because of the large costs involved in any use of mainframe operating systems, there were clear benefits in moving small to medium operations to large data centres where the proportionally lower operational costs of very large mainframe resources brought real savings to the smaller user. This is unlikely to be the case with open systems, because of different charging plans offered by suppliers of operating systems and utilities software. In the general area of Government-wide contracting, agencies should not be absolutely required to follow agreements which have been negotiated, because of the continuing change in the environment and the significant real reductions in cost which are occurring all the time. It should be sufficient that an agency demonstrates that it has achieved at least as good a deal as is available Government-wide. Similarly, such agreements should not be of too long a duration, as a good deal today may be unacceptable twelve months later in a different competitive environment. It is also important to note that with an interoperable wide area network, the network becomes effectively an immense single virtual machine. Therefore attention should be paid to the way in which lines, processors and storage facilities can be most effectively shared. This will have significance in the way in which shared back-up and disaster plans are formulated. The heart of the interoperable wide area network is the set of protocols which it will use. In the interest of simplicity, it is easy to argue that a single protocol is to be used. The Government did this with OSI in GOSIP, a move virtually ignored by the industry. Instead, it must be recognised that the TCP/IP protocols have a place in the network even if accepted OSI products such as X.400 continue to be used for specific purposes. In the same area, it is necessary that some major agencies abandon the assumption that the use of TCP/IP is an interim arrangement on the way to OSI. In the Library's view this can no longer be assumed, as the TCP/IP momentum, now that it has spread into the commercial sector, is such that a long-term TCP/IP based future is assured.

People Issues

The Issues paper refers to the changing needs of the users of IT in government agencies, especially noting the increasing sophistication and computer literacy of these users. While these needs and the way in which IT staff cope with the continuing changes merit attention, another aspect of education for IT staff should also be considered - the matter of client focus.

Traditionally, a significant number of people coming into the IT industry have been attracted to the 'problem solving' aspects of the work, and those with 'people skills' have been in a minority. Because of the increasing need to have IT staff interacting with their users rather than living in the back room, it is important that the education received by IT practitioners take account of this need for client focus. While this education is the responsibility of the agencies employing the IT staff, it should also comprise a part of the formal education provided at tertiary institutions.

The Library therefore recommends that:

'client focus' be included in the curriculum of IT courses in tertiary institutions.

Another concern is that, as stated above, this is a time of particular change in the ITT environment. The need for re-training of Government IT staff in Unix, client-server methodologies, Internet concepts and so on, will require a great deal of expenditure. A whole of government approach to this education would provide benefits to all agencies.

The Library therefore recommends that:

a 'whole of government' approach to funding education in new concepts and methodologies be pursued.

Outsourcing

In line with Government policy, the Library has tested the market for outsourcing its systems during the past three years. The ABN system was not considered for outsourcing because of its imminent replacement, and the Library believed that the limited term of the possible outsourcing would not justify the effort involved in an operational change of this nature.

The main system considered for outsourcing was the Library's financial management system, which runs on the same mainframe systems as ABN and OZLINE. Although the bids received from outsourcing agents for running the financial system were no doubt competitive with the notional costs of the internal operation, the Library could not justify transferring this operation to an external agent because of the fixed nature of the real costs of operating the mainframe system. That is, removal of the financial system would have cost an extra several hundred thousand dollars annually, but the mainframe costs would not have decreased.

This will always be a problem in justifying an outsourcing decision for part of an operation, especially as it is regarded as important to retain 'core' applications as an internal operation, and on the computer system used for these 'core' applications there will usually be capacity for a number of other systems. Outsourcing is therefore best considered on the basis of the total operations of an agency.

The Library has also had experience of insourcing. A small system that operated on a proprietary minicomputer system for several years at a service bureau, until the charges increased to such an extent that the Library had to consider whether it would be more economic to bring the system in-house. This analysis was done and it was found that the capital cost associated with installing the system in the Library would be recouped within eighteen months.

The Library believes that decisions on outsourcing, or insourcing, or rightsourcing should be taken with regard to the real costs of operation of the resulting system.

SUMMARY OF RECOMMENDATIONS

  • That an initiative be developed for the establishing in 1995/96 of a "open government pilot" project providing access to distributed electronic resources throughout the Commonwealth Government
  • That attention be paid to the implications of both Australian and New Zealand industry participation in Australian Government IT Purchasing
  • That a more efficient acquisition process be developed for the development of large projects, with particular attention to the Canadian system.
  • That 'client focus' be included in the curriculum of IT courses in tertiary institutions.
  • That a 'whole of government' approach to funding education in new concepts and methodologies be pursued.

Back to the top