Just as a big ship takes time to turn, the ATM industry is slowly arriving at the idea of centralised management of software applications within data centres. However, the idea is not new. The technology is mature and other sectors, such as finance and retail, have reaped the benefits for a long time. Happily, lessons learnt from these industries are helping ATM progress. First movers have already moved and some ANSPs are well under way with their plans. The recommendation within the recently published Airspace Architecture Study (AAS) to introduce ATM Data Service Providers (ADSPs) as a new category of service providers recognises that the management and storage of data can be separated from the provision of Air Traffic Services.
For those ANSPs reviewing their technology strategies, implementing a data centre may seem a natural next step. However, different businesses have different strategies, which will drive different decisions. Where one is seeking a contingency option, another will want to reduce their costs. The factors to consider before moving to a centrally managed ATM data centre were outlined by my colleague Isabel Franke-Chaudet in an article earlier this year. So, in this blog post I want to cover the different options for implementation, and how they may be assessed.
- Locally installed systems and software: Often used as the baseline, keeping the current ATM system locally stored, with peer to peer networks, may be the simplest option. Complex architecture redesigns are unlikely to be necessary and a new application would be straightforward to install (compared to a centralised data centre). However, when it comes to ease of upgrade, ANSPs need to implement upgrades and changes consistently across units to ensure harmonised processes across units. The main downside is that locally installed systems and software are not an efficient use of storage resources; each service procures and installs its own server equipment and therefore often buys more than it needs.
- Centralised management of software applications on a data centre: This option includes the installation of software applications on "standardised" hardware components at a central facility. Virtual links with control centres (En-route operations) and airports (Tower operations) provide technical harmonisation across units. Operating this model provides maximum flexibility and efficient use of space. It enables the ANSP to scale and flex their storage provision to meet demand more easily. The management of the data centre equipment can be contracted to a third party whose core offering is IT data centre support. On the downside, the design and implementation of this option would be complex and should not be underestimated. If floor space is not available within your current environment, then building a new purpose-built facility or renting space is something to consider. However, finding an appropriate location with adequate security is imperative.
- 3rd party managed service: Contracting a data centre and data processing to a 3rd party allows an ANSP to avoid many of the constraints of hosting and operating their own facility, such as cost, management and maintenance. Using the experience of data centre service providers will de-risk the implementation. The key to a successful solution is in defining a contract with a clear service level agreement with the 3rd party. The contract should also cover responsibilities for data ownership and liability.
- Software-as-a-service: Although not a data-centre model it achieves similar benefits. A cloud-based software as a service provides ANSPs with access to the relevant software as a package through the internet to deliver data processing without the need to own and manage a facility. Data ownership is maintained, but flexibility to modify hardware, or even the software, for changes in user requirement is lost. ANSPs must also consider whether security arrangements in cloud computing are adequate.
It is rare that a single option is a clear favourite and multiple variations of those described above could be viable. In these instances, a range of criteria are needed to understand which option is the strongest.
- Scalability: To what extent will the architecture enable physical hardware expansion without service disruption?
- Flexibility: To what extent can the software be dynamically adapted to meet demand?
- Implementation complexity: How feasible is the transition and implementation from the current situation to the selected option? The amount of tooling required to manage the configuration, integration and testing of some of the options could be significantly challenging.
- Business process changes: To what extent does your organisation need to introduce or change business processes and tools to support the design and operation of data centres, eg adopting recognised IT practises such as those documented in the IT Infrastructure Library (ITIL)? If centralised data storage, or a 3rd party managed service is the chosen solution, there will be no escape from significant changes to internal business processes
- Upgrade complexity: How rigid is the architecture against future developments?
- Target option vulnerability: How vulnerable is the solution to outside influence and/or attack eg natural disaster, cyber-attack etc
- Operating expenditure: How much will the solution cost to operate and maintain? This will include maintenance contracts, utilities, and software licenses, amongst others.
- Capital expenditure: How much will the solution cost to implement? This will include things like communication backbones, building infrastructures, and hardware.
- Connectivity resilience: How is the service level protected against inter-unit communications losses? Eg between a data centre and control centre
- Staff impact: To what extent will the staff be impacted by the move to a given solution?
The importance of each element listed above will differ between ANSPs. Weighting and tailoring the criteria is therefore an important part of the evaluation process to ensure the option chosen is suited to an ANSP's needs. Together, these elements provide the basis from which ANSPs can assess their options.
What is clear is that there is no golden bullet. Each element must be appropriately addressed to ensure future service provision supports an ANSP's business, operational, and technical needs. So, when it comes to deciding if a data centre is the right option, as is being found out by the UK's own parliamentarians, choosing to deal or no deal is not so simple.
COVID-19 and aviation: a step-change towards scalability and resilience
COVID-19 and aviation: protecting our people
COVID-19 and aviation: the re-start
From supply-mesh to supply security – managing cybersecurity in airport operations
COVID-19 and aviation: planning for the ‘new normal’
Preparing for the climate of the future
Three costly mistakes in ATM systems upgrade projects
Proactive vs reactive defence in aviation cyber security
The role of Human Factors in de-risking COTS implementations