Bridging the Digital Divide Using Personalized Service Level Agreements in the Internet of Things

Cathryn Peoples
November 11, 2020


The digital divide lives on in 2020 and many who could benefit from the Internet of Things (IoT) technology are simply not able to. Consider the elderly population as an example - almost half of the over 75s do not even use the Internet, never mind the IoT. This creates a problem in the modern world, given the range of services that are only being made available in an online capacity through the IoT.

Cost efficiencies can be achieved through operating such business models, but without all potential users on board with the approach, it is not one that can be exploited. The more traditional, paper-based approaches, therefore, need to continue to be used instead.

Why is this the situation we live in today? Why does the digital divide persist when internet capability is a critical backbone of society? One important reason is the complexity of service setup processes - several would-be users, making particular reference to those who are older, do not have the mental and/or physical capacity to establish a connection to online services. This is not only applicable to citizens who grew up without the internet, but also for those who used online technologies and services in their earlier lives – the evidence suggests that capacity to use technology declines with age, as discussed in [1]. Service setup procedures are largely manual, and interaction is required with a service provider to establish a connection; this is prohibitive for users without technical awareness or expertise.

Another reason for a lack of connection to the IoT is the financial cost of service provision. To participate in the online world, it is necessary to have at least one online device, which may have required a setup connection cost, in addition to either a monthly or pay as you go subscription rate. In relation to maintaining a connection, there is generally a limited range of services that customers may purchase - from an individual provider, there may only be three service tiers. In saying that IoT services are tiered, this refers to the fact that there can be different service packages available from ISPs, each of which varies in the quality of the service provided, the responsiveness of the support team, and the associated cost; a selection of these are summarised in [2] (Table 3). Customers, however, could have more widely-varying service needs than can be responded to using a restricted number of service tiers. It could therefore be natural to assume that there will be a similarly diverse set of service provisions to enable them to access the IoT in a manner that suits their needs. This, however, is not the case.

As a result, research has begun by the author to optimize the service provisioning process, through automating service setup and making services available which are tailored more closely to customer needs. The proposal is to achieve this by taking into account a selection of personal user characteristics, which are subsequently used to influence the service provisioned in ways not defined before. These personal characteristics capture detail on a customer’s tolerance of ad hoc changes to their service provision. This allows cost efficiencies to be exploited, which is an important factor in helping to bridge the digital divide.

With the customer tolerances known, the intention is to use these when considering service provision and management. Quality of Service is the aspect traditionally taken into account when defining service agreements, with an assumption that a customer will have subscribed to a service that meets their needs - if the level of service promised is provided, the needs of the customer will be responded to. However, depending on individual personal characteristics, customers can tolerate lower Quality of Service than is provisioned in the SLA. This is information that can be exploited in the service provisioning process, yet is not done at present.

In support of this mode of operation, it is therefore proposed that the process of service provision and management is achieved using a data ontology designed for the IoT – the model proposed in this research will be published in January 2021 [3]. An ontology in general describes a hierarchy of attributes, with leaves branching from higher sections of the tree in association with each category of data. Many ontologies for the IoT exist. However, the limitations, as discussed in [3], include that each ontology is typically provisioned for a specific domain or to respond to a particular operational challenge; a single ontology addressing the needs of all scenarios across the IoT does not exist. It is therefore to this gap which our work responds.

The objective of the ontology proposed is to allow sufficient data to be collected such that a Service Level Agreement (SLA) can be provisioned which fulfills customer needs. The proposed ontology captures Customer, Device, Attribute, Dataset, and SLA branches, with leaves including Risk (true/false) from the Customer perspective, and Interruptible (true/false) from the Attribute perspective. Both of these metrics can be used to identify if a customer can cope with a somewhat disrupted service – cost efficiencies can be applied for more flexible customers, matching their tolerances with the service provisioned, as opposed to the more traditional Quality of Service expectations. This approach applies a more personalized strategy to SLA provisioning.

The personalized SLA will be customized in both the service which a customer can tolerate, in the worst-case scenario, in addition to a service that a customer can financially afford. When a customer wants a completely reliable service under this model, they must expect the associated cost overhead, without efficiencies, to be exposed to this. When the customer can tolerate a less than perfect service, they can benefit from the cost efficiencies as a result of possible inconvenience, albeit an inconvenience they can cope with and have agreed to.

The digital divide exists where citizens want to access services, but are unable to. Thinking about our focus population, the elderly, we can remember that just because older users are not online, this doesn’t mean that their desire is not there. They are effectively excluded from a significant aspect of our society today, and the evidence suggests that this is having a negative mental impact, as discussed in [1]. Not being able to afford a service is one thing, but being able to afford it yet being unable to access it is another. There is an opportunity to adapt how services are provisioned so that the value which results from participating overcomes the costs - practical and financial - that are involved in setting a service up, and it is to this which the author’s research seeks to contribute.

This work is being carried out at the BT Ireland Innovation Centre in a partnership between BT and Ulster University as part of a five-year project, funded by Invest Northern Ireland [4][5].


  1. Peoples, C., Moore, A. and Zoualfaghari, M. (2020). A Review of the Opportunity to Connect Elderly Citizens to the Internet of Things (IoT) and Gaps in the Service Level Agreement (SLA) Provisioning Process. EAI Endorsed Transactions on Cloud Systems, 6(18), 1-8.
  2. Peoples, C., Rabbani, K., Abu-Tair, M., Wang, B., Morrow, P., Moore, A., Rafferty, J., McClean, S., Zoualfaghari, M. H. and Kulkarni, P. (2019). A Review of IoT Service Provision to Assess the Potential for System Interoperability in an Uncertain Ecosystem. IEEE SmartWorld, 1964-1971.
  3. Ed. Khan, M. A., Algarni, F. and Tabrez, M. (2021). Smart Cities: A Data Analytics Perspective. Available at:
  4. BT Ireland Innovation Centre Homepage. Available at:
  5. BT. (2017). BT Chooses Northern Ireland for £28.6 million innovation centre and creation of 50 graduate jobs. Available at:


Cathryn PeoplesCathryn Peoples received a B.A. degree in business studies with computing, an M.Sc. degree in telecommunications and internet systems, and a Ph.D. degree in networking from Ulster University, U.K., in 2004, 2005, and 2009, respectively. She is currently employed as a Research Associate at Ulster University working on Internet of Things (IoT) research. Cathryn is also employed by The Open University in the School of Computing and Communications within the Faculty of Science, Technology, Engineering & Mathematics as an Associate Lecturer in Software Engineering. She became the co-Editor-in-Chief of the EAI Endorsed Transactions on Cloud Systems in January 2020. Her research interests include cloud management, cross-layer protocol optimization, delay-tolerant networking, smart cities, and green IT.