Quality of Service for Edge-to-Cloud Continuum

Section 1. Workshop title
Quality of Service for Edge-to-Cloud Continuum

Section 2. Workshop topic
Current approaches on edge computing are not enough to address the forthcoming massive usage of edge computing in the frame of large Internet of Things (IoT) deployments. Essentially, the main goal in such scenarios is to ensure that the overall offered Quality of Service (QoS) fits the application needs over the edge or edge/cloud continuum deployments.
Processing speed and latency issues have been identified as the top barrier in this domain, while cost and reliability (meeting the provider Service Level Agreements - SLAs) are the top and second most important factors for evaluating edge and cloud services.
At the same time, network capabilities also play a very important role. Especially performance may hinder the adaptation of edge solutions for a variety of applications (e.g. in the Virtual/Mixed Reality applications), for which timely completion is key for the productivity of the company.

With this regards our project Pledger aims to research these topics and provide advancements using:

  • Rich data sets for better Machine Learning – Consolidating data in the cloud for sophisticated Machine Learning applications.
  • High Performance Computing – Leverage the Edge for computationally intensive analysis or machine learning.
  • Security – Trusted and secured end-to-end communications
  • Blockchain – Bringing Smart Contracts to the Edge.

Overall effective management of QoS in the Edge to Cloud continuum will impact a diversity of roles in the European Cloud market:

  • Edge computing providers, to enhance the stability and performance effectiveness of their infrastructures, through modelling and understanding of the overheads and optimal groupings of concurrently running services, runtime analysis and adaptation, thus gaining a competitive advantage
  • Edge computing adopters, to understand the computational nature of their applications, investigate abstracted and understandable QoS metrics, facilitate trust and smart contracting over their infrastructures and identify how they can balance their cost and performance to optimise their competitiveness and monitor their SLA, and,
  • Third parties to act as independent validators of QoS features in IoT applications, enabling new decentralised applications and business models, thus filling a large gap in the emerging edge/IoT computing market landscape.
    In addition, it is important to remark that in many cases, public authorities, especially in the case of procurement evaluation, may experience difficulties in ranking providers or comparing the abilities of the latter in terms of non-functional attributes (such as performance) that might be evident after the deployment of the actual service. Thus, a ranking mechanism can be of use, especially when coupled with formal verification of a provider’s history in QoS and SLA upkeep.

Section 3. Guiding questions

  • Edge/Cloud SLA monitoring tools
  • Edge/Cloud off-loading and trade-off techniques
  • Edge/Cloud computing benchmarking
  • Blockchain deployments on edge computing and IoT
  • Privacy, Trust and Security on edge computing
  • Network Slicing in 5G networks
  • Use cases in Edge QoS preservation: Smart city, Autonomous vehicles, AR/VR and Industrial IoT (IIoT),

Section 4. Contribution to R&D challenges
Service Level Agreements play a key role in Cloud computing market by being the mechanism that users have to enforce guarantees around performance, transparency, conformance and data protection when using Cloud services. While still these rich SLAs and its management techniques in single provider scenarios are not yet widely available in Cloud commercial providers, the consideration of diverse Cloud offerings, and specifically Edge computing pose higher levels of complexity to the issue. Fault tolerance, availability and performance are service management aspects still to be widely addressed in Edge computing. These are of specific interest due to the nature of Edge resources and their reliability. In the area of application off-loading among Edge and Clouds, typically resource management processes observe time, data, cost and energy management metrics. Edge QoS Latency, Cost, computation, data and application management, as well as, energy optimisation are key aspects still to comprehend to fully develop the Edge and Cloud markets. In addition, there is the need of research in QoS management in Edge environments that considers frequent loss of connectivity and low bandwidth as well as heterogeneity in resources and integration of special devices including hardware accelerators. This research is an important advancement towards the uptake and commercialisation of Edge-to-Cloud developments, exemplified by Pledger’s commercially-driven use cases in Manufacturing, Augmented Reality/ Virtual Reality (AR/VR) collaboration and Vulnerable Road User (VRU) Protection.

Section 5. Market and societal impact
The availability of tools for edge computing stakeholders and edge resource users that bring the ability to better monitor, measure and predict the dynamics of edge computing computation based on QoS and QoE particular requirements of the application has the potential to benefit all stakeholders in the Edge / Cloud value chain. Overall, the development of such research will boost the capacity of European innovators and industry to make use of high-end edge / cloud computing infrastructures that can support business applications with new features of QoS and QoE. It further facilitates adoption by commercial users (exemplified by use cases in Manufacturing and CAD design collaboration) as well as public organisations that impose specific procurement processes, exemplified in the Smart City/VRU protection Use Case. The latter case improves safety of citizens in the era of connected cars and illustrates that SLA monitoring mechanism can play a crucial role for our personal safety in case of critical services with latency constraints.

Section 6. Contribution to Sustainable Clouds, Cloud Federations, Cloud-To-Edge Computing, and EU data spaces
QoS and effective Edge / Cloud continuum service management are key research areas in the areas of Cloud Federations, Cloud-To-Edge Computing and can constitute a distinctive aspect for EU data spaces cloud and Edge infrastructure management.