Next Generation Cloud

Contributing to Atos innovation strategy with regards to Cloud and Edge computing

Computing is increasingly pervasive in all aspects from our life. Currently three billion people carry smartphones in their pockets: each one is more powerful than a room-sized supercomputer from the 1980s. Still in our collective mind-set computing is related to large data centres, or the laptop we carry with us. But closely related to the emergence of smartphones and our always connected society, computing is starting to be far beyond the smartphones in our pockets, it is increasingly and virtually, everywhere.

Thanks to Moore’s law, we are now at the breaking point in which computers start to be in our phones but also in our televisions, our cars, refrigerators and even dishwashers. And this is not expected to stop in the short term: smart fabrics, connected cars and roads, diverse forms of nano-computing, smart cities and robots will soon be part of our daily lives.  As “connected things” become more and more “intelligent things” there will be the need for these to go beyond rigid basic programming models to become fully networked computing systems capable of delivering advanced behaviours and interacting with their surroundings.

Soon sensors, actuators, robots, drones, routers, servers and cars will only be seen as particular forms of infrastructure elements. This way, computing will no longer be constricted to specific devices but will be virtually embedded and pervasive to everything, enabling an unprecedented computing continuum. 

Cloud computing initially emerged in the space in which we transitioned from an era in which underlying computing resources were both scarce and expensive, to an era in which the same resources started to be cheap and abundant (again, thanks to Moore’s law). Cloud computing has enabled the democratisation of computing, has provided the illusion of infinite computing and enabled the radical acceleration of commoditization of computing opening new possibilities for enterprise users of the technology to position IT as a strategic competitive weapon.

Today we are observing new forms of Cloud Computing, such as Edge and Fog, starting to break the Data Centre barriers so to provide novel forms of computing embracing computing power and data resources increasingly available everywhere. These are forcing existing Cloud computing environments that emerged as part of a centralisation paradigm to evolve to decentralised environments avoiding drawbacks of large data movements and latency, specifically found in IoT scenarios. These new forms of Cloud are making the Cloud concept evolve a more distributed approach in order to lead to better performance and enabling wider diversity of application and services, complementarily to existing multi-cloud and hybrid cloud models.


Gains in complexity of the connected “intelligent things” will pose specific requirements to Cloud Computing evolution. The self-contained and self-sustaining nature of these novel “intelligent things” resources combined with their size and energy harvesting constrains will require of novel computing and communication architectures beyond state of the art today.

One source of evolution that Cloud providers are already taking up is to further exploit hardware heterogeneity.  While initial heterogeneous offerings in public Cloud providers have focussed on delivering dedicated GPUs and FPGAs. Developing cloud computing technology compatible with the management of hardware heterogeneity will call for of finding ways to optimally endeavour heterogeneous special purpose processing units without losing the advantages of abstraction in utility based models such as development of sharing schemes and heterogeneity-aware scheduling at resource and platform levels.

The advent of fully distributed opportunistic collaborative Cloud environments is another source of disruption in order to fully exploit computational power of “intelligent things”. These environments are initially taking for in the evolution of Ad-hoc Clouds enabling smart collaboration among mobile devices. These build routes in ad-hoc networks and opportunistic computing. Further evolution of this concept will enable the creation of dynamic ecosystems, meshes or swarms of “Intelligent Things”. In these, resources capacities are complimented by connection to other objects in a community. These will allow the creation of dynamic eco-systems encompassing “intelligent things”, cyber-physical devices, edge and clouds, each of these adding to the collective capability and insight under the term Swarm Computing.

The main goal of this lab is to contribute to Atos research and innovation strategy with regards to Cloud, Edge and Swarm computing developments, models and architectures. You can find detailed information about the Lab's Computing Continuum vision, and the specific research topics identified for future work in these research areas here.

Main Activities: 

1- Advanced capabilities for IaaS, PaaS and SaaS

  • Accounting and monitoring
  • Autonomic resource management
  • SLA management
  • Multi-cloud environments IaaS, PaaS and SaaS
  • Experimental Facilities in Cloud
  • Eco-efficiency in Clouds and Data Centres


2- Service Management and Engineering: Advanced Service Architectures and SaaS

  • Cloud Service Composition, aggregation and orchestration
  • Service, Application and Data Marketplaces
  • Trust & Reputation Service Management
  • License Management


3- Edge computing

  • Heterogeneous / Things virtualisation management
  • Ad-hoc Cloud management
  • Application offloading
  • Local Cloud Management( Processors accelerations, QoS and traditional Cloud interoperability)
  • Data Management
  • Service Orchestration


  • Cloud Hybrid models: Interoperation, Portability, Federation and Brokerage
  • SLAs, Trust and License Management for Cloud environments
  • Integration of edge and mobile devices into decentralized Cloud architectures for IoT services
  • Energy efficiency in heterogeneous computing environments, Cloud computing and Data centers
  • Big Data Storage and scalability in Big Data processing
  • Cloud-based Experimental facilities
  • Autonomic and self-healing capabilities for Cloud management
  • Cloud Service composition, aggregation and orchestration
  • Cloud Marketplaces, Vertical markets, Added-value services and Applications
  • Cloud Standardization and Compliance
  • Scalability based on predictive models and including heterogeneous resources
  • Heterogeneous and autonomic resource management
Current Research Topics and Findings: 

1- Cloud Service Integration / Multi-Cloud

  • Improved mechanisms for QoS and SLA Managementz
  • Scalability
  • Multi-tenancy / Isolation
  • Cloud Migration
  • Application deployment Automation
  • Improved Cloud Monitoring
  • Interoperability and Multi-Cloud Provisioning
  • Cloud Networking
  • Economy, Cost, Trust and Reputation models
  • Cloud-marketplaces
  • Eco-efficiency

2- Edge (Fog) Computing

  • Edge (Fog) Management and heterogeneity
  • Across Edge execution orchestration
  • Interoperability and standardization
  • Off-loading optimization
  • Edge computing communication topology automation
  • Support for heterogeneous and geographically distributed systems

3- DeW Computing: Where the Edge meets the Cloud

  • Edge / Fog Service Management
  • Admission Control
  • Data abstractions and movement in Cloud and Edge computing
  • Edge Workload management

4- Development areas for Computing Continuum

  • Swarm management among IoE, Edge and Cloud Computing
  • Cyber-physical Cloud/Edge computing
  • Cloud/Edge Robotics



Adaptive Gateways for Diverse Multiple Environments
Edit project Link

Modular hardware and software gateway for IoT with support for protocol interoperability, device and data management, IoT apps execution, and external Cloud communication.


Cloud Brokerage Across Borders For Mobile Users And Applications
Edit project Link

Development of an integrated Brokerage Platform targeting federated clouds in order to support dynamic needs of mobile applications and users.


Business and IT-Cloud Alignment using a Smart Socket
Edit project Link

Introduction of the BPaaS concept that fulfills business process needs thanks to smart alignment techniques, packages BPaaS as “extended Cloudlets” autonomously deployable and including adaptive rules to appropriately react in a multi-cloud environment.


Data-intensive applications Improvement by moving daTA and computation in mixed cloud/fog environmentS
Edit project Link

Its goal is to propose a framework, composed by an SDK and an execution environment, which aims to overcome the barriers that now hamper the adoption of Cloud Computing and increase the adoption of Fog computing by exploiting the full potential of these two paradigms in a synergic way. 


INtegrating Distributed data Infrastructures for Global ExplOitation
Edit project Link

Development of an innovative cloud platform for the scientific community based on open source software and providing access without restrictions to a diversity of e-Infrastructures (public or commercial, GRID/Cloud/HPC).


Towards an Open, Secure, Decentralized and Coordinated Fog-to-Cloud Management Ecosystem
Edit project Link

The mF2C project sets the goal of designing an open, secure, decentralized, multi-stakeholder management framework, including novel programming models, privacy and security, data storage techniques, service creation, brokerage solutions, SLA policies, and resource orchestration methods.


Heterogeneous Secure Multilevel Remote Acceleration Service for Low-Power Integrated systems and Devices
Edit project Link

Development of an efficient heterogeneous cloud computing infrastructure, which can be used to seamlessly offload CPU-based and GPU-based tasks of applications running on low-power as well as more powerful devices over a heterogeneous network.


Symbiosis of smart objects across IoT environments
Edit project Link

Fostering a simplified IoT application and service development process over interworking IoT platforms.


Transparent heterogeneous hardware Architecture deployment for eNergy Gain in Operation
Edit project Link

Control of underlying heterogeneous hardware architectures, configurations and software systems while providing tools to optimize various dimensions of software design and operations.