Periodic Reporting for period 2 - SUPERCLOUD (USER-CENTRIC MANAGEMENT OF SECURITY AND DEPENDABILITY IN CLOUDS OF CLOUDS)
Période du rapport: 2016-08-01 au 2018-01-31
Distributed cloud computing raises many security and dependability concerns, due to increase in complexity and lack of interoperability between heterogeneous, often proprietary infrastructure technologies.
SUPERCLOUD aims to build a security management architecture and infrastructure for secure and dependable cloud of clouds that is: user-centric for customers to define their own protection requirements and avoid provider lock-ins; and self-managed to reduce administration complexity through automation.
Key results are infrastructures for security supervision, data protection, and secure network virtualization - validated through healthcare domain demonstrators: a Laboratory Information System, and an image processing and storage platform.
SUPERCLOUD aims to build a security management architecture and infrastructure for secure and dependable cloud of clouds that is: user-centric for customers to define their own protection requirements and avoid provider lock-ins; and self-managed to reduce administration complexity through automation.
Key results are infrastructures for security supervision, data protection, and secure network virtualization - validated through healthcare domain demonstrators: a Laboratory Information System, and an image processing and storage platform.
SUPERCLOUD proposes a new vision for secure and dependable distributed cloud computing (see IEEE Cloud Computing project paper). The SUPERCLOUD security plane and U-Cloud (user-centric cloud) architectural abstractions enable users to achieve provider independence for security management.
The computing framework includes components for virtualization and security self-management. For instance: the MANTUS/ORBITS virtualization framework to instantiate distributed U-Clouds specified in TOSCA, security services being "weaved" under user control; an extension of the NOVA micro-hypervisor for cross-layer U-Cloud security; an Intel SGX technology-based execution environment for strong isolation and trusted execution of computations; and the highly promising Cloud FPGA framework, to accelerate compute workloads and security management.
Self-management is based on orchestration of security services, including: OrBAC policy-based SLA management and authorization, from policy modelling to enforcement, notably security SLA negotiation, e.g. for firewall policies or availability in SDN networks; cross-layer and cross-provider autonomic security monitoring; geolocation-aware replication; and software trust management, extending to customers the experiences of providers during inter-provider trust assessment.
The data management framework provides dependability enablers leveraging state machine replication around the Hyperledger Fabric blockchain open sourced by IBM. Novel distributed protocols such as the XFT resilience model or the WHEAT Byzantine Fault-Tolerant (BFT) consensus protocol have been elaborated ─ the first BFT ordering service based on WHEAT was integrated into Hyperledger Fabric. Resilient distributed storage enablers are organized around the Janus multi-cloud storage service, with features covering disaster recovery, resilient key-value stores, or secure deduplication. Security and privacy enablers include cryptographic tools such as the SUPERCLOUD-ABE library for attribute-based encryption, or solutions for verifiable multi-party computation or data anonymization.
The network virtualization infrastructure, based on the SDN paradigm, gives tenants freedom to specify their own virtual network topologies, then deployed over multiple public clouds (e.g. Amazon EC2, Google Cloud) and private clouds (e.g. testbeds in Portugal and France). The Sirius network hypervisor and orchestrator interconnects multiple clouds while supporting the embedding of virtual networks specified by tenants ─ novel such embedding algorithms have been proposed to take into account user security and privacy requirements and scalability issues.
The infrastructure also includes components to support flexible chains of network security services or self-management of security incidents in the network. Regarding dependability, a logically-centralized SDN security architecture was proposed for secure and efficient control plane communications. Rama, a fault-tolerant SDN controller has also been developed to improve control plane resilience.
A number of SUPERCLOUD components were integrated into both demonstrators, and on the project testbed. Demonstrators were showcased successfully multiple times. Evaluation results show that the SUPERCLOUD infrastructure reaches fair maturity and performance levels, including cutting-edge user-centric elements comparing favourably or starting to be adopted in leading market solutions.
The computing framework includes components for virtualization and security self-management. For instance: the MANTUS/ORBITS virtualization framework to instantiate distributed U-Clouds specified in TOSCA, security services being "weaved" under user control; an extension of the NOVA micro-hypervisor for cross-layer U-Cloud security; an Intel SGX technology-based execution environment for strong isolation and trusted execution of computations; and the highly promising Cloud FPGA framework, to accelerate compute workloads and security management.
Self-management is based on orchestration of security services, including: OrBAC policy-based SLA management and authorization, from policy modelling to enforcement, notably security SLA negotiation, e.g. for firewall policies or availability in SDN networks; cross-layer and cross-provider autonomic security monitoring; geolocation-aware replication; and software trust management, extending to customers the experiences of providers during inter-provider trust assessment.
The data management framework provides dependability enablers leveraging state machine replication around the Hyperledger Fabric blockchain open sourced by IBM. Novel distributed protocols such as the XFT resilience model or the WHEAT Byzantine Fault-Tolerant (BFT) consensus protocol have been elaborated ─ the first BFT ordering service based on WHEAT was integrated into Hyperledger Fabric. Resilient distributed storage enablers are organized around the Janus multi-cloud storage service, with features covering disaster recovery, resilient key-value stores, or secure deduplication. Security and privacy enablers include cryptographic tools such as the SUPERCLOUD-ABE library for attribute-based encryption, or solutions for verifiable multi-party computation or data anonymization.
The network virtualization infrastructure, based on the SDN paradigm, gives tenants freedom to specify their own virtual network topologies, then deployed over multiple public clouds (e.g. Amazon EC2, Google Cloud) and private clouds (e.g. testbeds in Portugal and France). The Sirius network hypervisor and orchestrator interconnects multiple clouds while supporting the embedding of virtual networks specified by tenants ─ novel such embedding algorithms have been proposed to take into account user security and privacy requirements and scalability issues.
The infrastructure also includes components to support flexible chains of network security services or self-management of security incidents in the network. Regarding dependability, a logically-centralized SDN security architecture was proposed for secure and efficient control plane communications. Rama, a fault-tolerant SDN controller has also been developed to improve control plane resilience.
A number of SUPERCLOUD components were integrated into both demonstrators, and on the project testbed. Demonstrators were showcased successfully multiple times. Evaluation results show that the SUPERCLOUD infrastructure reaches fair maturity and performance levels, including cutting-edge user-centric elements comparing favourably or starting to be adopted in leading market solutions.
SUPERCLOUD virtualization allows flexible but efficient user-centric multi-cloud trade-offs between interoperability and security, with automation of protection. SUPERCLOUD also enables user-centric data encryption, with dependability guarantees across multiple administrative domains. Finally, SUPERCLOUD enables tenants to create virtual networks with customized topologies and addressing schemes spanning datacenters of multiple providers with the required isolation level.
SUPERCLOUD technology opens multiple paths for exploitation, focusing on 3 key markets: cloud brokerage markets to offer on-demand networks with fully à la carte security in the context of uberized virtualized network infrastructures; blockchain markets to guarantee strong and flexible data protection in multiple administrative domains, and target of strategic investments of key SUPERCLOUD partners across many business cases supported by a major open source project (e.g. Hyperledger); and cloud healthcare markets to meet security, data availability, and interoperability requirements, and to improve existing services and launch new products. More vertical markets may be supported as SUPERCLOUD provides an open framework.
For commercial exploitation, partners leveraged their business units to incorporate SUPERCLOUD results into existing portfolios and innovation management processes, to develop new products or services, or through startup creation, e.g. around the Janus storage technology.
To promote research excellence of SUPERCLOUD, partners contributed a great number of publications in top-class conferences such as OSDI, NDSI, CCS, EuroSys, DSN, CLOUD, CCGRID, or DBSec, or prestigious journals such as IEEE TPDS, DTSC, Cloud Computing, or Security & Privacy. Orange and IBM were interviewed by the IEEE to share the SUPERCLOUD perspective, resulting in a widely disseminated paper increasing project visibility. At EuroSys’17, partners organized the very successful ACM XDOM0 workshop on Security and Availability of Multi-Domain Infrastructures. Orange co-organized two editions of the SEC2 workshop on Cloud Security with SUPERCLOUD keynotes and papers. The SUPERCLOUD collaborative information platform was regularly updated, with publication of a newsletter. Training activities led by TUDA also fostered wider acceptance of SUPERCLOUD results, e.g. Summer School on Secure & Trustworthy Computing.
Many SUPERCLOUD components were open sourced, notably the Hyperledger Fabric, and the security monitoring framework. Standardization was also very active, e.g. on NFV at ETSI, or on encrypted objects and access control for healthcare at SNIA CDMI. Standardization efforts of SUPERCLOUD and H2020 SSICLOPS were synchronized to push recommendations on Trusted Inter-Cloud Computing at ITU-T.
Thus, SUPERCLOUD will enable user-centric and secure multi-cloud execution environments, enhancing cloud computing with innovative technologies and services, creating new business opportunities along several verticals of the multi-cloud ecosystem, uplifting Europe´s innovation capacity, and improving its competitiveness.
SUPERCLOUD technology opens multiple paths for exploitation, focusing on 3 key markets: cloud brokerage markets to offer on-demand networks with fully à la carte security in the context of uberized virtualized network infrastructures; blockchain markets to guarantee strong and flexible data protection in multiple administrative domains, and target of strategic investments of key SUPERCLOUD partners across many business cases supported by a major open source project (e.g. Hyperledger); and cloud healthcare markets to meet security, data availability, and interoperability requirements, and to improve existing services and launch new products. More vertical markets may be supported as SUPERCLOUD provides an open framework.
For commercial exploitation, partners leveraged their business units to incorporate SUPERCLOUD results into existing portfolios and innovation management processes, to develop new products or services, or through startup creation, e.g. around the Janus storage technology.
To promote research excellence of SUPERCLOUD, partners contributed a great number of publications in top-class conferences such as OSDI, NDSI, CCS, EuroSys, DSN, CLOUD, CCGRID, or DBSec, or prestigious journals such as IEEE TPDS, DTSC, Cloud Computing, or Security & Privacy. Orange and IBM were interviewed by the IEEE to share the SUPERCLOUD perspective, resulting in a widely disseminated paper increasing project visibility. At EuroSys’17, partners organized the very successful ACM XDOM0 workshop on Security and Availability of Multi-Domain Infrastructures. Orange co-organized two editions of the SEC2 workshop on Cloud Security with SUPERCLOUD keynotes and papers. The SUPERCLOUD collaborative information platform was regularly updated, with publication of a newsletter. Training activities led by TUDA also fostered wider acceptance of SUPERCLOUD results, e.g. Summer School on Secure & Trustworthy Computing.
Many SUPERCLOUD components were open sourced, notably the Hyperledger Fabric, and the security monitoring framework. Standardization was also very active, e.g. on NFV at ETSI, or on encrypted objects and access control for healthcare at SNIA CDMI. Standardization efforts of SUPERCLOUD and H2020 SSICLOPS were synchronized to push recommendations on Trusted Inter-Cloud Computing at ITU-T.
Thus, SUPERCLOUD will enable user-centric and secure multi-cloud execution environments, enhancing cloud computing with innovative technologies and services, creating new business opportunities along several verticals of the multi-cloud ecosystem, uplifting Europe´s innovation capacity, and improving its competitiveness.