The FELIX project Sustainability Bartosz Belter, PSNC FIRE Board Meeting, Brussels.BE October 2, 2015
What is FELIX A federation is a union comprising a number of partially self-governing regions united by a central ("federal") government under a common set of objectives * * Thierry Parmentelat, INRIA ( OneLab2 & Federation ) FELIX works on the interconnection of EU and JP SDN testbeds To increase mutual benefits of European and Japanese researchers by creating more complex environments for specialized research and experiments To create new opportunities for experiments due to geographical dispersion of testbeds Resources of the global EU-JP Federation becomes available for users of individual testbeds
FELIX results FELIX Federated Framework General architecture for federation of different Future Internet facilities FELIX Management Stack (FMS) Open Source implementation of the FELIX Federated Framework FELIX testbed Distributed experimentation testbed in which FMS is managing infrastructure slices over IT servers, switches and data links interconnecting testbed islands FELIX experiments Realization of defined federation use-cases which validates both FELIX Federation Framework and FMS software
FELIX Federated Framework General architecture Defines functional blocks Flexible and scalable Replaceable Resource Managers for Compute nodes (CRM), Network nodes (SDNRM), inter-domain connectivity (TNRM and SERM) as well as monitoring collector (MS) Hierarchical/recursive structure of resource orchestration (RO) Inter-domain network aware Uses NSI network services Allocates static links for experiments
Project phases Specification Use Cases User Requirements System Requirements FELIX Architecture Implementation Testbed implementation Software development and early testing Validation and Experiments Implementation of use cases in the testbed Experiments Validation of the FELIX Architecture Dissemination & Standardization (WP5) Architecture (WP2 / WP3) Implementation (WP3 / WP4) Experiments (WP4) Y1 Y2 Y3
Our thoughts on sustainability Different types of costs exist OPEX: Infrastructures require continuous funding to be available over long terms CAPEX: Infrastructures (hardware) require continuous update over time How to cover them? Own investments Public funding (e.g. H2020, National/Regional programmes, etc.) Incomes from experimenters, i.e. invoice for the usage of infrastructures Reality FIRE users might not be keen to pay for access to infrastructure Law restrictions in some countries may limit the commercialization of FIRE (e.g. infrastructures built from public funds cannot be offered as paid service on the market) Usually, partnering to form new alliances and submit for new co-funded projects to evolve and maintain infrastructures and FIRE services is the most viable path A good example: transition of some OFELIA infrastructures/islands in FELIX FELIX is not an OFELIA follow-up project, but some partners from original OFELIA involved Open Access from external users works fine, but only for a limited amount of time Supporting users requests and solve access problems requires extra resources (sort of help-desk) Lessons learned from previous projects are good to avoid errors and optimize the support process, but they are not always sufficient How to cover this extra cost? Still, funding is the most viable solution
Sustainability from other perspectives On sustainability of technologies Complying to standards or de facto standards increases chances to maintain infrastructure & management software over long terms at reasonable costs FELIX implements the so called common API, which is the agreed standard between GENI (US) and FIRE (EU) With FELIX this is now adopted in Japan We closely work with Fed4FIRE and are involved in GENI Takeout from our experience More efforts (and support) should be put in concertation and liaisons of FIREs to maintain the technological results of previous project Collaboration across projects and programs stimulates support for new technologies, as well as consolidation of APIs, architectures, shared tools
Sustainability from other perspectives (2) Sustainability vs. availability of inter-domain links In FELIX-kind-of-FIRE sustainability is also impacted by something beyond infrastructure owner premises, i.e. inter-domain/continental links NSI is now for free since PSNC, SURFNET, AIST and KDDI are involved in NSI development via GLIF and GÈANT What could happen in a different scenario, e.g. for a new infra owner willing to join FELIX or if NSI services would be on a pay-per-use mode? i.e. how much could it cost to experimenters to access a FIRE infrastructure across continents? Open proposal FELIX-kind-of-FIRE could be incorporated within GÈANT / GLIF mainstream service activities on SDN and experimental testbeds offer FIRE as part of a larger product portfolio for researchers (and possibly maximize CAPEX /OPEX savings for infra maintenance and evolution) Another opportunity continue supporting this R&D within FIRE+ and EU-JP Joint Calls
PARTNERS Poznan Supercomputing and Networking Center Poland National Institute of Advanced Industrial Science and Technology Japan Nextworks Italy Fundacio Privada i2cat, Internet I Innovacio Digital A Catalunya Spain SURFnet bv Netherlands European Center for Information and Communication Technologies Gmbh Germany iminds VZW Belgium KDDI Japan