5GinFIRE Open Call FAQs
5GinFIRE Open Call FAQs
- Is there any Available documentation?
- Is there any Testbed/Tools documentation?
Please have a look at the following links:
- Experimenter Tools (https://5ginfire.eu/facilities-2/)
- The testbeds (http://wiki.5ginfire.eu/5GinFIREtestbeds)
- Is the core network (i.e. Unified GW) provided as a functionality of the system and deployable through OSM?
Yes the UGW is the core LTE network.
So far the VNF UGW is not deployable via OSM MANO and B-COM is still working in the descriptors.
The PNF UGW is already deployed at Bristol site.
- Is it a hard limitation "Before awarding any grants to a third party, it will be checked whether the third party is a legal entity with a history of at least three years of commercial operations, and has not been declared insolvent." ?
- Which OSM version should be considered, TWO or THREE or FOUR?
Current infrastructure is based on OSM Release TWO, and the team is working on migrating to OSM Release FOUR. We foresee two possible scenarios:
- OSM Release TWO as production orchestrator and OSM Release FOUR as an experimental one.
- Only OSM Release FOUR as production and experimental orchestrator.
In general, we think it is advisable to assume OSM Release TWO/FOUR for experiments under category (1) in the open call, as they would be migrated to OSM Release FOUR as required; and OSM Release FOUR for categories (2-a) and (2-b) in the open call, as new features will be incorporated for further 5GINFIRE releases.
- How is the experiment lifecycle working? Will the experimenter be allowed to access directly the OSM launchpad/APIs, or the only interaction allowed is via the 5GinFIRE portal?
Please check the wiki http://wiki.5ginfire.eu/#the-5-gin-fire-architecture-technologies-services-and-testbeds. Normally the access will be only through the portal.
- Considering that we would like to instantiate multiple slices, are we allowed to execute them in parallel?
In 5GinFIRE,experiment is the abstraction that is available to experimenters for deployment. The 5GinFIRE MANO platform supports the deployment of simultaneous experiments, being their isolation guaranteed through the Virtualized Infrastructure Managers (VIMs).
- Is there a multi-site federated infrastructure, or the different testbeds are seen as individual sites? In the last case, where should be specified the site for deployment?
The technical solution adopted by 5GinFIRE for the MANO stack considers the utilization of a single orchestration domain, where an NFV orchestrator, implemented with Open Source MANO (OSM), manages and coordinates the creation of network services, being a network service generally defined as a composition of Virtualized Network Functions (VNFs) and/or Virtualized Vertical Functions (VVFs), generally referred to as VxFs. Each of these VxFs may be in turn deployed at any of the NFV infrastructures (NFVI) made available by 5GinFIRE partners. At the time of writing, three of these infrastructures have been enabled for experimentation activities through the 5GinFIRE MANO platform (an additional experimental infrastructure is under development at UFU):
- An NFV infrastructure at the global 5G Telefonica Open Innovation Laboratory (5TONIC), made available by TID (founding member of 5TONIC) and UC3M (member of 5TONIC);
- An NFV infrastructure located at ITAv;
- And an NFV infrastructure made available through a collaborative agreement by UNIVBRIS and BIO.
Each site running an experimental infrastructure deploys a Virtualized Infrastructure Manager (VIM), compliant with the OSM software stack. The different sites providing the NFV infrastructures are interconnected, and the experimenter can request, when defining an experiment through the 5GinFIRE portal, which experimental infrastructure should support the execution of each VxF.
Is inter-VNF communication based on classical virtual networking like in OpenStack?
At this moment, all the sites that provide an experimental infrastructure are using a Virtualized Infrastructure Manager (VIM) based on OpenStack. Virtual networking for VNF communications at each site is enabled by its corresponding OpenStack VIM.
Is OpenStack the VIM/NFVI used in the sites? Which version specifically?
Yes. The current version is Openstack Ocata.
- Are we allowed to upload specific images on the VIM/NFVI while implementing VNFs?
Uploading VM images on the VIMs is part of the development/validation process of VNFs, so this will be allowed upon request.
- Will the experimenter get access (i.e. SSH) to the deployed instances? If yes, how? Via VPN or public IPs?
The 5GinFIRE platform support the controlled access of experimenters to their reources under experimentation (i.e., VNFs). This is done through a VPN access to the platform, and via ssh to the VNFs.
- In case of VPN, are there some firewall rules?
Yes, firewall rules are configured to control the access to infrastucture resources and deployed VNFs.
- Will you accept proposals with potential modifications to the OSM tool? If yes, which version should be considered, TWO or THREE?
YES this falls under category 2a. for version see answer above
- Will you accept proposals with potential modifications/extensions to the 5GinFIRE portal?
YES this falls under category 2a
- How are those extensions supposed to be integrated? Should we consider potential modifications to the components, or should we only use APIs (hopefully with SDKs you may provide)?
They can be either services or APIs usage, APIs extensions or modifications. Proposers are free to define
In case of extensions based on APIs, are you providing an API description which we could consider for our initial solution design? Please let us know also about any restrictions due to any RBAC mechanism you utilise.
- Regarding the Portal there is an API. Please check the http://wiki.5ginfire.eu/portalarchitecturedesign/portal-api section
- Regarding OSM there are the ETSI NFV APIs specified and the correspponding OSM version to support
Currently we have only noticed that the experiment lifecycle mainly deals with category 1 types of experiment, therefore it is not clear how those potential extensions could be realised and be deployed on demand on top of your platform.
The extensions can be deployed as services or they can reuse existing APIs (Portal, OSM). Proposers, though, are free to define the deployment they need
- Will some support be provided about the description files definition?
Please check this Wiki. The descriptors are standardized in format and packaging by ETSI specifications
- Is a successful applicant expected to attend 5GINFIRE review meetings and / or onsite visits (e.g. at the Testbed sites, for integration)?
Beside review meeting, we would like to invite all successful proposers to a kick-off meeting in May 2018 as well as tutorial session at EUCNC conference (to be confirmed) if appropriate. Reserving some additional resources for additional travels, e.g. to Testbed sites seems reasonable.
- The budget for the open-call allows the new partners to buy equipment?
In principle we do not specify a limit for equipment purchases. This means that this could be done within reasonable limits. The spirit of the call is to put the money into work and not into infrastructure or other equipment.
The main point is whether the purchase is duly justified and that the rest of the budget request is leaving sufficient room for executing the work or operation (in case of 2b).