Questions from the Request for Tender process

To stimulate an open dialogue with companies interested in the ARCHIVER project, all information given in answers to questions raised by potential suppliers during the Request for Tender period will be documented and published in this section.

This page will be regularly updated as new questions are answered.

Clarifications

C1.Link to OMC Results

The DOI of the “State of the Art, Community Requirements and OMC Results Report” provided in the Request for Tenders (section 2.4) and the Functional Specifications (section 3) is incorrect. On both documents, the DOI goes to: https://doi.org/10.5281/zenodo.3613577. The correct DOI is: https://doi.org/10.5281/zenodo.3618215.

C2. Impact of Brexit on the Tender evaluation process

The European Commission has confirmed that for the purposes of evaluating responses to the ARCHIVER Request for Tenders, the United Kingdom shall be considered as an EU Member State. This is relevant in particular to Compliance Criteria CC.2 (Annex 3: Compliance Questionnaire, section 2.2) which states that “The majority of the R&D services (as defined in section 3.1 of the Request for Tenders) over all three Phases, including in particular the principal researchers working for the PCP Contracts, shall be located in EU Member States or Horizon 2020 Associated Countries.

C3.      Revised deadline and project timeline

In light of the ongoing Covid-19 outbreak, the ARCHIVER Request for Tenders deadline is extended to 14 April 2020 – 4:00 pm (expressed in Europe/Zurich time zone).

The timeline for the PCP described in section 3.3 of the Request for Tenders is updated below. An updated version of the Deliverables and Milestones document is annexed to this clarification (made directly available on the Lead Procurer’s e-tendering platform, and on the RfT document available by download from the project’s website).

C4.      New revised deadline and project timeline

In light of the ongoing Covid-19 outbreak, the ARCHIVER Request for Tenders deadline is extended to 28 April 2020 – 4:00 pm (expressed in Europe/Zurich time zone).

The timeline for the PCP described in section 3.3 of the Request for Tenders is updated below. An updated version of the Deliverables and Milestones document is annexed to this clarification (made directly available on the Lead Procurer’s e-tendering platform, and on the RfT document available by download from the project’s website).

The Buyers Group continues to assess the impact of the Covid-19 outbreak on the ARCHIVER project and reserves the right to amend the project schedule to take this into account. Any amendments to the project schedule will be communicated during Call-offs.

Questions & Answers

Q1.You allocate points in the technical evaluation of bids for Open Source solutions but I understood that IP can remain with the selected vendor. Can you please clarify?

A1.      The two are not mutually exclusive. The Results generated as a result of the PCP Contracts will belong to the Contractor. ARCHIVER favours Tenders that commit to making these Results available under Open Source licensing conditions.

Q2.Would there be only one contractor selected for the Pilot Phase?

A2.      The aim is to have at least two contractors during the pilot phase. Please refer to section 3.4 of the Request for Tenders for further information.

Q3. Is it possible to update our Tender in CDS before the response deadline?

A3.      Yes, this is technically possible.

Q4. What if none of the Tenders meet the stated minimum requirements?

A4.      The requirements will not be modified. A Tender that does not meet the minimum stated requirements will be discarded and cannot result in a PCP Contract.

Q5.      Is the project scope limited to the domains described?

A5.      The Results developed during the project are intended to respond to the long-term data preservation and archiving needs of public sector scientific research organisations generally, not just in the four scientific domains of the four procurers (for which deployment scenarios are provided).

Q6.      With reference to the Deliverables and Milestones document, point M1.2 requests a basic capabilities platform for archiving and preservation. Which are these basic capabilities?

A6.      In this context, M1.2 is requiring access to a "demo version" of your current solutions, in terms of current functionality. This demo access shall reflect the baseline state of the art, with up to 100 TB of capacity.

Q7.      With reference to the Deliverables and Milestones document, point M1.2 mentions that the prototype platform need to have up to 100 TB capacity. Is this the total capacity for all datasets or is it an individual dataset maximum size?

A7.      The 100 TB access are required by the Buyers Group for assessment of the current solutions. At this stage, they shall reflect the total capacity for all datasets and assessments.

Q8.      With reference to the Functional Specifications section 3.2, is there any technical specification of the client side to deploy applications on it if needed?

A8.      All the requirements are functional and do not necessitate the use of any specific technology.

Q9.      With reference to the Functional Specifications sections 2.3, 3.3 and 4.3, compliance with CoreTrustSeal or ISO 19363 is specifically requested by means of a self-assessment to be delivered (these deliverables seem to be indicated in the Deliverables and Milestones document). In terms of the other regulations/standards (ISO/IEC 27040 for storage security, ISO/IEC 19086 for cloud SLAs, FAIR principles, GDPR, etc.) against which tenderers are requested to demonstrate compliance according to the Functional Specification document, no specific deliverable has been stated within the Deliverables and Milestones document. When and by what means are the tenderers expected to demonstrate compliance against those? More generally, for all relevant regulations/standards, will it be enough to demonstrate compliance by means of self-assessments by each tenderer?

A9.      The correct standard mentioned in the received question shall read ISO 16363 and not 19363. D1.6 is not asking for conformance but instead for a self-assessment of the tenderer. In other words, where is the tenderer situated in the process that leads to compliance either in the CoreTrustSeal or ISO 16363 frameworks? Concerning the additional aspects referred in the question, the tenderer shall demonstrate the technical and organisational measures in place in order to comply with the process foreseen in 19086 for SLAs, how the FAIR principles are being followed (there is no standard for FAIR yet, but your attention is drawn to the outputs of the EOSC FAIR WG (https://www.eoscsecretariat.eu/working-groups/fair-working-group, which may be used as a reference) and how is GDPR being respected "by design" respectively for the Architecture Design Document (D1.4), Prototype in Phase 2 (D2.4) and Pilot in phase 3 (D3.4).

Q10.    How virtual price will be considered to select consortium for PCP?

A10.    The virtual price is the price that the Buyers Group would pay for the services in a scenario of exclusive development (please refer to section 8.1 and 8.2 of the Request for Tenders). The Buyers Group will verify that the actual price is lower than the virtual price, reflecting investment made by the Tenderer in the project on account of the risk- and benefit-sharing approach of the PCP Contract. This precludes the possibility of ‘state aid’. Other than this verification, the virtual price is not taken into account for the evaluation of the Tenders.

It should be noted, however, that the quantities/number of hours and resources indicated by the Tenderer in the breakdown of the virtual price for Phase 1 in Annex 5 represent a commitment on the part of the Tenderer.

Q11.    With reference to section 4.1.2 of the Request for Tenders, does ‘data preservation and archiving services’ include the cloud provider?

A11.    ‘Data preservation and archiving services’ does not include the simple provision of infrastructure for storage in the cloud. Infrastructure provision in the Petabyte range will address Layer 1 (please refer to section 2.1 of the Functional Specification). Data Preservation and Archiving software services will be for Layer 2-4. The degree of integration across infrastructure providers and the archiving and preservation software layers will define the real innovation that can be achieved.

Q12.    With reference to Annex 5, where should prices for subcontractors be indicated?

A12.    The tables in the ‘summary’ tab and the ‘price breakdown’ tab pertaining to the ‘other Consortium member’ shall be duplicated as needed, and renamed as ‘subcontractor’ or similar in the case of a subcontractor.

Q13.    With reference to Annex 4, section 2.2, could you please clarify if the tenderer can include third-party services that run on top of my infrastructure?

A13.    Your attention is drawn to the definition of the term “Solution”, which means a service or set of services for archiving and data preservation following the OAIS reference model, deployed in the cloud or in a hybrid model, able to handle petabytes of data and sustained high ingestion rates (approximately 1-10 GB/s). Third-party services can be included in the resulting solution if they comply with the requirements defined in the PCP Contracts. Specifically, the need to be available through a central EOSC service catalogue access point so that a researcher has access to the full set of ARCHIVER services, being able to trial them, evaluate their functionality for her/his specific research field and, in the future, purchase them with a clear costing model taking into account all the associated cost influencing factors and, if needed, be able to repatriate and recover research data seamlessly to another location by the end of the contract and usage period.

Q14.    With reference to Annex 4, section 2.3, could you please clarify what is meant by ‘plausible arguments’ in the phrase “The Tenderer shall describe the solution that will be designed making plausible arguments about the demanded volumes at scale”?

A14.    In other words, it is expected that the tenderer will describe the technical capabilities and scalability of the resulting solution in terms of the defined criteria.

Q15.    Could you please indicatively specify how many members the Buyers Group has and how many people will use the helpdesk ticketing system?

A15.    The Buyers Group has four members (CERN, DESY, EMBL and IFAE-PIC). Please refer to section 2.5.2 of the Request for Tenders. The expectation is to have an approximate number of two technical members of each of the Buyers accessing the ticketing system.

Q16.    With reference to Annex 4, section 3.3, could you please clarify what is meant by “The merit of the commercialization plan and the impact of the proposed solution”?

A16.    Your attention is drawn to section 8.1 of the Request for Tenders. The contractor is expected to commercially exploit the Results developed during the PCP. This Award Criteria will be used to evaluate the tenderers ability and readiness to commercialise the Results. The tenderer shall therefore describe how it will commercially exploit the Results and bring a viable product or service to market, including proposed licensing and software arrangements. The Tenderer will explain why it has chosen this approach to commercialisation. The Tenderer will indicate its target market and the reasons why it is targeting this market. The Tenderer shall estimate the time to market.

Q17.    Regarding storage capacity, for phase 1 the requirement is 100 TB. What is requirement for phase 2? What is requirement for phase 3? It is mentioned key functional requirement for archiver R&D services is tens of PBs with linear growth over the years: is this a requirement once the services are in “production” mode, i.e. after phase 3 completed and in the scope of a future procurement contract?

A17.    In general, there are no strict requirements in terms of volumes through the phases 2 and 3 but we do expect agile feedback cycles to test deployments. In the prototype Phase, Functional testing will start, and the capacity availability shall be staged and shall not be over 1 PB. In the Pilot phase, performance and scalability will be among the main aspects to test, storage capacity shall not be over 5 PB. During this phase, particular focus will be given to the viability of the implementation of exit strategies.

Q18.    With reference to section 2.3.1 of the Functional Specifications relating to connectivity between datacenter and research sites, are you requesting dedicated leased line between datacenter and each research sites (CERN, DESY, EMBL and PIC)? If yes, what is the speed for each leased line? Could you provide more information about connectivity of datacenter via GEANT network? Are you considering connectivity between datacenter and research institutes via internet? What would corresponding speed?

A18.    Contractors are NOT required peer to the Buyers Group Data Centers directly.  In order to ensure large end to end network capacity (in multiples of 10 GB/s) between the Contractor’s facilities and the Buyers Group data centres. Contractors are required to peer to the GÉANT network. Please find more on this topic in the Functional Specification document which reads Peering to GÉANT is essentially supported in four modes, depending on the use case:

• Connection to GÉANT via an NREN, to an NREN PoP location
• Direct connection to GÉANT at one or more of its PoP locations
• Connection at a GÉANT Open Exchange (London, Paris)
• Direct connection at one of the following Internet Exchanges (Amsterdam, London, Vienna, Frankfurt, Milan) - N.B.: the interconnection between the Tenderer and the GEANT network will be a dedicated, private interconnection on the IX premises, and will not use the existing shared connectivity between GEANT and the IX.

The Buyers Group will consider equivalent alternatives to peering to GÉANT (in terms of bandwidth capacity, performance and egress/ingress conditions) presented by the Tenderer. In this case, the Tenderer must elaborate on details concerning the network connectivity options available offering equivalent service features. These details must include information about egress/ingress charges during the project and after the project in case of future purchase of the resulting services.

Q19.    Milestone M1.2 references a demo platform with basic capabilities. Independently from the basic capabilities of the demo platform, are the connection and authentications with IdP to test the Service Provider Auth process necessary for Phase 1?

A19.    One of the overall aims in Phase 1 is a written detailed solution design, a system architecture, and a technical design of all the components. This implies that it is necessary to describe how network connectivity will be established from the contractor’s services to the GÉANT network, but not to have the demo platform already connected. Concerning Authentication and Authorisation, the Buyers Group except Tenderers to have Authorisation a& Authentication schemes in place that can be assessed and tested in order to understand the effort to integrate the Tenderer SPs with the IdPs in eduGAIN or other research community AAI services that will be available in the EOSC (European Open Science Cloud) context.

Q20.    If the connection to Geant is necessary, will the buyers connect to the demo platform using their connection to Geant network? If the connection is not necessary, can an Internet connection be used?

A20.    The GÉANT connectivity requirement is explained in Q18. The requirement is necessary to be in place from the Prototype Phase. This said, the time needed to fulfil the requirement should not be underestimated. The Buyers Group expectation is that selected Contractors have their plans for connectivity in place and are communicated during the Design Phase.

Q21.    What is the limit or estimation of Storage TB (TeraBytes) for phases 2 and 3?

A21.    Please refer to Q17.

Q22.    What is the estimation of bandwidth growth for the connection for phase 2 and phase 3?

A22.    In general, the Buyers Group estimate a minimum bandwidth of 10 Gbps from the Tenderer data centre to the GÉANT network and a minimum of 40 Gbps may be required during the testing periods in the pilot phase. Test scheduling will be performed as required in agreement between selected contractors and the Buyers Group.

Q23.    Could you please specify the contractor data centre location? Is it Geneva?

A23.    The contractor data centre locations are of exclusive responsibility of the Tenderers. If the questions refers to the Buyers Group data center locations : they are in Geneva in the case of CERN, in Hinxton (UK) for EMBL-EBI, Hamburg (Desy) and Barcelona (IFAE/PIC).

Q24.    In phase I, it is required the ability to store files and datasets of up to 100 TB. Could you please specify the TB quantity for Phase II? Could you please specify the TB quantity for Phase III?

A24.    Please refer to Q17.

Q25.    With reference to the deployment scenarios PIC Large File Remote Storage, PIC mixed file remote storage, and the Babar experiment, do you have an estimate of how many instances would be needed?

A25.    The number of instances to be purchased will depend on the type of data to be archived, on strategies to prevent vendor lock-in, and on payment schedules. So, for example, an instance may be purchased each year to archive 300 TB of data acquired from an instrument over that year that needs to be preserved for 5 years, using a single payment mode and a suitable SLA. Or an instance may be "purchased" once to archive all high-level data products from a completed experiment which need to be preserved for decades, using a yearly payment model with an agreed price revision policy and a suitable SLA including an exit strategy.

Q26.    With reference to the deployment scenario PETRAIII_EuXFEL Data Archiving, it is mentioned that the total archiver for each individual photon scientist’s data archiving use is 5 TB. Do you have an estimate of the number of users?

A26.    DESY has currently a few thousands of individual users for all three accelerators in operation with an expected increasing number of users interacting with an archive system of 500-700 per year.

Q27.    With reference to the deployment scenario PETRAIII_EuXFEL Data Archiving, you reference the need for manual data archiving for beamline experiments (pages 3 and 4). Do you have an estimation of how many beamlines there will be?

A27.    The current 3 (facilities) photon science accelerators will represent ~50 beamlines. We expect beamline common archive policies (largely influenced by lab wide data policies and facility specific extensions on that) to be grouped into roughly 20-30 larger groups dealing with common archiving activities.

Any further question can be sent to procurement.service@cern.ch using the form contained in Appendix C of the Request for Tenders before 10 March 2020 at 16:00 (expressed in Europe/Zurich time zone).