APIS operates on well-supported operating systems and other core infrastructural software and is using hardware and software technologies appropriate to the services it provides to its Designated Community.
The technical infrastructure of APIS is in two different locations:
a) The archive backoffice, including its website, is allocated in ICS-UL infrastructure, whose the main characteristics are as follows:
A VMware vSphere solution was implemented with a cluster composed of two ESXI 6.7 hosts (managed by vCenter Server Appliance 6.7) and an HPE MSA2050 shared storage that supports all infrastructure and production virtual machines.
The network for the entire infrastructure has its own VLAN, for management and for the production part.
A backup solution was set up with a Storeonce3100 unit, working with Veeam Backup, and reinforced with a NAS QNAP unit in production at other ULisboa facilities, which works as storage for remote file backups.
In addition, an IRS server (Insight Remote Support) was installed, which aims to collect the alarms of the HPE equipment and, in an automatic way, inform the manufacturer about the incidents that occurred. With this application it is possible to check the type of contract and its validity for each serial number.
The entire infrastructure is directly supported by two APC Smart UPS X3000 permanent power units, configured with the “PowerChute” management application.
We can simplify it as follows:
|(2) Servers HP DL360Gen10 (a)||Physical hosts|
|(1) Unit HPE MSA2050 (b)||Virtual infrastructure storage|
|(1) Unit StoreOnce 3100 (c)||Backups Storage|
|(1) Unit NAS QNAP (d)||Backups File Storage|
|(2) UPS APC (e)||Power supply units|
(a) Two servers with Intel Xeon Silver 4114 CPU (2 Processors), 128GB of memory and ESXI 6.7 operating system.
(b) Storage with SAS connection, two disk controllers and eight 2TB SAS 10K disks.
(c) 8TB capacity repository.
(d) One NAS QNAP unit with four 6T SATA3 disks each and two more hot-spare disks.
(e) Two UPS units with an output capacity of 3000VA
b) RCAAP is based on OpenSource DSpace which is compliant with the Dublin Core metadata scheme and the OAI-PMH protocol; the system has OpenSource operating system Linux, CentOS distribution and redundant infrastructure with high-availability which is provided by FCT/FCCN.
In page 6 of this document we explain the relation between partners (also in shared technology), including the obligations and roles that are in the contract (2012).
• Information about the implemented standards (DC, MODS, METS):
https://wiki.duraspace.org/display/DSPACE/Home and http://www.dspace.org/why-use
All data is securely handled by the archive in APIS google drive, where data folders are organized according to OAIS model. We may also refer to the ISO 16363 compliance of RCAAP, being the main conclusion: “The internal audit revealed a high level of maturity regarding the fulfilment of the normative requirements (…) and being well prepared in terms of infrastructure, reliability and security in the preservation of digital objects.”
• The repository performs annual assessments on the capacity and forecasting (6-month weighted). The assessment concerns disk space, CPU usage, memory usage and bandwidths consume.
• The information about the installed software is kept in a private wiki connected to the project. All software documentation used can be found here:
• The software is based on the DSpace community.
More information here: http://www.dspace.org/why-use.
• Information about the installed hardware:
The virtualization system that supports RCAAP virtual machines is based on the CITRIX XenServer solution, which implements a virtualization solution of enterprise-class. The solution offers all critical resources needed for any server implementation and virtualization.
The FCT|FCCN virtualization service is focused on business continuity and it implements a geographically redundant architecture (technical rooms; SE03 and GRID) composed by two Clusters of autonomous virtualization, without a single point of inter-cluster failure. In case of unavailability of one of the data centres, the interoperability between the two clusters establishes a Disaster Recovery Plan (DRP), which guarantees the continuity of service. The Recovery Time Objective (RTO) maximum level is 6 hours and the Recovery Point Objective is 24 hours.
The technical infrastructure of the repository provides for protection of the facility and its data, products, services, and users.
In 2015, the repository (RCAAP) conducted a compliance assessment with ISO 16363. It was under a large-scale audit initiative audit of the Institutional Scientific Repositories of Portugal that covered around 28 repositories which revealed reliability in the management and preservation of digital objects.
List of installed processes and critical processes:
In the framework of RCAAP, the critical processes identified were those with direct impact in the:
• Compliance with the obligations established with the adherent entities;
• Legal compliance.
Impact analysis – relation between the time spent and the impact caused by the inexistence of the process.
(Note that this risk analysis has been developed by RCAAP for all the repositories and not specifically for APIS. Additionally, APIS does not store sensitive data up to now.)
The risk analysis is the result of a partnership with the TIMBUS project (Timeless Business Processes and Services – http://timbusproject.net/).
The context model shows the dependencies of the system – both in terms of infrastructure, software, operating system, data model and obligations – and then, after defining the requirements, a risk model was made. This was intended to know the risks that may affect the system in order to promote the preservation of RCAAP.
The methodology entails the following phases:
• Risk identification
• Risk analysis
• Risk assessment
Based on the work carried out in partnership with the TIMBUS Project, the model defines:
• The elements used to assess the risk
• The metadata used to describe the elements of the model
The figure shows the risk matrix after the implementation of the mitigation strategies pointed out in the TIMBUS Project (controls).
The risk identification can be found in ANNEX 1.
The Contingency Plan intends to anticipate and manage the impact of an eventual disaster with RCAAP services. According to ISO 16363 – 5.2.4, the repository must have a contingency plan that includes at least one copy of all digital information placed in a geographically distant location (including a copy of the contingency plan).
In this sense, recovery plans were made for 4 scenarios:
• C1 – Unavailability of the HOST
• C2 – Permanent loss of the HOST
• C3 – Unavailability of the Backup
• C4 – Violation of the integrity of an Archive Informational Package (AIP)
For each scenario, there are specific instructions and plans of action to replace the services.
There are also a set of procedures related with the following typologies:
• Service maintenance and content preservation
• Content deposit
• Support to the community
• Risk management
ANNEX 1 – RISK IDENTIFICATION
We should still highlight the following references concerning APIS outsource partners’ security certifications:
Trust certification of FCCN: https://docplayer.net/18228041-Trusted-certificate-service-tcs.html
RCTS Certificates is the service that ensures the supply of several types of digital certificates to the entities connected to RCTS. This service promotes the security, authenticity, confidentiality and integrity of information sent and received via the Internet. FCCN guarantees this service through the TCS (Trusted Certificate Service), provided by GÉANT Association (involved in EOSC and FAIRsFAIR project): https://www.fccn.pt/en/security/rcts-certificates/.