Communications, com puting and data services

During the LHC operation era com puting for particle accelerators and the experi­

ments has evolved into a service for a world-wide user community. Adopting products and best practices that emerge from an ever-growing Information and Communica­

tion Technologies (IC T) industry has proven to be a cost and performance effective path to serve the community. Large-scale science projects used to be a driver o f IT infrastructure developments [220- 225]. A future particle collider research facility can again be a case and a driving force for advanced IC T developments in areas that go beyond the hardware and software domains, which dominated the particle accelerator projects o f the 1980’s and 90’s.

A set o f general services comprising wired and wireless networks, desktop, mobile and centralised com puting for all users, various storage system tiers, software and data provisioning, authentication and authorisation, assistance and consultancy, training and much more, form the backbone o f services for individual business units (see Fig. 5.14) . Depending on the activity type, different levels o f quality o f service apply. The particle collider experiments require an elastic data com munication and

the other hand, involves various different networks for data, voice and video com ­ munication as well as infrastructures comprising embedded and real-time computing facilities.

Service units o f the facility such as the fire brigade and medical services, secu­

rity and site protection, environmental monitoring, safety-related systems and indus­

trial installations require dedicated IT services. In particular, remote monitoring and intervention has become an important means to reduce service level agreement costs for industrial plants over the last ten years. Finally, administration units ranging from human resources and finance to different types o f workflow systems need to be appropriately served whilst taking into account the length o f the construction and operation phases which will last for decades. For all domains, resilience, data protec­

tion, cyber-security, technology evolution and migration and long-term data accessi­

bility are topics that call for a dedicated organisation to ensure proper coverage with an appropriate mix o f in-house personnel, external suppliers and industry/academ ia partnerships.

The geographical extent o f the services and long-term sustainability o f a new large-scale particle-collider research facility call for a shift o f activities, traditionally covered by individual detector and accelerator engineering groups, to a business ori­

ented scientific IT unit. This approach allows tenders, contracts and operation to be optimised at an organisation-wide scale as well as influencing the value o f member state contributions favourably for the member countries, the organisation and the world-wide community. In particular, this approach for the construction and opera­

tion o f data centres can be beneficial for experiment users, accelerator engineers and users with generic needs.

Computing and interconnect technologies are evolving rapidly and generation changes need to be expected. At the same time the optimum cost effectiveness is continuously swinging between buying and leasing. This ever changing IT environ­

ment can best be accom m odated by a continuous cost/benefit analysis considering all users in the organisation, carried out by a team which is working closely with industry on one side and with the users on the other.

Embedded and real-time com puting including programmable logic controllers are also a concern for a technology infrastructure that is characterised by its longevity and thus dominated by operating costs. Given the significant increase in the num­

ber o f devices for a future collider, further standardisation, coordinated testing and certification and procurement and m aintenance/repair services, available to all users o f the organisation, will help to improve the cost effectiveness. A large-scale parti­

cle accelerator building on decades o f engineers’ experience presents an ideal case for an openly available architecture and platform for supervisory control to inte­

grate the diverse subsystems. A system that can evolve with emerging “Internet o f Things” (IoT ) products and yet unanticipated device technologies can create impact far beyond the particle accelerator community.

Cyber-security plays an increasingly important role in IT systems and embed­

ded com puting is no exception to this. The use o f processors, operating systems and embedded Web servers in the m ajority o f programmable laboratory equip­

ment such as simple digital I /O devices, measurement instruments, oscilloscopes and autonomous robots already require a well organised infrastructure. This is supported by a process that leads to a secure environment on one side and which has the least

1268 The European Physical Journal Special Topics

possible impact on usability on the other side. For example, it should allow the pos­

sibility o f developing and deploying across network boundaries, create islands and sandboxes to limit potential harm, have transparent virus and malware checking and isolation, have system updates that have little or no impact on work efficiency, pro­

vide coordinated rollback and much more. This evolution is expected to continue and the IoT approach will also require an organisation wide vertical integration o f services across the horizontal user domains [177].

Cooperation on IT standards, technology developments and organisation with other research facilities which have similar requirements (e.g. DESY, ESRF, ESS, Fermilab) needs to be strengthened. Synergies with other scientific domains (e.g.

astronomy and radioastronomy facilities, material sciences with light sources and FELs, astrophysics installations such as neutrino and gravitational wave observato­

ries, particle accelerators for medical applications and nuclear fusion experiments) can be developed to lead to more effective operation o f world-wide IT services for research. Activities spawned by DE SY on front-end com puting hardware [226] and CERN 's openlab [227] are good examples for such initiatives.

Possible services include not only fibre optics and data centres that can be shared with external partners but can also include mobile com munication and the coopera­

tion with emergency and rescue services in a cross-border context. C E R N ’s activities concerning the establishment o f the T E T R A radio com munication system in the region are a first step in this direction [228]. C E R N ’s particular status as carrier­

neutral Internet eXchange Point (C IX P ) [229] will gain importance in a technology ecosystem that becomes ever more dominated by profit making organisations which are building and operating global com munication infrastructures. These collabora­

tions may help to ensure that non-profit making organisations continue to operate independently and becom e less dependent on infrastructure operator priorities, which may favour financially stronger commercial clients.

Beyond the regional scale, the LHC programme has shown the value o f a world­

wide com puting and communication infrastructure to make the research data avail­

able to scientists in all participating countries [230]. The success o f a future particle collider programme will rely even more on international participation. Many poten­

tial participating countries are not part o f C E R N ’s global vision today and some are still technologically underserved. The SESAME light-source under the auspices o f UNESCO is a great example o f such an initiatives [231]. The capacity to plan, develop and implement a world-wide inclusion policy goes beyond the scope o f the high energy physics community and is therefore also a good opportunity to establish a com m on strategy with other scientific disciplines. A timely development o f such an initiative will help raising the interest and acceptance level o f large-scale investments in an infrastructure focusing on fundamental scientific research.

Considering the fast evolution o f information technologies in all domains [232], the long term cost impact o f in-house technology developments and their potentially limited large scale impact, it is prudent to base architectures for the particle acceler­

ator and the experiment data processing environments on industrial hardware, soft­

ware and service infrastructures. The particular needs o f an FCC-scale facility may, however, also represent attractive test-beds for emerging technologies. Co-innovation projects with industrial partners during the early construction phase, permitting pre­

commercial procurement initiatives that can lead to high-performance infrastructure services at a com petitive cost are one way to optimise this situation [233].

A preliminary cost-benefit-analysis o f the LH C /H L-LH C programme [234]

revealed that more than the impact value generated by training coresponds to more than one third o f the infrastructure’s cost (sum o f captital and operation expen­

ditures). The IC T sector represents an ideal case for training at large with ever growing societal and industrial demands. Early stage researchers and engineers are

mon optimisation, development o f standards and best-practices and co-innovation with the research infrastructures as a demonstration case. These activities would be carried out in low risk environment for industrial partners o f any size. The scaling up o f the open, industrial SC A D A platform PVSS around the year 2000 is one example o f the success o f this approach [235]. Eventually, the com pany was integrated in the SIEMENS group and the software has been re-branded as W inC C O A to becom e the SIEMENS flagship SC A D A system on the global scale.

Finally, long-term data availability [236] has becom e an important aspect o f ensur­

ing the lasting impact o f the facility [237]. The accessibility o f several decades worth o f raw LHC data, all metadata and previous analysis results has turned out to be a m ajor topic for the IC T community. W ith a future particle collider, the time span will extend to the end o f the 21st century, calling for evolving data storage and man­

agement systems that serve the worldwide particle physics community for more than 100 years. Considering the continuous evolution o f data formats, the ever-changing particle detectors and a highly dynamic user community, data quality management is a primary topic. The value o f a particle collider research facility depends directly regulatory-compliant way, everything which contributes to safe operation. This sys­

tem also includes the procedures associated with the different operating conditions encountered during the lifetime o f the collider. A high-level computer-based safety management system integrates underlying safety related functions, including fire detection, oxygen deficiency detection, smoke and helium extraction systems, fire extinction systems, access and authorisation management, door supervision and control, video surveillance, radiation monitoring, conventional environmental mon­

itoring, evacuation signalling, supervision and control o f lifts, communication with people in underground zones, emergency lighting and acoustics and communication with emergency services (fire fighting, rescue, healthcare providers, public and private security forces). The sub-systems function autonomously.

There are two complementary systems for personnel protection in the under­

ground areas: the access safety system and the access control system. During machine operation, the access safety system protects personnel from the hazards arising from the operation o f the accelerator and from the beams. It acts through interlocks on important safety elements o f the accelerator. B y interlocking these elements it is pos­

sible to establish the right accelerator and equipment conditions in order to allow authorised personnel to access the underground installations or to allow the restart o f the accelerator equipment when access is finished. W hen the accelerator is not operating with beam and is in access mode, the access control system allows the pos­

itive identification o f any person requesting access and ensures that all pre-requisites and authorisations for that person are valid. For operational a n d /or safety reasons, the access control system also limits the number o f people present simultaneously in the underground areas.