Future SOC Lab

The lab for the Future of Service-oriented Computing (Future SOC Lab) was a place for scientific exchange in the areas of in-memory computing, cloud computing, and non-CPU (GPU) computing. The lab provided IT resources free of charge to conduct research activities, supported diverse topics within a heterogeneous IT landscape, and connected the research community with the industry.

In 2020, we celebrated 10 years of Future SOC Lab. Up until then, more than 650 projects by 115 institutions from 30 countries have been conducted. Meanwhile, the Future SOC Lab has been advanced to the Data Engineering Lab.

Project and Industry Partners

Equipment

In cooperation with industry partners, the Future SOC Lab provided the following hardware and software.

  • Compute Power
    • High-RAM Multi-core Machines
      • HPE DL980
        • 80 cores CPU
        • 2 TiB RAM
      • HPE DL560
        • 80 cores CPU
        • 1.5 TiB RAM
    • Distributed Computing
      • 1000 Core Cluster
        • 22 nodes, each:
        • 12 cores CPU
        • 128/256 GiB RAM
      • HPE Moonshot
        • 10 cartridges, each:
        • 8 cores CPU
        • 64 GiB RAM
    • Virtual Computing
      • Virtual Machine Cluster
        • 3 nodes, each:
        • 48 cores CPU
        • 1 TiB RAM
  • GPU Power
    • Nvidia K GPU
      • K20/K80 (HPE/SuperMicro)
    • Nvidia DGX-1
      • Tesla V100
      • 1000 terraflop
      • 40960 cores
  • Storage and Backup
    • HP 3PAR (as part of HP CC)
      • ~100 TiB storage
  • Networking
    • 10G/40G interconnect
      • Fujitsu and Brocade switches
        • 10G server connection and storage connection
        • 10G storage connection
        • Internet uplink via X-WiN of DFN (German Research Network)
      • 40G router interconnect for core switching

  • SAP HANA

    • For single HANA installations on shared hardware, the following hardware was available:

      • Fujitsu RX900 S1
        • 8 x Xeon X7560 CPU
        • 1024 GiB RAM
        • 6 x 900 GB HDD
        • 1280 GB Fusion-io ioDimm3 SSD
      • Hewlett Packard DL980 G7
        • 4 x Xeon (Nehalem EX) X7560 CPU
        • 1024 GB RAM
        • 2 x 300 GB HDD
        • 1280 GB Fusion-io ioDimm3 SSD

Team

The Lab was mainly managed by two employees, who provided the reliable and well-usable infrastructures. Bernhard Rabe and Dr. Tobias Pape set up and maintained the infrastructures, answered user questions, and advised on how to optimize using the lab.

Events

Project results were exchanged on regular, but also sporadic events. The events served as an exchange platform between project and industry partners.

Project Procedure

Applicants and projects were asked to adhere to the following procedure.

1. Notice Call for Proposals

We invited applicants to work on their research questions using the IT landscape of the Future SOC Lab. Proposals could be submitted for the upcoming phase.

Topics of Interest

Topics of interest included but were not limited to:

  • Machine learning
  • Blockchain
  • GPU computing
  • Application containerization
  • Unikernels
  • Service-oriented Computing (SOC)
  • Microservices
  • Cloud computing
  • In-memory database technology
  • Multi-core architectures
  • On-demand delivery models for business applications
  • and other …

Resource Utilization

After approval of a project proposal, the research projects received free of charge access to the IT infrastructure of the Future SOC Lab until the next HPI Future SOC Lab Day (approx. 6 months). Projects could additionally request exclusive access to certain resources.

Roles and Expectations

When submitting a project proposal, we asked for contact information of two persons.

  1. Principle Investigator (PI): Usually a Chair holder / Professor / Lecturer / Research Group Lead / or similar, had to accept the terms of use on behalf of the institution. This person had access to the Future SOC Lab portal and, among other things, could manage the project: Process the request, upload documents (e.g., reports and posters), request access for team members, and submit continued project requests.
  2. Contact Person (CP): Usually a scientific employee / research project lead served as the contact person for the project. This person had access to the Future SOC Lab portal and had the same rights as PI. In addition, they was responsible for communicating information and reminders to all team members.

Another role is the role of the team member. Team member had access to the IT resources, but not to the Future SOC Lab portal.

Email communication

We used an email distribution list for information that all projects needed to receive. This list contained the Principle Investigator and Contact Person. The Contact Person was responsible for ensuring that all project participants were informed.

2. Receive Decision

The steering committee of the Future SOC Lab, which comprises representatives of the HPI and the industry partners, decided about the approval of project proposals. This happened during the Future SOC Lab Day.

3. Access Resources

Accepted research projects were granted access to resources. If further resources were needed during or after a project, a new project proposal needed to be submitted.

At the end of a project phase, all project-related data has been deleted and access to resources has been deactivated.

4. Document Project Results

At the end of the project phase, projects had to submit a technical report and a scientific poster on their research activities. The reports were published once a year through the university publisher. The posters were presented at the next Future SOC Lab Day.

Reports, posters and presentation were asked to include the project idea, the project activities, the resources used, the results, a discussion, and next steps.

5. Attend Future SOC Lab Day

Research projects presented their work to the Future SOC Lab community and all interested parties. The event gave insights into latest research questions and trends in IT companies.