SORCER @ US Air Force Research Laboratory

One of the missions

…of the United States Air Force Research Lab (AFRL) is to develop and assess technologies for next generation aerospace systems. To address this issue the Multidisciplinary Science and Technology Center (MSTC) within AFRL’s Aerospace Systems Directorate is developing physics-based design exploration and technology assessment methods and processes to support Air Force leadership decisions on potential system capabilities and technology investments. The Multidisciplinary Science and Technology Center is using and developing the Service ORiented Computing EnviRonment (SORCER) to address the computational framework requirements for distributed collaborative design [1].

Improving quality

To develop better quality of air vehicles it is necessary to analyze a significant number of configurations of potential prototypes. An exact analysis of each of these configurations is very complex and time consuming.According to US Air Force research the development of innovative and quality air vehicles requires a physics-based distributed collaborative design since the current approaches based on empirical data and formulas does not offer the required level of accuracy. Such a change unveils a whole new level of complexity – the computing complexity of physics-based distributed modeling. SORCER provides the innovative federated service-oriented platform to cope with the computational complexity the complexity of physics-based modeling an, the complexity of the IT infrastructure.

Saving time

  • Drastically reduce design cycle time and time-to-market, with the potential to reduce the cost of product creation – SORCER supports collaboration among distributed business teams. The service cloud can host all your business tools, thus reducing the cost and time of product creation while increasing reliability and quality. Once any of these tools are wrapped as services, models can be integrated and assembled into a federated workflow process and AUTOMATICALLY executed.
  • Higher fidelity models based on physics – to achieve an optimal design it is required to analyze a very large number of potential configurations. At the beginning of the process when the number of analyzed configurations is large, lower fidelity models are used for quick computations; however, as the number of configurations decreases higher fidelity models are used. The higher level of accuracy is achieved not only as a result of changing design parameters – but in most cases there are new parameters taken into consideration (i.e. parameters related to thermal expansion, g-force etc.) that were not analyzed before. Increased number of disciplines considered, increase the chaining and couplings of disciplines, performing design optimization considering aerodynamic, structural, and control effector design variables simultaneously requires the power of the SORCER platform to generate significant time saving.
  • Work creativity and service-oriented mash-up – only SORCER provides front-end programming and modeling languages for federated service-oriented computing [3]. It allows creating new service applications instantly and mashing-up on-the-fly results of a significant number of various engineering applications.
  • Work creativity and service-oriented mesh-up – only SORCER provides front-end programming and modeling languages for federated service-oriented computing [3]. It allows creating new service applications instantly and meshing-up on-the-fly results of a significant number of various engineering applications.
  • Decentralized federated service-oriented network environment – SORCER distributes calculations into its service-2-service (peer-2-peer) federated network leading to a much higher scalability and reliability than existing centralized computing platforms. SORCER federates dynamically all services available in the SORCER service cloud making them parallelized as needed with self-load balancing using so-called space computing developed initially for supercomputers [2]. Another timesaving comes from autonomic service administration by on-demand service provisioning/un-provisioning.

The abovementioned features of SORCER allowed the US Air Force Research Laboratory (MSTC AFRL/WPAFB) to shorten the design optimization processes 12 times [1].

Cost reduction

The essential part of cost reductions is the result of saving time described above; however, additional savings may be achieved as follows:

  • Replacement of a significant part of physical prototyping (consuming time and money) by computer simulations in a physics-based computational model. The computations are executed in the service cloud managed by the SORCER platform.
  • Better utilization of computational resources. The design optimization processes require an enormous computing power; however, through the dynamic distribution of service providers over all computers running the SORCER service cloud as well as via on-demand provisioning, new service providers can be added instantly on additional underutilized hosts (i.e. after hours) to run additional time consuming projects.

[1] R. M. Kolonay, „Physics-Based Distributed Collaborative Design for Aerospace Vehicle Development and Technology Assessment”, in 20th ISPE International Conference on Concurrent Engineering: Proceedings, 2013, p. 198;[2] Sobolewski, M., Burton, S., & Kolonay, R. (2013). „Parametric Mogramming with Var-oriented Modeling and Exertion-Oriented Programming Languages”, Proceedings of the 20th ISPE International Conference on Concurrent Engineering, C. Bil et al. (Eds.), ISBN: 978-1-61499-301-8 (print), 978-1-61499-302-5 (online), IOS Press, 2013 pp. 381-390;[3] M. Sobolewski M. and R. Kolonay, R. (2012) „Unified Mogramming with Var-Oriented Modeling and Exertion-Oriented Programming Languages”, Int. J. Communications, Network and System Sciences, 2012, 5, 9.