Abstract. System modules are considered as the main parts of specialized computing systems. An approach to development of scalable embedded computers based on self-similar system modules stacked on top of each other without extra cables, is presented. The existing system module specifications are optimized for using a single module per one embedded computer that makes it difficult to increase system performance by using multiple modules without significant reconstruction of the carrier board. The proposed alternative approach uses isomorphic (self-similar) modules that can be mated using planar board-to-board hermaphrodite connectors. If assembled in a stack, these modules form a compact computing cluster. New approach also introduces a dynamic system connector pin allocation based on routing system signals through a FPGA. The dynamic allocation improves the efficiency of system connector pins utilization, and allows to get a wider range of hardware interfaces implemented using fewer pins of system connector while maintaining the modules compatibility with various application systems. This is achieved by flexible software controllable FPGA configuration depending on the module neighbors. Special system microcontrollers inspect overall system compatibility by using data stored in electronic passport of each module, so that the computing system can be automatically inspected for conflicts and correctly configured before powering up the main processors.

Keywords: system module, mezzanine, carrier board, embedded computer, system bus, system microcontroller, FPGA-based switch, dynamic configuration of system connector

References:

1. Zhao H., Dong Z.Y. International Conference on Information System and Artificial Intelligence (ISAI), IEEE, 2016, рр. 488–491.
2. Min G., Xiucai Z. 10th International Conference on Electronic Measurement & Instruments, IEEE, 2011, vol. 2, рр. 99–103.
3. Kabović A.V., Kabović M.M., Ćelebić V.V. 22nd Telecommunications Forum Telfor (TELFOR), IEEE, 2014, рр. 423–426.
4. Davis T. et al. International Journal of Next-Generation Networks (IJNGN), 2010, Vol. 2.
5. Vásquez J. et al. EPICS IOC Based on Computer-On-Module for the LNL Laboratory, 2015.
6. Pedretti D. et al. Custom Hardware Platform Based on Intel Edison Module, 2015.
7. Klimeck D. et al. 29th International Conference on Application-specific Systems, Architectures and Processors (ASAP), IEEE, 2018, рр. 1–4.
8. Wegerson M., Straub J., Marsh R. University of North Dakota Graduate School Scholarly Forum, 2015.
9. Samson J.R. Aerospace Conference, IEEE, 2011.
10. Giraudi L. Development of a SMARC module for an ADAS system based on the i. MX8 processor, Politecnico di Torino, 2018.
11. Herbrechtsmeier S., Rückert U., Sitte J. Advances in Autonomous Mini Robots, Springer, Berlin, Heidelberg, 2012, рр. 101–112.
12. Dashevsky V.P., Bizin М.М. Proceedings of the TUSUR University, 2015, no. 3(37).
13. http://www.congatec.com/fileadmin/user_upload/Documents/Technical_Documents/etxspecv3.02.pdf. (in Russ.)
14. COM Express Carrier Design Guide, https://www.picmg.org/wp-content/uploads/PICMG_COMDG_2.0-RELEASED-2013-12-06.pdf.
15. https://www.sget.org/fileadmin/user_upload/SMARC_Hardware_Specification_V200.pdf. (in Russ.)
16. https://www.sget.org/fileadmin/file_management/SDT02/Qseven-Spec_2.1.pdf. (in Russ.)
17. https://pc104.org/wp-content/uploads/2015/03/PCI104_Express_v3_0.pdf. (in Russ.)
18. http://www.stackpc.org/content/files/specs/Specification_StackPC_rev_1_2.pdf. (in Russ.)
19. Nusev G. et al. International Conference on Decision Support System Technology, Springer, Cham, 2018, рр. 157–168.
20. Dashevsky V.P. Industrial Automatic Control Systems and Controllers, 2019, no. 1, pp. 3–14. (in Russ.)
21. http://suddendocs.samtec.com/catalog_english/lshm.pdf. (in Russ.)