ISSN 0021-3454 (print version)
ISSN 2500-0381 (online version)
Menu

2
Issue
vol 67 / February, 2024
Article

DOI 10.17586/0021-3454-2017-60-4-295-301

UDC 519.687.1

SCHEDULING SETS OF COMPOSITE APPLICATIONS WITH SOFT DEADLINES IN TIME WINDOWS OF HETEROGENEOUS COMPUTING ENVIRONMENTS

K. O. Bochenina
ITMO University, Saint Petersburg, 197101, Russian Federation; junior researcher


N. A. Butakov
ITMO University, Saint Petersburg, 197101, Russian Federation; junior researcher


A. V. Boukhanovsky
ITMO University, Saint Petersburg, 197101, Russian Federation; Director


Read the full article 

Abstract. The loading performance of heterogeneous computing resources under user deadline requirements to results of calculation necessitates the use of specialized planning algorithms of the tasks execution. Such algorithms should consider both the specifics of the distributed environment (the level of utilization, types of resources) and possible data relationship between parts of the design scenario. The proposed algorithms implement per-task, phasic and clusterization approaches to planning sets of composite applications with a soft deadline requirements in conditions of partial availability of computing resources. Comparative efficiency of the developed algorithms for scheduling sets of the composite applications on heterogeneous resources is investigated. Presented experimental results demonstrate that: a) the use of information about the availability of resources and the tasks completion timing can significantly improve the scheduling quality; b) clusterization approach is superior to stage and per-task according to integral criterion of scheduling efficiency.
Keywords: scheduling algorithm, cloud media, composite applications, resource load scheduling, time windows

References:
  1. Goble C., Roure D.De. Fourth Paradig. Data-intensive Sci. Discov., Hey T., Tansley S., Tolle K. (eds.), Microsoft Research, 2009, рp. 137–145.
  2. Zhao H., Sakellariou R. Parallel Distrib. Process. Simp. IPDPS’06, Washington, IEEE Computer Society, 2006, рp. 14.
  3. Bittencourt L.F., Madeira E.R.M. J. Grid Comput., 2009, no. 3 (8), pp. 419–441.
  4. Masko L et al. Scheduling moldable tasks for dynamic SMP clusters in SoC technology, parallel processing and applied mathematics, PPAM 2005, Poznań, Poland, 2005, Springer, Berlin, 2005, no. 3911, pp 879–887.
  5. Hirales-Carbajal A. et al. J. Grid Comput., 2012, no. 2(10), pp. 325–346.
  6. Bochenina K. Procedia Comput. Sci., Elsevier Masson SAS, 2014, vol. 29, рp. 509–522.