Where academic tradition
meets the exciting future

Dynamic Horizontal and Vertical Scaling of Multiple Cloud Services in Soft Real-Time

Benjamin Byholm, Ivan Porres, Dynamic Horizontal and Vertical Scaling of Multiple Cloud Services in Soft Real-Time. TUCS Technical Reports 1182, TUCS, 2017.

Abstract:

We approximately solve the problem of dynamic horizontal and vertical scaling of multiple cloud services in soft real-time. We recognize that this is a generalized bin packing problem combined with a linear assignment problem. We formalize the problem domain and develop an autonomous resource management system with soft real-time constraints. The system incorporates a genetic algorithm with quadratic expected time complexity for solving the packing problem, providing a service deployment plan with an optimal number of servers, and a successive shortest path algorithm with quadratic expected time complexity for finding an optimal way of realizing the obtained deployment plan.

Files:

Full publication in PDF-format

BibTeX entry:

@TECHREPORT{tByPo17b,
  title = {Dynamic Horizontal and Vertical Scaling of Multiple Cloud Services in Soft Real-Time},
  author = {Byholm, Benjamin and Porres, Ivan},
  number = {1182},
  series = {TUCS Technical Reports},
  publisher = {TUCS},
  year = {2017},
  keywords = {Cloud computing, Consolidation, Elasticity, Packing},
  ISBN = {978-952-12-3560-3},
}

Belongs to TUCS Research Unit(s): Software Engineering Laboratory (SE Lab)

Edit publication