Task grouping and optimized deep learning based VM sizing for hosting containers as a service

Journal of Cloud Computing

Advances, Systems and Applications

Table 6 Comparative Assessment of Deep-ConvLSTM_FPO Scheme With Task Size 100

Number of Iteration	VM size selection technique	Hurst exponent+Markov transition	ICLB	OM-FNN	VM size IaaS multi-tenant public cloud	Deep-ConvLSTM GD	Deep-ConvLSTM ADAM	Proposed Deep-ConvLSTM_FPO
Resource Utilization
5	0.0397	0.0381	0.0325	0.0291	0.0289	0.0286	0.0283	0.0275
10	0.0567	0.0407	0.0406	0.0387	0.0377	0.0374	0.0370	0.0365
15	0.0498	0.0454	0.0454	0.0408	0.0400	0.0387	0.0381	0.0375
20	0.0451	0.0430	0.0405	0.0399	0.0397	0.0380	0.0369	0.0355
Response Time
5	18188	13184	12420	8456	6225	892	567	277
10	17784	14780	11985	8578	6193	945	599	269
15	17801	15798	11980	8635	6118	987	634	263
20	17758	16755	11982	8743	6129	1034	678	262
Task Rejection Rate
5	0.2055	0.1978	0.1848	0.1808	0.1785	0.1765	0.1728	0.1565
10	0.2214	0.2055	0.1985	0.1848	0.1794	0.1764	0.1737	0.1659
15	0.2413	0.2254	0.2054	0.2016	0.1990	0.1958	0.1937	0.1875
20	0.2855	0.2541	0.2325	0.2257	0.2143	0.2137	0.2110	0.1987
Makespan
5	0.7876	0.7578	0.7145	0.6965	0.6876	0.6754	0.6543	0.6367
10	0.7976	0.7755	0.7356	0.7154	0.7078	0.6865	0.6754	0.6145
15	0.8156	0.7987	0.7467	0.7357	0.7267	0.7076	0.6875	0.5967
20	0.8478	0.8257	0.7765	0.7578	0.7476	0.7245	0.7087	0.5788