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

Journal of Cloud Computing

Advances, Systems and Applications

Table 9 Comparative Assessment of Deep-ConvLSTM_FPO Scheme With Task Size 400

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	0.1522	0.1460	0.1347	0.1331	0.1328	0.1312	0.1301	0.1279
10	0.1842	0.1675	0.1525	0.1494	0.1476	0.1437	0.1401	0.1325
15	0.1798	0.1714	0.1568	0.1537	0.1519	0.1501	0.1500	0.1479
20	0.1963	0.1748	0.1619	0.1595	0.1568	0.1537	0.1517	0.1488
Response Time
5	68881	58878	46085	40123	23119	15457	981	634
10	68910	61906	46115	40475	23292	15843	1036	697
15	68765	62762	45951	40734	23171	16284	1091	619
20	68984	63981	46207	40982	23250	16790	1158	693
Task Rejection Rate
5	0.4585	0.4413	0.4326	0.4235	0.4126	0.4087	0.4037	0.3855
10	0.4659	0.4585	0.4488	0.4458	0.4325	0.4268	0.4179	0.3959
15	0.4854	0.4785	0.4585	0.4529	0.4458	0.4379	0.4268	0.4014
20	0.4986	0.4876	0.4785	0.4654	0.4587	0.4479	0.4368	0.4113
Makespan
5	0.8765	0.8578	0.8236	0.7967	0.7865	0.6654	0.6467	0.6256
10	0.8578	0.8467	0.7976	0.7867	0.7755	0.6976	0.6654	0.6578
15	0.8367	0.8076	0.7765	0.7654	0.7578	0.7367	0.7245	0.6976
20	0.8087	0.7865	0.7578	0.7467	0.7367	0.7268	0.7226	0.7156