This article provides a comprehensive review of wide and ultrawide bandgap power electronic semiconductor devices, comparing silicon (Si), silicon carbide (SiC), gallium nitride (GaN), and the emerging device diamond technology. Key parameters examined include bandgap, critical electric field, electron mobility, voltage/current ratings, switching frequency, and device packaging. The historical evolution of each material is traced from early research devices to current commercial offerings. Significant focus is given to SiC and GaN as they are now actively competing with Si devices in the market, enabled by their higher bandgaps. The paper details advancements in material growth, device architectures, reliability, and manufacturing that have allowed SiC and GaN adoption in electric vehicles, renewable energy, aerospace, and other applications requiring high power density, efficiency, and frequency operation. Performance enhancements over Si are quantified. However, the challenges associated with the advancements of these devices are also elaborately described: material availability, thermal management, gate drive design, electrical insulation, and electromagnetic interference. Alongside the cost reduction through improved manufacturing, material availability, thermal management, gate drive design, electrical insulation, and electromagnetic interference are critical hurdles of this technology. The review analyzes these issues and emerging solutions using advanced packaging, circuit integration, novel cooling techniques, and modeling. Overall, the manuscript provides a timely, rigorous examination of the state of the art in wide bandgap power semiconductors. It balances theoretical potential and practical limitations while assessing commercial readiness and mapping trajectories for further innovation. This article will benefit researchers and professionals advancing power electronic systems.

Cloud computing infrastructure helps users to minimize cost by outsourcing data and computation on-demand. Due to the varying user needs in terms of computation power, storage capacity, etc., cloud providers offer various machines to choose from, to maximize the intended need. In this paper, we disprove several common conceptions regarding the performance and cost of cloud by experimenting on instances of two different families (compute and storage optimized) of the most popular cloud platform, Amazon Elastic Compute Cloud (EC2). Our analysis shows the interesting finding that, for the machines of the same configuration, storage optimized instances have lower disk read-write speed than compute optimized, which does not completely reflect the claim made by Amazon in all cases. Additionally, storage optimized instances have notable performance difference among them. We also identify that the I/O performance of same instance type varies over different time periods.