An epitaxial substrate having a 2D material interposer, the epitaxial substrate extending along an epitaxial interface direction, wherein the epitaxial substrate includes: a polycrystalline base substrate having a superficial layer, a wafer bevel, and a back surface, wherein a difference in coefficient of thermal expansion between the polycrystalline base substrate and MN or GaN is not greater than 1.510.sup.6 C..sup.1 in a direction parallel to the epitaxial interface; a multi-orientation 2D ultra-thin material interposer arranged on the superficial layer of the polycrystalline base substrate, wherein the multi-orientation 2D ultra-thin material interposer has a top layer, a lattice constant of the top layer being highly matched with that of AlN, AlGaN, or GaN; and an AlN, AlGaN, or GaN-based epitaxial layer, which is epitaxially grown on a portion of the multi-orientation 2D ultra-thin material interposer distant from the polycrystalline base substrate.

A laser-based light source includes a material arranged on a package base adjacent to a laser diode chip and an optical element coupled to the material. The optical element is aligned to receive electromagnetic radiation from the laser diode chip. The optical element includes a wavelength conversion material and is configured to receive at least a portion of the electromagnetic radiation emitted by the laser diode chip. A reflective material surrounds sides of the optical element.

A method for flattening a surface on an epitaxial lateral overgrowth (ELO) layer, resulting in obtaining a smooth surface with island-like III-nitride semiconductor layers. The island-like III-nitride semiconductor layers are formed by stopping the growth of the ELO layers before they coalesce to each other. Then, a growth restrict mask is removed before at least some III-nitride device layers are grown. Removing the mask decreases an excess gases supply to side facets of the island-like III-nitride semiconductor layers, which can help to obtain a smooth surface on the island-like III-nitride semiconductor layers. The method also avoids compensation of a p-type layer by decomposed n-type dopant from the mask, such as Silicon and Oxygen atoms.

Top-of-rack (TOR) switches are connected to a network fabric of a data center. Each TOR switch corresponds to a respective rack of the data center, and is configured to provide access to the network fabric for computing devices mounted in the respective rack. A request is received, from a client device via a portal, to configure a first rack of the data center. Configuration data is received, from the client device, for a first virtual network to be accessed by a first computing device mounted in the first rack. In response to receiving the configuration data, a first TOR switch of the first rack is configured, which includes associating the first virtual network with the first TOR switch.