Migrating to a cloud computing platform such as Microsoft Azure offers organisations of all sizes and industries a myriad of significant benefits, ranging from reduced IT costs to greater scalability, enhanced security features, simplified management, and more. An increasing number of organisations are therefore considering the prospect of moving their operations to the cloud as they seek to unlock a competitive advantage.
However, a cloud migration can sometimes pose a formidable technical challenge for an organisation, especially if carried out on a large scale.
“The first challenge is bridging on-premises networks to Azure, and achieving that with security, reliability and scale,” says Brent Yates, CEO of SwXtch.io. “Azure is very capable of moving packets of data reliably at high speeds as long as you need Vanilla transmission control protocol or user datagram protocol standards. However, we often hear from customers that clouds do not support the necessary networking features required to implement their specialised workloads. This frustration is due to customers expecting cloud networks to work like on-premises networks, even though they don’t.”
SwXtch.io provides CloudSwXtch, a virtual overlay network, which lets an organisation moving to the cloud build a high-performance network on top of its cloud network without changing its application code. As Yates explains, this can help ease the transition between an on-premises and cloud network.
“We make the cloud look like an organisation’s on-premises environment,” says Yates. “We help businesses turn on the same features in cloud networks that users see from physical switches, routers, network interface cards and servers, and give them the power to configure these network features and capabilities to their specifications. Adding CloudSwXtches to their cloud networks gives businesses control of the network and full visibility into packet flows that aren’t available through Azure’s network monitoring tools.”
A crucial feature of a CloudSwXtch network is its ability to enable multicast on the cloud. Multicast is a method which allows a producer of data to send that data to multiple destinations without having to physically send multiple copies or manage the list of destinations. This means that the sending application doesn’t have to know or deal with how many destinations exist, and also that any destination can dynamically leave and join the flow of data. In addition, the network is only moving the packets that are necessary, which results in increased efficiency.
In an on-premises environment, multicast is handled by hardware switches and routes. However, this is difficult to replicate in the cloud.
“In cloud networks, you don’t have direct access to the physical switches and routers, so multicast is not available,” explains Yates. “CloudSwXtch, however, implements multicast over the top of the cloud networks by creating a layer of switches (the CloudSwXtch). Along with our xNIC software, CloudSwXtch gives our customers a high-performance network that they can control with full standards compliant multicast and no software changes are needed for their applications.”
For many organisations, multicast is a necessity that represents a major barrier to cloud migration and its many benefits.
“Many industries require multicast because their existing workloads require it, so its absence makes it impossible for them to move their systems into the cloud,” says Yates. “Multicast also makes designing and testing certain communications applications much easier, which is why engineers design systems around it. Finally, industry standards require it. In media and entertainment particularly, the SMPTE 2110 standards require multicast as a distribution mechanism. This is because multicast allows multiple streams or content sources to be independently distributed, and then rejoined and synchronised later in the process.”
With CloudSwXtch, an organisation can also use multicast to enhance the cost-saving benefits that help make cloud technology such an attractive proposition.
“By moving a portion of their workload into the cloud, a business can optimise on-premises versus cloud spend,” says Yates. “However, this is not possible for a lot of workloads that are dependent on multicast – unless multicast works in both environments. By arranging their network topology to send a single copy across links that are billed based on the amount of data, then doing the fanout at the far edge, a business can minimise their data transfer fees.”
This article was originally published in the Summer 2023 issue of Technology Record. To get future issues delivered directly to your inbox, sign up for a free subscription