Choosing a deployment environment
Pexip Infinity supports a wide range of network scenarios, from small, local, private deployments to large, global, cloud-based implementations. How many Conferencing Nodes you deploy, where they are located, and the hypervisors you use will depend on the makeup and requirements of your individual organization.
While your Pexip authorized support representative is the best person to advise you on the most appropriate deployment environment, below are some of the factors to consider.
On-premises deployments use your own network, hardware and hypervisors. You will need to source appropriate servers to host the Management Node and Conferencing Nodes (for more information see Server design recommendations). You will also need to consider whether the majority of users are connecting from within your private network or whether you will have users connecting externally, and therefore if your Conferencing Nodes need to be publicly accessible. If you already have a mature IT network with datacenters in appropriate locations, this may be a suitable option.
Cloud deployments use Microsoft Azure, Amazon Web Services (AWS), Google Cloud Platform (GCP) or Oracle Cloud Infrastructure to host some or all of your Pexip nodes. This option does not require initial investment in hardware; instead you must have an agreement with the cloud service provider, and you pay for usage.
You can deploy Pexip Infinity using a combination of on-premises and cloud, with the Management Node and some Conferencing Nodes deployed within your own network, and other Conferencing Nodes deployed using a cloud service and/or in the Pexip Private Cloud, for increased flexibility. You can even set up Dynamic bursting to a cloud service to give you temporary increases in capacity whenever you need it.
There is also the option to use the Pexip Private Cloud, where Pexip deploys some or all of your Transcoding Conferencing Nodes on your behalf in the form of Pexip Smart Scale locations. This option is useful if you want the privacy and security advantages of a private cloud deployment without having to set it up yourself.
You don't have to stay committed to one particular deployment environment, and you can always increase or decrease the number of Conferencing Nodes as your capacity requirements evolve. You could, for example, start with an on-premises deployment and then add some cloud-hosted nodes later.
For more details on the on-premises and cloud options, see Network deployment options.
For on-premises deployments, you need to install a hypervisor on your servers before you can deploy the Management Node and Conferencing Nodes. Pexip Infinity currently supports four hypervisors:
- VMware is an independent product with a range of features, dependent on the licensing options you choose.
- Microsoft Hyper-V is included with Windows Server and also offers a range of features.
- KVM and Xen are open-source options that are not as feature-rich.
For more information, see Supported hypervisors.
The capacity of your Pexip Infinity deployment, in terms of how many calls can be handled at any one time, is largely determined by the number of Conferencing Nodes you deploy. To optimize the conferencing experience, we recommend that, where possible, you locate Conferencing Nodes close to each of your main user bases. For example, if the majority of your users are located in three main offices then you should consider concentrating the majority of your Conferencing Nodes among those three locations. If however the majority of your users are spread across many geographic locations, you should deploy Conferencing Nodes evenly among your datacenters.
You must also consider the role of your Conferencing Nodes, in terms of which nodes will be accessed by the client devices and endpoints that need to participate in conferences. In a large Pexip deployment with 5 or more Conferencing Nodes, or where you need to transcode media in multiple locations such as within a DMZ, you should consider deploying Proxying Edge Nodes in addition to your Transcoding Conferencing Nodes.