Way back when I left the world of network engineering and became an analyst (circa 1999) my first assignment was to write a technology brief on the topic of IP multicast. At the time, IP multicast was emerging as a means to optimize the delivery of streaming video across enterprise data networks. In most cases this was to support the broadcast of town halls or executive addresses to employees. IP multicast solved a very real problem for one-to-many broadcasts: How to ensure that streams didn’t overwhelm the network.
Consider a scenario without multicast: A company broadcast to share the latest product news is streamed via a webinar app. Without multicast, each attendee individually connects to the webinar and receives their own audio and video stream. Considering that a high-definition broadcast could require between one and four MBs of bandwidth, one can easily see how a large-scale company event, with thousands of participants receiving their own individual video streams, could quickly overwhelm the WAN and potentially even the LAN(especially Wi-Fi). With large numbers of simultaneous video broadcasts, the potential for congestion exponentially rises.
In the early days of networking, the solution to this problem was IP multicast. With IP multicast, network administrators configure routers and switches to optimize and pare streams so that rather than each endpoint individually connecting to a streaming server, the network would transmit a single stream to each location and individuals would connect to their local switch or router to obtain the stream. While it sounded good on paper, early IP multicast implementations were often buggy and required constant attention from network managers. It wasn’t unusual for networks to be locked down, and an “all-hands-on-deck” situation to occur during one of these town halls as network administrators fought to keep the stream alive. Obviously this approach wouldn’t scale given the large scale use of video streaming in our current world.
Over the years IP multicast has been replaced with enterprise content delivery networks, or ECDNs. ECDNs function in much the same way as IP multicast, but rather than relying on configurations within routers and switches, ECDNs use appliances or virtual software that sits on top of the data network, plus a variety of approaches including caching and peer-to-peer technologies to replicate streams only to those who wish to receive them. Going back to our earlier scenario, with an ECDN, employees attending a company webinar would connect to their local ECDN caching server, or receive the broadcast from other devices on their local network rather than connect to the webinar across the WAN.
On the public Internet, CDN services from companies like Akamai and Cloudflare have long been used by streaming services (e.g. Netflix, Hulu, etc.) to reduce network impact and improve delivery performance for mobile and residential subscribers. In the enterprise, ECDNs provide a number of additional capabilities the enterprise might require, including secure stream delivery, monitoring and performance management, and the ability to optimize delivery of additional one-to-many data flows for IT upkeep such as application updates.
ECDNs lost their luster over the last few years due to the pandemic driven shift to remote work. However, as companies have increasingly brought employees back to the office, and as video streaming application use continues to increase, so has the interest in ECDNs. Metrigy’s Workplace Collaboration MetriCast: 2022 global study of more than 930 end-user organizations found that almost 30% of participating organizations were already using an ECDN while another 32.2% planned to deploy one by the end of 2023. Of those already using ECDNs, 44% planned to increase their utilization. Those using ECDNs were seeing tangible benefits with an average WAN savings of 22%.
Late last year the ECDN market became active with Zoom launching its own ECDN (Zoom Mesh) and Vbrick acquiring ECDN vendor Ramp to broaden its already existing ECDN capabilities. A year and a half before that Microsoft acquired a Peer-to-Peer ECDN vendor by the name of Peer5 to use its WebRTC-based technology to optimize delivery of streams from Microsoft Teams live events. Other leading vendors in this space include Hive, Kollective, and Riverbed.
If your organization is returning to the office, and you are looking for ways to reduce the network impact of video streaming, application delivery, and content delivery while improving voice and video performance, an ECDN may be right for you. Evaluate competing services based on their ability to support your video streaming apps and services, and their flexible delivery models that provide support for cache and peer-to-peer delivery. Consider ECDNs as a way to avoid the cost of additional bandwidth by optimizing the bandwidth that you have today.
ABOUT METRIGY: Metrigy is an innovative research firm focusing on the rapidly changing areas of Unified Communications & Collaboration (UCC), digital workplace, digital transformation, and Customer Experience (CX)/contact center—along with several related technologies. Metrigy delivers strategic guidance and informative content, backed by primary research metrics and analysis, for technology providers and enterprise organizations.