Automating the Edge with Robotics

Edge compute IPOs are here, signaling future demand for data center capacity

Edge compute IPOs are here, signaling future demand for data center capacity

Tech IPOs have been dominating the headlines for financial news in 2019, and not always for good reasons. Uber and Lyft have both been on a downward slide since their IPO debut, in part owing to earnings reports showing losses in the billions of dollars. For the kind of potential growth story investors can love, Wall Street might want to look to edge computing IPOs instead.

Out of 24 tech IPOs so far this year, two edge computing IPOs have already occurred in 2019: Fastly and Cloudflare, although it would have been easy to overlook them as players in the content delivery networks (CDN) market. They do provide content delivery services, and they are the first two “CDN” vendors to go public since ChinaCache in 2010. (Ironically, ChinaCache is being de-listed this year from the NASDAQ exchange as others join.)

CompanyYear of IPO
Akamai1999
ChinaCache2010
Limelight2007
CDNetworks2005
Fastly2019
Cloudflare2019
*CDNetworks was acquired by ChinaNetCenter in 2017.

The near-term opportunity for Fastly and Cloudflare is the CDN market, which experienced solid growth in 2018 and is expected to grow another 6% to reach $5.3bn in 2019 according to our forecast at Edge Research Group. Looking ahead, edge computing will be a big factor in the market’s growth two to five years down the road.

But how are Fastly and Cloudflare tied to the edge computing opportunity?

The evolution of the “Edge”

There’s a strong historical and technological connection between the object caching service Akamai first offered in 1998 and today’s vision of edge computing. In fact, almost twenty years ago Akamai and others were referring to a CDN as an edge network because CDNs place resources in several dozens, even hundreds of datacenters around the globe. This gives them an advantage in proximity (and therefore, latency) to end users compared to a cloud provider with datacenters located in a relatively low number of key cities like San Jose, CA and Ashburn, VA.

Edge computing is also about using compute, network and storage services to provide better end user experience of applications and services. It’s no surprise then that for almost two years

now customers and investors have watched as many CDN vendors embrace the terminology of edge computing and serverless computing in an attempt to pre-empt the ever-expanding reach of cloud providers like Amazon Web Services and Microsoft Azure.

So, what’s the real difference between the earlier CDN edge and a modern edge compute service? In a word: control.

Creating a modern, programmable edge

In the past, CDNs maintained absolute control of their compute and network resources in edge PoPs (points-of-presence): clients had little control over resources in PoPs because running software at the edge could disturb existing services.

Now as CDNs begin to embrace edge compute, they are providing their clients with the ability to run custom code on the CDN infrastructure. Generally speaking, these services are used to augment the capabilities of an existing application, including content delivery, image resizing, global load balancing and more. They are also event driven, meaning that there is a request from a user, and a response by the application.

A modern, programmable edge service will go further by including policy-based decisions on the exact “edges” that will run an application. The ability to support stateful applications-i.e., applications that store data over time.

Fastly and Cloudflare positioning themselves for new edge services

The first new wave of functionality followed the serverless trend which allows developers to run containerized code on fully managed infrastructure; services like Cloudflare Workers offered a set of pre-defined Javascript functions for developers to use. Cloudflare evolved the service to allow developers to program in other languages and is now one of the first to support stateful edge applications with Workers KV.

From the start, Fastly allowed developers to control CDN functions to a greater degree than many competitors. Although Fastly pivoted earlier than other CDNs to the edge computing message, a fully programmable edge service is still in development. Their underlying technology, called Lucet, has been released as an open source project, though. Lucent adapted the Web Assembly compiler and runtime to work on servers (it was originally developed for running programs in a browser). This approach has the potential to provide better performance and enhanced security than some other approaches to edge application development.

The Datacenter Impact

Fastly and Cloudflare consume a fair amount of datacenter space and bandwidth. Fastly has servers in 64 datacenters around the globe and 52TB/sec of capacity. Cloudflare has a presence in 194 cities around the world.

Cloudflare and Fastly will follow Akamai’s lead in placing more servers in more datacenters in order to maximize overall performance. For several years now, these companies have been gradually deploying resources outside of the top tier datacenter markets. Akamai, the largest CDN service provider, has already invested in edge datacenter providers like EdgeConnex, which has built datacenters used by CDN providers (among other customers) in the 1 to 10MW capacity range. Looking at the next wave of services from this new generation of edge computing providers suggests they will soon be among those datacenter customers that deploy in micro-datacenters located in buildings and cell towers while typically requiring under 1MW of capacity.

In short, IPOs from companies like Cloudflare and Fastly are showing that edge computing is happening now. Tracking their revenue growth will help provide a roadmap for growing datacenter capacity at the edge.

A version of this article appeared in Interglobix Magazine.

Related Posts

Article Topics

 | 

Comments

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Featured Edge Computing Company

Edge Ecosystem Videos

Automating the Edge

“Automating

Deploying AI Models at the Edge

“Deploying

Latest News