The business internet is broken. But there’s hope to fix it by rethinking its core value.
Discover why the business internet is broken and how a reimagined approach can fix it.
The business internet is broken. Wait, what? Enterprises run their entire business on the Internet. More than 5 billion people are on the internet, and business internet traffic has increased fivefold in the last seven years alone.
But Graphiant’s 2023 State of the Network Edge survey paints a clear picture. When asked how well the top three types of internet connectivity fared at various important metrics, network architects flunked the entire lot:
If enterprises aren’t getting the performance, security, privacy, and scalability they need, it’s fair to say the business internet is broken. The question becomes, why? How did we get here?
We need to return to the internet’s beginning to understand how we got here. The original internet (then called Arpanet) had one singular purpose: Connectivity. There were no services in the Cloud (in fact, there was no Cloud). Traffic was peer-to-peer.
So, two milestones were reached when Charley Kline typed a short message from his computer at UCLA to Bill Duvall’s computer at Stanford Research Institute. First, the world achieved its first internet connectivity. And the system crashed after typing just two letters (“Lo” of the word “Login”)—the world’s first internet outage.
Humble beginnings, to be sure, but consider this: The internet was born just three months after humans walked on the moon. The Apollo mission gave us so much, but it’s hard to argue that the internet hasn’t had the bigger impact.
Consumers Take Over
It’s difficult to remember a world without the internet (impossible, for 40% of the population born after the internet emerged). But for 22 years (1969 to 1991), the internet was restricted to government, research, and higher education.
That changed when the Internet opened to the public in 1991. Tim Berners-Lee invented a crucial piece of software called the World Wide Web. It is the foundation that allows websites to be built and consumed. It is the “www” in a website’s address. It is often shortened to just “the web.”
And with that one invention, consumers took over the internet. As websites grew (from 1 in 1991 to 17 million by 2000), the internet shifted from peer-to-peer to client-server. Instead of UCLA connecting to SRI to directly transfer some data, millions of consumers (clients) are connected to a smaller number of very popular websites (servers). The 1990s incubated the world’s largest online companies (Amazon launched in 1995, Google launched in 1998, and Facebook launched in 2004).
To a smaller extent, this was also happening for business users. For example, Salesforce launched in 1999. But in these early days, businesses were more interested in using the internet as an underlay network to connect different parts of their business (data centers, branch offices, remote workers).
Tension emerged between businesses (whose traffic was still predominantly peer-to-peer) and consumers (whose traffic was client-server). Yet, the consumer client-server model won out because most internet traffic was consumer traffic flowing to massive sites like Facebook, Google, and Amazon.
Then, in late 2004, something revolutionary emerged, pushing business to the client-server.
In November 2004, Amazon introduced AWS, and the Cloud was born. This was the impetus for businesses to adopt the client-server model. Important business SaaS apps like Microsoft 365, Google Workspace, and Zendesk came online. At the same time, businesses began shifting enterprise data and applications to the Cloud.
Total cloud spending is nearing $700B. Business internet traffic has shifted to peer-to-peer (between enterprise resources) and client-server (to Cloud apps like Microsoft 365).
But a different trend is making client-server more difficult: massive growth in internet traffic. More than 150 exabytes of data transit the internet every month. That’s 150 billion gigabytes and is forecast to grow 24% a year. And that growth is causing a reexamination of where we stage data and applications worldwide.
The Three Laws of Internet Traffic
Data and workloads have lived in the data center or the cloud for a long time. But that’s changing due to three fundamental laws:
- Laws of physics: It takes time for data to travel from point A to point B – that’s just simple physics. The further the distance, the longer the trip takes. This introduces latency. And while latency is annoying for consumers looking at Facebook, it can be a critical problem for business apps like communications, financial trading, and telehealth.
- Laws of economics: Bandwidth is expensive, especially for the traffic volume we see today. Where data and applications are staged has large implications for bandwidth costs (and who bears those costs).
- Law of the land: As the internet has grown, privacy concerns have also grown. In the last decade, laws have emerged worldwide to address data sovereignty. Things like GDPR in the EU and CPRA in California mandate strict rules on storing and moving personally identifiable data.
These factors challenge the traditional client-server model. The highly concentrated model of billions of users connecting to a small number of servers (low hundreds or thousands) is replaced with a model that moves data and applications closer to users. This is creating exponentially more staging nodes around the Internet.
Let’s see how this works. Business uses last-mile connectivity to access their closest regional service provider. The regional SP acts as the centralized distribution point for the business. Some services are served locally (perhaps Microsoft 365 has a presence in that regional SP). Some services require branching out from the regional SP to travel to wherever the required data or application is located.
Essentially, we’re moving to a model where the business connects to an application exchange service with last-mile connectivity. It’s complex, and it creates an enormous amount of work for the enterprise. But what is coming next is what has truly killed the business Internet.
The Edge Will Eat the Business Internet
Tom Bittman, a Gartner analyst, wrote a controversial blog predicting workloads and data would move away from the Cloud and out to the edge. Today we are seeing his prediction come true.
A sizeable and growing set of workloads (and data) are better off running at the edge. It may be for latency reasons, data sovereignty, or too much data to backhaul to the centralized Cloud.
Analysts say the edge computing market will grow at 48% CAGR to become a $445B market by 2030. But as this happens, is the edge eating the Cloud as Bittman predicted? Not so much – the Cloud is also forecast to grow during this same period.
But the edge is eating the business Internet.
The internet of 20 years ago had a highly centralized client-server model of hundreds of millions of users (i.e., clients) accessing a relatively small number of massive websites like Facebook (i.e., servers). The internet of today has moved workloads and data to, well, everywhere. It has slid back to a peer-to-peer network.
But peer-to-peer over today’s internet requires complex tunnels for every peer-to-peer connection. How can enterprises manage all of these tunnels as connections explode? They can’t, of course, which is why the business internet is broken.
But it can be fixed.
Fixing the Business Internet
To fix the business Internet, we have to understand one fundamental fact: The core value of the Internet has changed.
When the internet was conceived, its value was connectivity. You didn’t need to build a network – you relied on the internet for that. You simply connected to the Internet, and in exchange, you got connectivity.
But the core value of the business internet today is accessibility. Everything a business needs is connected to the Internet. The fact that we use the internet for connectivity to those things we want to access is no longer the Internet’s core value.
Using the internet for connectivity is the core problem. To fix the business Internet, we need to replace the connectivity component of the Internet with something better.
Here is what that looks like. The last mile is all we need for accessibility. Every entity (Cloud, SaaS vendors, businesses, IoT devices, etc.) will use last-mile connectivity to access the Internet.
But we replace the middle mile with a private network as-a-service. A network with guaranteed security, data sovereignty, performance, and scalability advertises services it can directly connect with instead of relying on transit providers. Most importantly, a network that doesn’t require the business to build tunnels for every connection.
Instead of days or weeks to painstakingly build new peer-to-peer connections, the business simply provisions the connection in seconds from a simple cloud portal.
And remember when we discussed how we’re moving to a model where the business connects with last-mile connectivity to an application exchange service? That becomes the job of the new middle mile.
The business internet is broken, but we can fix it by understanding what to change.
This article was originally published on Spiceworks