Why do the 5 points for Network Architecture 2020 matter?
1. Application Delivery
The network is merely a construct to provide access to - to deliver - applications to end users. Application delivery is the purpose of any network. Without this, the network is pointless, an expensive waste of time, copper and silicon.
Whether the application is streaming video, web browsing, gaming or anything else, networks in 2020 must excel at delivering them.
How can the network excel at delivering applications? Although this varies by application, the consistent factors are:
- Speed - that is, uplink and downlink data throughput, sufficient to support the application. If the application needs 5 Mbps average downlink capacity, the network had better be capable of providing it. Otherwise, stuttering applications and unhappy users lie ahead.
- Latency - or rather, the lack of it. User feedback consistently shows low latency as key for web browsing and real-time applications, and is more beneficial than increased bandwidth in most cases for these applications.
- Application-specific optimisation - where possible, any special sauce that can be added such as video compression or data de-duplication to give that one application a useful boost in user experience.
- Quality of service - we have this today, right? Well, yes and no. We have relatively blunt mechanisms to implement quality of service, often relying on strict priority levels assigned to individual network packets. Smarter, more application-aware quality of service will greatly benefit us all in 2020.
For the network, application delivery isn't just job 1 - it's the only job.
Without it, the network is meaningless, no matter how fast, cheap or big it is.
Smarter networks that focus on application delivery for specific high-value applications will overtake their beefier, more-provisioned rivals.
2. Service Agility
New applications are being created and launched faster than ever on today's networks, many requiring significant infrastructure to support them.
Think of launching a new gaming service, something to rival the likes of Steam. You need a Content Delivery Network (CDN), where the huge volumes of data that are the games are cached as close to your users as possible.
Next, you need a robust, scalable shopping and payments infrastructure. After all, it's no good if people can download your games when they can't buy them from you in the first place.
What if you're also providing multiplayer online game servers? You need a whole bunch of them across the country (or world) to give your players the lowest latency possible.
Sounds like a lot of work, right? It absolutely is. Now think of the secure gateways and other expensive one-trick boxes you'll need to make this work. 6 months later, you need to switch gears and do something different.
Oh dear, half of the investment you just made is sunk. Do not pass go.
To get more competition and better services, the barrier to creating new services needs to be lowered. Less time, less money and less people must be needed in 2020.
3. Topological Flexibility
In 2020, networks will need to adapt and work with their surroundings, as well as be able to handle unexpected changes.
Think of a mobile phone network. Today, a single large macrocell typically covers an area. To bring greater capacity, particularly to urban areas, the topology must change to a greater number of smaller cells, deployed closer to their users.
If the network is unable to accommodate topological changes like this, it will be forever at the mercy of prevailing trends. When a huge investment is required at a fundamental level to accommodate topological change, you can be almost certain it will be at the back of the queue, all the whilst losing ground to rivals.
For another example, think of the vision of millions of Internet of Things (IoT) devices connected to the network. Taking a kilobit of data from a million devices is a very different topology to taking a million kilobits from one device.
Ultimately, a lack of topological flexibility stunts the capabilities of the network for the future. How many service providers want to spend billions building a network to find it's out of date in two years?
4. Economical Expansion and Operation
This may seem obvious, but it's worth emphasising. After all, if providing network service to one area or improving coverage in another is technically feasible but far too expensive to make business sense, it won't happen. Everyone loses.
Building Network Architecture 2020 around technologies that provide manageable deployment and operational costs must be a key objective to ensure that the network can be where we need it, bringing the benefits to as many people and places as possible.
5. Value Chain Advancement
Okay, it's not very fashionable to suggest that service providers should make more money. Wait, stop throwing things, let me explain.
Today, network connectivity is a given. People expect it, in various grades and speeds, but it is essentially a commodity. Providing access to applications, via a hugely expensive national or international network, is considered the lowest-value part of the chain, with the applications themselves at the top.
So why is this a problem? Well, becoming a commodity means the prices for connectivity are going to be pushed down, at a time when we need service providers to invest to ensure they are prepared for 2020.
If service providers can't move up the value chain - providing their own high-value services in addition to pure network connectivity, such as service provider-owned video services, two things are likely to happen:
- Network connectivity gets more expensive for everyone.
- Investment in new network infrastructure slows down.
It's safe to say that neither of these are things we want.
So, there you go. To fully meet the needs of 2020, Network Architecture 2020 needs to fulfil all 5. Otherwise, missed opportunities and unhappy users await.
what's the next step?
Explore some of the technologies that can help us fulfil all 5 points and prepare our networks for 2020.