Networks used to be more complex to deploy but had a simpler architectural design. Data and applications were either run by the provider, by the customer on their server, or a mix of both.
With the rise of cloud, ‘where is my application hosted?’, and ‘how do I connect to it?’ are two of the most important questions in networking.
Employees are also no longer where they were before: now they are on the move, remote working or in the office. Distribution and Cloudification have become the buzzwords of the networking industry.
Table of Contents
What is a distributed network?
A distributed network is any network that is not contained within a defined area (e.g., an office, a factory, a campus).
Your office will have a network, but it is not distributed. But if you had a network that spanned many offices and locations, it would be a distributed network.
Before, a distributed network would have been one spread across many business locations.
Network distribution’s definition changed during the Covid-19 pandemic.
Now the idea of a distributed network must factor in houses, regional and home offices, and more.
What is cloudification?
Cloudification is the adaption of applications and process to cloud computing.
This involves the process of taking applications that were locally or privately hosted and moving them to run in or utilise public or private cloud datacentres.
Cloudification has become a very common process as many companies want to downsize their on-site datacentre capabilities and take advantage of the computational power of large, centralised datacentres.
These datacentres and networks also often come with large inter-regional internet backbones that can allow large organisations to easily send and share data across their organisation around the world.
Why does cloudification matter to distributed networks?
Cloudification is the adaption of applications and process to cloud computing.
This means taking applications and moving them to run in or utilise cloud infrastructure.
Cloudification has become a very common process. Many companies want to downsize their on-site datacentre capabilities. This allows them to take advantage of the computational power of large datacentres.
These datacentres and networks also often come with large inter-regional internet backbones. This allows large organisations to send and share data around the world more easily.
What are the most common network problems service providers face?
Common Network Problems for Service Providers
- Distribution and Network Quality
- Networks have become more distributed in the past 2-3 years.
- This has led to a huge variance in network quality across home users and office networks.
- It is hard to control network quality for remote workers. There are tools available like SD-WAN and business broadband, but these can be expensive.
- Even if you do plan your network for distribution, there will always be edge cases. There is no longer ‘one use case’ that can define your network structure.
- Routing and Remote Access
- Companies’ application portfolios have become more distributed into the cloud. This means considering how to route traffic to handle this new diverse ecosystem.
- It is not enough to have all traffic routed through the company VPN.
- Companies that do this can often find their VPN services overloaded by non-priority traffic.
- In response, many companies are taking a more critical approach to handling VPN traffic.
- A solution many are turning to is internet breakout. This is direct routing of traffic onto the internet. It means non-critical applications do not need to use the VPN.
- Hosting and Location – Route to Cloud
- ‘Route to Cloud’ has become an important term for service providers to consider in the past few years.
- This is the route that an employee takes to reach the cloud services they use. It is no longer enough to connect someone and provide secure access. Providers must think about what cloud services a user is accessing and how they connect to them.
- This means evaluating all applications to ensuring there is a smooth access plan for each and building policies that ensure that the routing is not the cause of QoS issues.
- Resource Competition
- The number of cloud applications used by each organisation continues to rise. This also means more and more bandwidth is being used to fuel these services.
- Often non-critical applications are competing with critical applications for bandwidth. e.g., A cloud storage update happening during an important video call.
- This competition will continue to escalate as more applications move to the cloud.
- Device Related Issues
- The wide variety of devices of all shapes and sizes at companies is a challenge for many providers. BYOD was growing before, but remote working has accelerated this trend.
- While this isn’t a network-specific problem, it does impact connectivity providers.
- Sometimes the device itself is at fault for the quality of the connection. But without uniform devices or device policies across a company it can be hard to diagnose the issue.
- This can lead to the network taking the blame for device issues.
Often ‘the network’ isn’t the problem
When looking at the issues above, we are often talking about problems that aren’t caused by the network. This is a recurring issue experienced by networking service providers. The growth in applications, remote working, and BYOD has created many problems for networks.
It has also led to an explosion in data generated that can monitor and report on network quality. The networking industry is responding to its challenges by generating more data. This is helping it diagnose specific issues. Customers can then hear the specific problem and ensure it is resolved.
It’s not only network players that are doing this. Several application vendors are looking to build network monitoring tools into their platforms. Soon, some popular applications will be giving network quality feedback.
Moving beyond that, there are some common solutions to the above problems that are worth discussing.
What solutions are networking service providers building to improve network issues?
Solutions that networking service providers are building to improve network issues
- Internet Breakout
- Internet breakout is the routing of traffic to the internet instead of a company network.
- Companies achieve this by placing an internet access point close to a user. Traffic is then offloaded to this access point rather than utilising the company network.
- For an office that is part of a WAN this will involve special routing policies. It might involve an SD-WAN device that is capable of sorting traffic by application type.
- For remote workers, this will likely be a policy about when and how to use your VPN. The company will want to ensure that personal traffic, for instance Netflix and Video Calls, does not use the VPN.
- This approach is being enabled by newer routing approaches like SD-WAN. These solutions can differentiate traffic from applications. This allows for varied treatment of applications based on company policy and need.
- Critical traffic that needs the highest quality can use the expensive private network. Less important or network-sensitive traffic can use the internet. This smooths operation across different applications that run concurrently.
- This practice is getting more and more common. Many of the larger providers of applications such as Microsoft 365 are now building internet gateways. These provide a direct path for connections from the internet into Microsoft 365. Microsoft is even recommending that some companies use the internet as the default for Microsoft 365.
- Better Cloud Routing
- Following on from the idea of better routing for systems like Microsoft 365, it’s important to understand the route from the end user to the cloud service they are trying to access.
- In the previous point we talked about Microsoft recommending the internet. It can do this because there are many datacentres hosting the service, and many access points to these datacentres.
- This makes it likely that the closest access point to the 365 cloud is closer to the end user than the corporate network. If the traffic went to the corporate network first, it would be doing an inefficient route.
- Due to this trend, larger network providers and vendors are building agreements for co-location and direct routing to common cloud solutions. This guarantees that they can connect traffic to the closest datacentre to the user. This will reduce overall latency and improve end-user experience.
- Intelligent Routing
- Both points above are examples of intelligent routing. This is the use of data to determine the best route for traffic.
- This is going to become a more and more common practice as networks are generating more and more data. Networking vendors are taking this data and using Machine Learning tools to optimise their products. This data is also enabling real-time network monitoring and error-correction.
- An intelligent network that has many methods of handling traffic is more resilient and efficient. One that can do this automatically leads to drastic improvements in network quality.
- SASE for Security
Will networks have problems in the future?
All the solutions discussed above share a trait: they imagine a future where networks fix themselves. In this future, direct network intervention becomes rarer; your network knows what is wrong and suggests solutions to you.
At the same time, as we are discussing this we have the continued rollout of higher bandwidth connections across the globe. Many countries in Europe have high-speed broadband as a given. 5G is also becoming more widespread, and backbone lines are being expanded by hyperscalers every year.
In a world where networks are smarter, and bandwidth is widely available, there is one big question about whether networks will struggle at all?
A home with a 1Gbps connection in the Netherlands will not struggle, even if many people are working from it. If that is the quality of a home connection, then a business one should be able to handle magnitudes more.
If a vendor can deploy a solution that can use intelligent routing to deploy and optimise the network, and local bandwidth from any internet provider can handle the service, where is the role of the specialist service provider?
The answer, as always, is that this utopian image is far from being a reality. There will always be local issues that need local companies and local support.
However, service providers cannot rest easy as the ‘local choice’, even with Cavell’s market research showing a preference to buying local amongst telecoms buyers. Since the local provision of physical connectivity is not as valuable a commodity as it once was, service providers need to be making sure that they are harnessing the trends above to provide an optimised network experience. Many will also find success in offering cloud services themselves.