Skip to main content
As someone who is constantly keeping up with what’s happening in the telecoms and IT technology market, it seems that every second sentence I read has the word cloud in it. And with recent announcements by companies such as AT&T that they are moving their 5G core to the cloud, the cloud is now telecoms!
In fact, the cloud has been telecoms for some time. Virtualised functions were available for years, whilst cloud on-ramps have been connectivity destinations for as long as the clouds existed.
The Flexera State of the Cloud Report 2021 has stated that 92% of enterprises have already adopted a multi-cloud strategy. My question is, how are IT and network folks actually architecting their telecoms network when it comes to enabling business functions to be accessed and to inter-work these days?
In particular, when there are so many options available? And with technology changes afoot, SD-WAN, cloud providers options and so on, how do you architect your business to operate efficiently and effectively?
Start with the applications first
To me, when you start to design or re-design your network to be a business operations enabler, you don’t start with ‘the cloud’ or the data centre, not even the office. You start with the applications and then the consumers. And by consumers, I don’t just mean people. I mean other apps and systems that depend on data from other apps and systems to perform their function.
This approach isn’t really anything knew, but it feels like it’s current. As one industry colleague I was talking to recently told me that he had been speaking to his CEO about networking or IT and he immediately asked for a full systems audit – wanting to know where his systems were, what their functions were and how they interacted. Nothing out of the ordinary really, but for some larger, more mature companies, a herculean task to start with.
But is it worth it to do so?
Companies who have been around for some time have probably experienced “network creep”. Cloud creep and even systems creep have a tendency to cause the need for more network in some way shape or form.
Whenever a systems audit is performed, you are not taking full advantage unless you use your audit to perform a network assessment to realign a fit-for-purpose architecture. Or as I call it, an Application Orientated Architecture.
What makes an Application Orientated Architecture?
Networking is one of four cornerstones of business operations. The other three being the housing environment (whether tin or walls), the code & operating system function and people. Any one of these four cornerstones can impact the performance of business operations.
Often, networking can be forgotten as less important, or “we have the internet and that’s enough”. Especially when deployed by 3rd party telco service providers, who focus on the network they want you to have rather than what you need.
But once you have your systems and/or application audit performed, there is a simple next step – define the interfaces between each of the systems and applications and categorise them.
Are they mission critical? Are they core and are they non-core interfaces?
The simplest of architectural diagrams can be created. Map that against your current network and then start looking into the network tools that are now available as substitutes for your existing services – you might be surprised!
Several years ago at Epsilon, we went through this same process ourselves. The most painful part was clearly identifying that our mission critical CRM system was on-prem and only public internet facing in Singapore! Building a mission critical core network that included our CRM changed the user experience. It also increased the inter-application performance within our business system ecosystem.
Another interesting result as an example: Designing a multi-service yet integrated network becomes possible when you choose and use the right technology for the interfaces required.
Removing a full MPLS mesh and replacing it with a combination of SD-WAN and ‘as a service’ networking. In this case, point-to-point ethernet services become maintainable by users, with both services having visible controllers you can work on yourself. And you still get the support of MPLS as most carriers are still running an MPLS underlay for their ethernet services.
Now don’t get me wrong, the physical aspect of networking will never go away. Designing your application orientated architecture means you need to map out where the systems and applications reside. This is your physical point-to-point architecture. You also have interface classifications to determine the type of network you need in place between housing environments, as well as any redundancies and disaster recovery plans.
Once you have these designs and have included your user access designs, you can finally look at overlay technologies that can run across all the transport technologies to give you an end-to-end network controller, including cloud networking capability. At the same time, layering in security, integrated public and private networking and end-to-end performance control all from within the cloud’s VPC/VNET environment.
The proof is in the ROI
These types of new overlay controllers have recently been reviewed by Forrester in a Total Economic Impact Study on the cost savings, business benefits and overall ROI of deploying this type of new network solution. The results were overwhelmingly positive – 222% ROI!
From my point of view, the simplicity of taking time out to review your business operations ecosystem, gives you the power to work on realigning and improving your business. The benefits are tangible, from performance improvement to cost saving as well as tighter security. Oh yes, security! That’s a story for another day.
Mark Daley, Director, Digital Strategy and Business Development, Epsilon Telecommunications
Close Menu