Buy vs. Build – The Network Edition

8 October 2018

As the juggernaut of Cloud continues to plough mercilessly forward and as the number of tools at the disposal of developers continues to grow, the conundrum of ‘Buy vs. Build’ in lieu of a Digital requirement appears to be leaning ever in the favour of ‘Build’. And is this really such a bad thing?

Well, without going into the murky, calculated depths of when it is better to do one or the other, the net result of your decision should come down to Total Cost of Ownership (TCO). Of the many benefits that are difficult to monetise, having more control of the asset and being able to iterate upon it can often be the deal breaker. If you have followed your Agile principles closely then you should have a solid platform on which to continue to improve, safe in the knowledge that unless your staff retention policy is truly shocking, even a changing of the guard will not have affected your decision.

But would I advise of any scenario, ever, in which building a Network and Connectivity Platform yourself is the right answer?


Whilst there are several network items available in the Cloud (routers, firewalls, load balancers, proxies etc), they are part of a greater ecosystem that has become the dark art of the Technology Sector. These items require an initial 3 phase knowledge to develop and implement a true architecture.

1 – How do I configure the network item itself?

2 – How do I configure the network item inside the Cloud environment?

3 – How do I configure the Cloud environment?

If you can master these 3 items, you are still only half way there. We haven’t even started talking about what is outside of the Cloud environment yet. And for those who think that everything outside of Cloud is ‘just the Internet’ I am afraid you are very wrong.

Finding people who can deliver upon all of these specialist areas is incredibly rare. Network skills are an ever-dwindling pool, a phenomenon caused by a combination of offshoring, a lack of investment by telcos in their staff and a belief that the Internet is the answer to everything (you should by now be seeing a theme in my thoughts on the Internet). So here is my first reason for why ‘Build’ is a bad idea.

The guys you need are unicorns. I know of two of them. Just two. Your decision to ‘Build’ needs these unicorns. But you can’t just build it, you need to keep building it. You need to change it. To improve it. To iterate. If you are lucky enough to snare a unicorn, (and buyers beware of the pantomime horse with a horn on its head!) it is highly unlikely you will be able to keep them as they will be in such high demand. This puts a huge risk on your decision.

Then there is physical tin and wires. It is an oft-joked about scenario amongst technology people that those less technically savvy think that Cloud is just ‘in the sky’. Unfortunately, it is my experience that some of the more ‘savvy’ may not be grasping the punchline when it comes to Networks.

There are many things that you can do with the aforementioned network items inside of the Cloud environment. However in order to build a platform that is appropriate for almost any business, there will be considerations that require physical connectivity. This is likely to mean that you need to get involved with MPLS providers, Cloud Fabric Providers, Managed Service Providers and Hosting Providers. The list does go on. If you go down this path you are going to start to need to look after physical equipment and that of course is moving away from the position you most likely wish to be in, just managing your Cloud environment from your desk.

You don’t necessarily need to tread the path of physical connectivity, and yes, through gritted teeth I concede that you can leverage the Internet for a lot of things. If, however you want an architecture that meets the key principles of being the lowest TCO (the reason you decided to ‘Build’ and not ‘Buy’) then it is highly unlikely you can avoid the physical elements. So to my second reason why ‘Build’ is a bad idea.

The TCO should be measured against numerous points. Inside of your Cloud environments, anything that you cannot do from your desk should carry an immediate flag in the ‘Build’ column. If working with physical elements is the right thing to do, and I stress again that it rarely isn’t, then your decision should definitely be to ‘Buy’.

Networks are not sexy. They never have been. They never will be. They are however absolutely fundamental to everything we do in Technology. If they break, then all of the great things built on top of them, relying on them, will come tumbling down like a house of cards.

Looking to take greater control of your Digital Destiny should not be frowned upon. Every requirement should be meticulously scored against the TCO and the outcome should be the appropriate business decision. I strongly believe this will continue to lean towards ‘Build’ and that this is a positive thing but I foresee no horizon where this includes anything to do with networks.

Now I must tend to those two unicorns.


Click here to find out more about Cloud Gateway and how it can help you and your business

Justin Day