The future of collaboration is cloudy
From the archive: written before or outside the Architect Tomorrow newsletter, so it may be out of date.
Sorry for the terrible pun! But I promise it is relevant…
Cloud offers unique web scale capabilities that are out of reach for most organisations to achieve on premise. Why unachievable? Well its not just cost - its the fundamental cloud architecture that allows for new workloads and computing intensive capabilities.
Examples of this are search and machine learning. You are probably aware that the only way search engines like Google or Bing can offer the lighting fast experience is through a distributed architecture. This relies on having hundreds of thousands of server instances. Most organisations can’t afford to have thousands of servers for a capability that doesn’t differentiate themselves in the market (if its your core business system then there may be more of a business case).
Machine learning is really interesting - take for example Delve which is part of Microsoft Office 365:
“Delve proactively highlights key information that’s relevant to you based on what you are working on and colleagues you work with closely”.
Again you can only achieve this with web scale architecture which is only cost effective when shared. You could think of it a bit like a football fan - you buy a match ticket for a single seat in the stadium - you’d never be able to buy the entire stadium just so you could watch your favourite club by yourself. Maybe that is a rubbish analogy - but then I am not a football fan!
The other angle to consider on this is connectivity, mobility and security. In order to securely expose internally hosted collaboration platforms to mobile devices a lot of effort (and on going maintenance) is needed. Firewalls have to be configured, servers have to be patched and updated, regular security penetration testing is needed (and fixes put in) and you need to support all those servers. When the systems were within the corporate boundary / firewall businesses could choose to run the risk of not keeping up to date; but that is no longer a realistic approach to cut operating costs*.
So why go to all that effort and cost for a capability that is commodity and largely off the shelf? Particularly when that capability continues to evolve (and updates come to the cloud faster and on a more frequent basis), devices continue to change and business models continue to be disrupted.
I must confess when I first heard about cloud a few years ago I thought it was vendor hype and just a way to lock in less spiky revenues from services. Now that the killer applications are coming and the business case is becoming more compelling; the conversation on internal systems seems to be naturally moving from:
“we can’t use cloud” to “why wouldn’t we host this in the cloud?”
Therefore don’t be surprised when cloud platforms completely morph out of their original on premise origins. Ultimately offering their users greater productivity than those stuck using ageing on premise platforms.
* Generally it can be a false economy to completely ignore upgrades - leave it too long to keep current and the costs become significantly more than the smaller upgrades would have been.
Originally published on LinkedIn. Comments and discussion live there.