When Robots Beam Down From the Cloud


cloud-high-low-ft

theaucitron/Flick



It’s been over 20 years since Larry Ellison made the case for cloud computing. Some speculate he only invented the idea to get at Bill Gates, the dominance of the PC and the then-newly released Windows 95. Even if the motivation had a personal edge, the point is that he had a strong case.


As always, the argument is about economics, and the issue is: how many people benefit under the new regime as compared to the old?


When Ellison first espoused the model, he likened it to things like power plants, reservoirs and recycling – cases where it makes sense to share and where we all consume the output. Aside from the out-of-character, dangerously socialist undertones, Ellison were simply saying cloud computing costs less.


The idea was quickly ridiculed. Journo comment went for the jugular, saying, “The day Larry Ellison composes sensitive Oracle memos on this new fangled device is the day I…”


Well. How far we have come since then. Not only do people compose sensitive memos on the “new fangled devices,” we use them to keep unreleased film scripts and actual finished films. Plus, it seems all celebrities keep their nude snaps on these fangled devices too. It has even invented a new form of blackmail where we, the public, get to see (or not see) these items unless demands are met. So, the sensitive data argument has been laid bare and beaten by the good ol’ cost reduction paddle.


Ellison always argued it was the network that should be regarded as the real value – not the device. And to this we now have much consensus. The bean counters like networks because they deliver everything as a service (EaaS). This means capex is abolished, everything is a running cost, and there are no more bulging balance sheets groaning under the weight of old iron and tin.


However, apart from the security question, there is an issue that still generates corporate jitters, and that is interoperability. It is the question of the basic plumbing: how do these things get connected together? Say I want two or three clouds? More? How do they connect with non-cloud things? In short, how do they interoperate? Especially in an emergency?


Most businesses want more value from assets, not less. For that reason, connections between on-premise and cloud, and between clouds, and even within a single cloud, is key. Otherwise we fear a padded cell that eventually rains value out of our businesses. Without certain freedoms it risks becoming a One Size Fits All as a Service (OSFAaaS) prison, in which we are forced to survive on a single service Pangea with no way off. Where is the balance?


Businesses want great value but they need choices, flexibility and some individuality. Like it or not, we also all compete at some level, even on how we pay sales bonuses, deliver parcels and make telephone calls. Put another way, we even configure SAP differently.


One of the most exciting things about this situation is the intervention of our now-familiar software robots. They are late to the party but will make a crucial point.


Software robots, or computers/virtual machines/processing capabilities, are repurposed to mimic humans and the way humans orchestrate and operate other software applications. These software robots become fantastically relevant in the cloud context because they offer choice.


They are totally promiscuous – like humans. They will use any service, like humans; can switch and distribute workload across services based on congestion and latency, unlike humans; automate like machines, unlike humans; are available from the cloud, unlike humans; and can scale, almost at the speed of light, very unlike humans.


Software robots offer a form of “operating system” for the cloud, providing freedoms from one provider to another, or between complex blends of configuration. One could even arbitrage service levels with the right software robots. This looks a lot less like prison.


Punters will truly welcome these software robots. They dispel the fear of being trapped and having to make economic choices that don’t seem to make total sense. Legacy on-premise systems can deliver value. Moreover, the cloud choice is not irrevocable. It means on-premises capacity is used, as needed, can be enhanced, mixed and even replaced with cloud services as required.


Software robots also accelerate the adoption of dedicated cloud switch over. They are a technology that is used by the operations teams, so they offer an immediate increase in the resources capable of carrying out these types of initiatives. It also helps that the business teams are capable of taking data sections (literally through GUI windows) to the new cloud service, without the extremity of uprooting and performing a complete ETL database transplant. This sheds an enormous amount of weight and risk.


Because robots mimic humans, they future-proof against new releases and upgrades at any link in the chain. This flexibility means cloud becomes less like prison and more like a service buffet. Who could dislike that? It means people can relax and enjoy the balance sheet improvement. Ironically, they may even accelerate the race towards a service Pangea.


Jason Kingdon is chairman of Blue Prism.



No comments:

Post a Comment