This is a guest post for the Computer Weekly Developer Network blog by Matt Smith.
Smith is chief technical officer for UK, Germany, Nordics and South Africa region at Software AG — the company sells big data, integration and business process technologies designed to drive operational efficiency, modernise IT systems and optimise processes.
1 banana, 2 banana
To start with, please think of bananas.
Yes, the curved yellow fruit beloved of athletes and regularly bought by millions of shoppers.
It is often said that if a supermarket doesn’t sell any bananas within a given ten minutes on any chosen day, then something is seriously wrong with the balance of life.
But how can a store manager stacking beans seven aisles away know whether his or her banana-sales are drooping?
Further, how can that same manager do something about it immediately?
Fruity algorithms, oh-err missus
The answer, as I’m sure he or she would tell you, lies in event processing and advanced algorithms.
(CWDN Ed — Do store managers really a perception of advanced algorithms, or are they more concerned about Mrs Miggins in aisle 16 asking if the tuna is dolphin-friendly? Oh okay we get it, Software AG is painting a picture of the connected future.)
This should not sound like a sexy sales gimmick and it isn’t — it’s more scientific and involves analysing all the data that pours into and out of a supermarket throughout its working day.
By putting these technologies to use, it’s possible to match the current level of sales against the historical patterns, without anyone having to sit watching a screen — or a diminishing stack of bananas.
How banana logic works
As soon as the till receipts indicate an unexplained plunge in banana sales, the algorithm triggers a stock-take, which can be conducted using in-store sensors. If it turns out that stocks are low, an alert is automatically sent to the manager’s smart badge and he knows he or she has to drop the beans and crack on with the bananas, restocking those shelves.
This admittedly, is not the most sophisticated demonstration of how the smart data layer works, but like bananas, the technology is a potent combination that boosts energy.
Just as bananas are attractive to athletes because they have three different types of sugar, the smart data layer combines event processing, complex algorithms with time-based rules, and low-latency messaging.
Having the right algorithms in place means the data can be combed 50 or 60 times faster than in a conventional system.
Yet until very recently it was only the banks and trading institutions that could afford to crunch through mountains of many different kinds of data in real time and then use it to make split-second decisions. Of course they need this capability to trade in the markets.
It’s not just retail
Now, however, every sector has the chance to use the same technology, thanks to the falling cost of memory and the vastly increased power of parallel processing chips.
Telecoms, manufacturing or retail – they can all put this technology to work, storing data and using it in a completely transient way. A credit card company can employ it for quick-fire marketing campaigns, for example, and then if the data is no longer needed in that form, just chuck it in the bin.
In fact, the smart data layer is going to be indispensable in the drive towards omni-channel retail. It will automatically mash through data from multiple systems, providing the connections and fast decisions about inventory that guarantee full shopping baskets in-store or online.
Nobody needs to have a big shed full of expensive servers to do this work. It means that even medium-sized businesses can now enjoy the benefits of the smart data layer, conducting tasks that would formerly have been beyond their pockets.
So when someone senior tells you they are getting into the smart data layer, they are definitely not bananas. Their business is about to receive a big energy boost.