The Connected Enterprise

by CAKE LABS Articles 19 December 2016

In this Innovation Session, Chaminda Bandara from the Menu Core team spoke about the Connected Enterprise.

The Connected Enterprise is a concept that takes several different systems and shares the data amongst it. Google is a perfect example of getting this concept right.

1995 saw the battle of the search engines. Several players competed for the right to be king, yet over 20 years later, it’s clear that Google has emerged as the sole victor, crushing all other competitors. Let’s take a look at a particularly interesting case study.


AltaVista vs Google

AltaVista was a search engine back in 1995. It quickly became a popular brand, but lost ground to Google over the next few years. It eventually was bought over by Yahoo! in 2003, but 2013 saw the product finally die, as its service was shut down by Yahoo!

One of the factors that may have played a part in AltaVista’s decline, was the way they handled data. They decided to handle data, such as search information in an isolated manner. Google went down a different path where it identified relationships between search results, resulting in where it is today. It has evolved into a search engine that is able to predict and suggest information to their users in a convenient manner. What Google did is to make sure that the core of their business remained connected.

The above diagram shows an example of an enterprise system. As the diagram illustrates, all components are connected, and what this ultimately means is, the relationship of one component is shared by the other. So if we take a piece of data from the Sales Channel, we could plot what its relationship should be with Marketing and the Supply Chain. A system like this makes it very easy to take critical business decisions, and even make predictions based on the information available.


Graph Databases

In most cases, where a system already exists, connecting all the databases is no easy task. One cannot simply throw away the existing databases and start anew as all existing users will be lost too, if this were to happen. Graph Databases offer a simple solution to this. The multiple system can be kept as is, but only the connected data between the systems are extracted and published into the graph database.

Walmart is a good use case for the successful use of a graph database. In the Walmart website, if you go to buy Product A, Walmart searches their database for similar buys in relation to the product you bought, and will suggest Product B and C as well.

Another important factor when going for a graph database, is picking which algorithm you are going to use. There are several options available to you, but let’s take a look at a few key ones.

The Page Rank algorithm is one that was perfected by Google. The page rank algorithm assigns a value to any point in the graph, so if anyone were to search for something on Google, google goes through past records (which have already been assigned values), and ranks them via their weight. The highest weighted result will be what you see as the first result on the first page, and the rest, all sorted via their weight, follows.

Next we have the Centrality algorithm. This works by the algorithm calculating the shortest path between 2 separate nodes, so if you take 2 nodes, the algorithm works out the minimum spanning fee. The algorithm also uses a weighted system, by adding weight to the nodes depending on their traction.

If you take the diagram below, the nodes marked in red have the highest centrality value, and the ones in green has the lowest.

This algorithm is very useful in the micro biology industry. Say you want to find the common factors in, for example a virus, you can put the different factors into a graph, and the algorithm traverses through the nodes. This will ultimately give out the common factors in a virus.


CAKE as a Connected Enterprise

For us at CAKE, moving into a Connected Enterprise will be a big challenge, but might ultimately reap huge benefits. Currently our existing CAKE Platform has several different applications, that are hosted in different severs, and maybe even databases.

What this all means, is that if we want to figure out relationships between the different sets of data, it will prove to be a very tedious task. However, should we move into the world of the Connected Enterprise, it might open up several different business opportunities. Having a comprehensive, connected database, will help in all decision making, as with this kind of data, market prediction becomes far simpler. Helping consumers decide what kind of food they might want to try by analyzing trends, the consumer’s previous sales information, and even of the food item they are buying, would be a path that could open up.

Using the connected enterprise methodology could also be used to help CAKE merchants by improving business decisions through the analyzing of trends in a certain geographic region and providing comprehensive Business Intelligence reports.


The world of the Connected Enterprise holds a lot of promise and challenges, but it could change the way businesses are run, if done right.




Leave a Comment