Stone Soup
Overzealous Salesforce.com and Hadoop fans are selling stone soup; don’t buy it.
The head of enterprise architecture at a Fortune 500 company told me a story today about Salesforce.com.
A stranger comes to town with an empty cooking pot. He starts a fire, fills the pot with water, drops in a single large stone, and puts the pot on the fire. A villager asks what he’s doing. “I’m making soup with my magic stone,” he says. “I’ll share some if you help improve its flavor.” The villager agrees and tells his friends. The villagers bring vegetables, meat, and spices, and add it to the soup.
The stranger removes the magic stone. He shares the delicious soup. The villagers are amazed; the “stone soup” is the best they’ve ever had!
The moral of the story? Fool suckers into doing your work for you, and take credit.
Our enterprise architect, whose name is withheld to protect his identity, shares a truth: over-centralization is oversold. Centralizing all a companies data increases complexity, because non-value add work is created to connect systems to the central source. And, once centralized, tool options are reduced; the least common denominator reigns. At his company, execs have drunk the Salesforce Koolaid that the delicious soup of Einstein AI and Tableau are one-size-fits-all. But our hero is no fool. “My team will have to bend like a noodle to get all our data in Salesforce. It will take years,” he said. He and his team were the villagers, left to do all the work.
Paul Moxon compared centralized data to the Stockholm Telephone Tower. This tower, shown below, was built in 1887 to connect telephone lines in Stockholm, Sweden. That central tower had 5,500 connections. What a mess!
Don’t build a Stockholm telephone tower out of your data.
There’s a better way: data virtualization. Virtualization provides agile, secure, scalable access to data without moving it. Without ETL. Without paying Salesforce millions of dollars. Without tons of extra work for little value. It uses in-memory techniques to abstract, federate and cache data sources to deliver approximately the same performance as the source. Any tool designed for a database can use this virtual fabric: PowerBI,Tableau, Spotfire, data science tools.
Less work, less money, your choices.
Firms use data virtualization to quickly create fast, secure, reliable access to data. KBTG Bank in Thailand uses data virtualization to expose data services in days instead of six months. When you create hundreds of services a year, virtualization provides competitive advantage from agility. Data engineers spend time adding value, not fetching ingredients for stone soup.
So don’t fall for the stone soup trap. Virtualize, don’t centralize.
