SAP HANA powers Walmart’s Data Cafe
Authored by: Michael Diehr
Each week, more than 245 million customers visit Walmart’s nearly 11,000 stores under 71 banners in 27 countries and e-commerce websites in 11 countries. As the world’s largest retailer, Walmart captures an immense amount of data about its business.
We realized that there was tremendous potential value in our data and wanted to unlock data-driven innovation throughout the company. However, our existing data warehouse was limited by its legacy infrastructure, reducing its value to our people. A few of the crucial limitations were a lack of speed, flexibility and timeliness.
Our response was to design a thoroughly modern data warehouse which would provide near real-time data, unlimited granularity, at least three years of company data, and the ability to support thousands of concurrent users. Speed was a special concern. We decided that no query, no matter how complex, should only about eight seconds to complete. We called it the Data Café.
To accomplish this, we realized the SAP HANA platform was great fit for us. Not only can it handle our 200 billion rows of data, it can also insert our hourly updates to the system at up to three million rows per minute. (With no cubes or data structures to worry about, the data is immediately available for analysis.) Perhaps most importantly, it allows users to build any query and get an immediate response. In fact, 94% of queries complete in less than two seconds.
With the Data Café, Walmart is now putting information at the fingertips of thousands of employees, unlocking a world of data-driven innovation across the company. For example, with near real-time data and self-service queries, our merchandising teams can spot trends or opportunities and take advantage of them within a few hours.
This is fantastic.
Amazing capabilities with so much data.
Agree Russel. You can find more details on the HANA Innovation Award entry HANA powers Walmart's Data Cafe : View Idea
Voting starts on Apr 7th to choose the finalist 🙂
Great Use Case and it shows how the largest customers can leverage our HANA platorm for truly innovative capabilities. 200 billion rows of data....AMAZING
Thanks for sharing Rukhshaan!!! I also appreciate Michael to highlight this business uses in HANA In-Memory Database. This is another business case where database is no more a bottleneck, when thinking the superior reporting and decision making, business forecast/prediction met on time and finally self-service access to all information for high volume data. I would say this is today's real business case but guideline for tomorrow's digital transformation scenarios (where still is the question "how big would be the Big-Data for your self service BI"?). I believe this is very helpful blog for HANA use cases for modern EDW environment.
Hi Ravikar
Thanks for the feedback. Have you looked at the entries in this year's SAP HANA Innovation Award - there are some great examples of how HANA is being used for delivering self service BI
SAP HANA Innovation Award 2015: Ideas
I would recommend starting with the entries in the Technology Trailblazer categories as they are focused on Big Data but also browsing the Analytics Wizard ones for more self-service BI
Ruks
Hi,
As you mentioned above that IT team no longer create cubes, aggregates, and data-joins. What do you exactly mean? Was not the data stored and loaded in HANA? Was it virtually accessing ERP? Was this a "BW on HANA" scenario or independent HANA system?
Thanks
Here is the answer that Michael provided on the award entry page - reposting it here:
" Previously the IT team had to build cubes, aggregate the data and put it in a format that would allow the queries to work. This would take time and delay the 'real-time' need for analysis on POS sales. With HANA, the data is loaded into HANA without the need for any aggregation, cubing or joining data. The data is loaded into HANA in it's raw format."
Hope this helps
Hi Rukhshaan, Thanks for the post, I've also looked at the innovation nomination.
Some of the solution detail is a little "light" - I assume this is NOT a BW on HANA solution?
Was HANA Live used for modelling rapid deployment? Perhaps custom HANA modelling?
Which client tools enabled end-user self-service - Excel, Analysis Office, Lumira, WEBI?
Would be great to know a little more.
I'm curious what you use to get data into HANA.
SLT, Data Services, HANA smart data integration/access/streaming, something else?
Kind regards,
Dan