Driving Operational Analytics - An SAP Journey with MapR

Contributed by

10 min read

SAP HANA DB is part of the HANA platform and it also acts as the underlying database for all current and future SAP products. SAP continues to gain market share as their strategic shift to ‘cloud’ continues. While HANA DB is a leap forward as an in-memory, relational database, and serves the platform well, there are still newer-generation technologies that are more efficient when it comes to handling broader scale and performance requirements. SAP HANA DB prescribes strict measures for maximizing efficiencies and non-functional requirements such as HA, all of which have a direct impact on the cost of operating a HANA DB environment in the context of today’s data-rich enterprises.

New usage scenarios such as IoT, or increasing industry regulation that requires keeping more data for a fixed time, have meant a rapid increase in the amount of data that is being stored and this growth has a material impact on both direct and indirect costs (hardware, licenses, data center costs etc.). So, is there a way of leveraging the full value of SAP HANA and the S/4HANA suite while addressing data volume and variety in a lower-cost, new-generation way?

What follows is the description of three stages of a data journey that can take you from cost containment and retirement all the way to broader, enterprise-wide innovation much more quickly than you might imagine and take your business to new levels of competitiveness and new levels of relevance to your customers, suppliers and partners.

Three Stages of a Journey

Stage 1 – Driving cost efficiencies into the SAP HANA environment

  • Cold SAP HANA data are off-loaded to the MapR XD Distributed File and Object Store; a new-generation, scale-out-on-commodity-hardware, data platform
  • SAP HANA has access to all of the hot data it continues to store – and – the cold data that resides on MapR
  • Cold data can be loaded back into the SAP HANA persistence layer from MapR at any time – data movement is bi-directional and data availability is on-demand

Stage 2 – Leverage SAP HANA data with additional data sources for deeper insight

  • SAP data can be combined with additional structured, and semi-structured data for richer context and meaningful analytics
  • MapR’s converged data platform now plays the role of providing new data services for new solutions and applications that leverage the combined data
  • MapR’s converged data platform provides all the core functionality to drive new data services. In addition to the MapR XD Distributed File and Object Store, MapR Event Store provides Global Events Streaming for SOR-grade real-time data capture and analytics – and – MapR Database provides best-in-class operational data capabilities for unstructured data

Stage 3 – Put new-generation data capabilities at the service of SAP HANA

  • Any or all data can be landed first in MapR where it can be consolidated, enriched and processed
  • Only the prepared “hot” data are moved/copied over to SAP HANA, so that high performance SAP HANA workloads can be maximized
  • As a converged data platform, MapR fulfils the role of both data lake and data hub with some unique benefits that arise from those highly integrated competencies

Why take the Journey?

Timely off-load of HANA data into MapR XD leverages massively lower-cost data storage without sacrificing the non-functional capabilities that your business expects from critical enterprise applications such as SAP. With much lower cost per TB, much more data can be stored which means you can avoid data reduction processes in HANA and data scientists and analysts can leverage ALL the data, leading to more accurate predictions and analytics. Regulatory data can be stored in a cost-effective way but, importantly, be retrieved at any time for analysis.

With realistic ROI expectations of less than 12 months, cost savings can be realized quickly with capital released for other critical enterprise investments. With new data generated or being made available every day, hour, and increasingly, every minute and second, enterprises have the opportunity to use this data to drive exceptional customer experience, critical business decisions and new-generation digital products and services that can keep their organization and brand, relevant and competitive.

Why is MapR necessary for this journey?

To work harmoniously with SAP HANA as one of the world’s most trusted, enterprise IT systems, there needs to be parity on data security, high availability and other non-functional requirements demanded by an enterprise for their business-critical data. MapR’s enterprise capabilities guarantee business continuity and preserve access controls, for example, across the entire tiered data estate. Disaster recovery, data mirroring and high-speed snap-shots align with the same stringent SLAs expected of SAP. Multi-tenancy, auditing and data lineage enable organizations to apply and adhere to strict data governance guidelines.

With the major premise of cost reduction underpinning an off-load strategy, there would be little point in moving data to a file store whose clusters will sprawl and whose loose alignment with the other technologies required for complete new-generation data capabilities (NoSQL databases for high-scale structured and unstructured data and event streaming capabilities for example) will lead to increasing numbers of data silos. New-generation data silos are ultimately no more desirable than old-generation data silos. MapR has industry leading scalability, performance and reliability combined with much simpler cluster architecture; a massively more efficient foot-print made possible through our “no name-node” architecture.

A single cluster of MapR Data Platform can be used for multiple different workloads and can serve unlimited applications with all the data they need to consume at any required data velocity. MapR’s open standard POSIX interface enables direct integration with any enterprise application making the three stages of a journey outlined here, possible with other enterprise applications and systems so those systems, too, can finally expose their data value to the business in a much more potent way through MapR’s new generation data platform.

And finally, MapR is SAP proven. It underpins SAP’s Consumer Insight 365 platform as well as being the software defined storage solution for the HEC platform. SAP understands the MapR difference and that difference, and the way it complements SAP, can be leveraged by SAP customers too.

What does the destination look like?

A global chemicals company have multiple HANA clusters (one in each region) in which they store data from their production lines. They previously had no orchestration layer sitting above these clusters but now use MapR as the centralized data platform to combine both real-time and batch analytics to give a new level of visibility across the supply chain organization; the actual, current status of their global production.

In Germany, supermarkets have to store transaction data for seven years to prove, at short notice, that there was no collusion or price fixing for certain product categories. The authorities are able to request data at any time and the supermarket must return it within one to two days. One customer is keeping the most recent transaction data on HANA while off-loading older data to MapR. The SLAs required by the anti-trust regulators are met across the entire hot-to-cold data landscape.

A media company sells bundles of products as a subscription to corporate customers and ahead of a renewal discussion, the sales team wants analytics on how many times each of the bundled products have been accessed so they can determine where the value is for the customer and identify products that the customer has not been using. Clickstream data is being moved to MapR for cost efficiency and MapR will in future generate the reports for the sales team to strengthen their negotiating position and reduce customer attrition.

MapR XD is deployed as a multi-functional, scalable data store that stores all of the data, content, metadata and digital assets. All of this data can now be leveraged for analytics such as machine learning through exposure to the plethora of new analytics technologies, algorithms and libraries that are available today. Cost can be retired through the replacement of large, inefficient footprints of legacy storage and the built-in enterprise features of MapR further lower TCO while improving data governance and delivering a secure, multi-tenant environment that realizes minimal data movement and maximum insight.

With MapR, you can finally embark on a journey that delivers both cost efficiencies – and – new generation innovation. A journey that provides a real-time bridge between your trusted enterprise infrastructure and new generation capabilities. All while avoiding unnecessary complexity and the recreation of data silos and excessive data motion that have characterized the inefficiencies of the past.

This blog post was published September 25, 2017.

50,000+ of the smartest have already joined!

Stay ahead of the bleeding edge...get the best of Big Data in your inbox.

Get our latest posts in your inbox

Subscribe Now