aaabbc - Fotolia
With the IT industry still fizzing with the possibilities of big data, it has become a battleground for suppliers of different components in the enterprise architecture.
While data warehouse suppliers say they provide the path to better analytics, application vendors are claiming similar benefits. IT departments should be right to pay attention, but they need to invest with their eyes open.
Infor, the world’s third largest enterprise application supplier, is now promoting the benefits of analytics in its applications.
“The idea is to take data, no matter what the source, and aggregate that in one place and build the analytics on top of it,” said Infor CEO Charles Phillips.
Sky Vault supports contextual applications that sit inside enterprise resource planning (ERP) tools and allow users to scrutinise data in any application or database on the network, perform analytics on the data and draw down the results using Json, the lightweight data-interchange format.
For example, users of Infor’s LN ERP systems could draw on supplier shipping and logistics data from GT Nexus, a cloud supply chain management network owned by Infor, and use it to inform decisions about inventory management in the ERP system.
Infor’s analytics tool Infor BI Analyzer sits behind the supplier’s Ion integration layer, which can draw data from multiple sources, including Infor rivals SAP and Oracle, the world’s two leading enterprise application companies. It uses the Amazon Redshift database, which Infor says can rapidly scale.
Read more about ERP data analytics
- How democratised BI tools in ERP application suites promise to finally make ERP analytics a reality.
- Centralised business apps, such as ERP and CRM, can frustrate, but also make a business lean. Data analytics has a fresh Six Sigma role to play.
- Manufacturers are realising that long-term sustainability requires the ability to exploit machine-learning algorithms by employing predictive analytics.
Sky Vault also includes Hadoop for machine data, while Infor BI Analyzer uses an in-memory database for speed.
“We can extract data from anything. Ion has adaptors to SAP and to Oracle, so we can develop feeds from any apps in Sky Vault. We can take that data and, once it is in Sky Vault, render that in real time as a context app, making it relevant to whatever the user is working on,” said Phillips.
Gartner research director Nigel Montgomery agreed it was a sensible idea because many larger ERP users will run multiple systems from multiple suppliers. There could be advantages to allowing analysis between ERP systems, providing the architecture is not too rigid, he said.
“Analytics should be embedded, not bolted on the side, and should use data from throughout the whole business. Infor is doing the right things,” he added.
“The danger is you create another monolith. That is the problem with all of these systems. If you look at SAP and [in-memory database] Hana, it’s going to give you speed, but it is still a big block of architecture.
“Unfortunately, all of the ERP players seem to believe that they have to own everything for it to be successful. In fact they could partner very effectively,” said Montgomery.
He also pointed out that if businesses are attempting to analyse live data across a range of ERP systems, it will only operate as fast as the slowest system. As a result, businesses may struggle to take advantage of the very fastest in-memory databases.
MDM a stubborn problem
Meanwhile, moving analytics into ERP environments does not remove the challenge in creating consistently defined data across applications, known as master data management (MDM). MDM projects can be notoriously difficult. They require agreement from multiple stakeholders and are not glamorous to lead. Nonetheless, they are vital for enterprise-wide analytics projects to succeed.
“If you take a multi-system approach, you now get into MDM and a whole bunch of issues, not related to Infor, which might make it difficult to succeed. Not least of which, not all ERP systems are open enough to pull data from,” said Montgomery.
Senior IT professionals could be forgiven for a certain sense of déjà vu. MDM has been a perennial challenge in creating and exploiting an enterprise data warehouse (EDW), which collects and stores data from across the enterprise for analytics and reporting, in an initiative separate from live applications.
A raft of internet technologies that allow data sharing across applications has made it easier to create an enterprise-wide view of performance without building an enterprise data warehouse, he said.
“A ‘post-modern’ approach to ERP takes account of multiple systems. It is not an issue, so long as it is a strategic decision. The problem arises when businesses just take best of breed and bolt it on,” he added.
But the end of the enterprise data warehouse is not near. Notwithstanding the considerable investment already sunk into these systems, real-time processing on live data in not always necessary, and is expensive to do.
“Most companies work in batch mode,” said Montgomery. “Real time costs money, and you don’t need it for everything. You may need to model processes from historic data, and having it there in real time is no advantage. It is a case of horses for courses.”
There may be advantages to running real-time analytics in ERP systems, drawing data from across and outside the enterprise. But IT departments are likely to continue to invest in parallel projects in the EDW for the foreseeable future.