In-memory’s role in logistics
In the IT world that exists outside logistics, there are a few more buzzy concepts than in-memory computing (IMC).
The dual heavyweights of enterprise resource planning — SAP and Oracle — have both unveiled in-memory database technologies across their range of solutions in the past year. Expect more specialized providers to do likewise.
So what exactly is in-memory computing? And more relevant to this space, what might its impacts be for users of logistics and transportation systems?
First question first. In-memory basically allows data to reside in a computer’s physical memory rather than on a hard disk or external drive. The theory is that data is then more easily accessible to manipulate in various applications, sort of like keeping all your cooking implements on the countertop saves time over fishing them out of various kitchen drawers.
I attended SAP’s launch of its HANA in-memory database technology in New York in December 2012. It was a glitzy affair, with coordinated live telecasts in three locations, including SAP founder Hasso Plattner.
Plattner described the essence of in-memory as increasing speed and decreasing waiting time for processes to occur. He said the rise of mobile computing necessitated both changes, with users wanting business processes to occur with the same speed and clarity as those of mobile phone functions.
It wasn’t immediately clear how HANA might impact the supply chain modules that SAP offers, other than providing users with the potential for these faster processing speeds. At the time, I sort of envisioned the change a company might undergo in migrating from its existing database to HANA as one might see in tossing a laptop bought in 2002 and replacing it with one manufactured in 2013.
As it’s been later explained to me, the advantages of in-memory, from a supply chain perspective, come from the ability to better game plan. Faster processing using in-memory technology would allow, for example, a user of a transportation management system to run countless routing scenarios in quick time, and closer to the event upon which a transportation manager has to execute.
It’s also easy to envision freight brokers being able to use in-memory power to more quickly process and produce instant and accurate rates.
Theoretically, this puts more power in the hands of the transportation manager to decide which scenario is most appropriate. As Plattner said, users of enterprise systems have become conditioned to the rapidity of functionalities on their mobile devices. It’s a topic we’ve broached in this space before.
Oracle has its own in-house in-memory technology, called Coherence, as well as the TimesTen platform it acquired from Hewlett-Packard in 2005. A range of smaller companies have developed in-memory capabilities, presumably to underlie the databases of more niche providers that don’t have the investment wherewithal to compete with SAP and Oracle.
The next step is solutions developed specifically with in-memory in mind. Just as logistics solutions providers have been forced to adapt their offerings to cloud environments (at least those that weren’t initially founded on a cloud basis), they will have to adapt to environments where ERP systems and all subordinate modules will run much faster due to in-memory technology or, better yet, a combined migration to cloud and in-memory technologies.
The possibilities seem endless, but there are some potential pitfalls. It’s not hard to imagine the amount of data that faster processing generates could quickly get out of hand. The era of big data has arrived, and it’s only going to get bigger. Managing that data in an in-memory environment should be a priority.
“In-memory applications are best for processes that require speed of response or are processor- or database-intensive, so planning applications benefit the most,” said Will McNeill, a principal analyst with Gartner. “Running any kind of logistics plan will be about 10-times faster. Scenario planning is another.”
McNeill said he doesn’t expect many vendors to tailor any specific supply chain applications to in-memory computing capabilities in the near term, with the possible exception of Oracle and SAP. If there is near-term movement, he said, it will start in broader business intelligence scenarios.
But the gains to be made in planning seem tangible.
Toby Brzoznowski, executive vice president for the supply chain design solutions provider LLamasoft, wrote in an April commentary for Supply Chain Digest that cloud and in-memory computing have enabled a much richer degree of planning than was available before, by “removing the barriers that have traditionally limited supply chain design to a desktop application used by a single modeler.”
In-memory provides the robust engine to run so many complex scenarios.
On a wider basis, Gartner posited in April that in-memory is “racing towards mainstream adoption.”
“(In-memory computing) is an emerging paradigm enabling user organizations to develop applications that run advanced queries or perform complex transactions on very large datasets at least one order of magnitude faster — and in a more scalable way — than when using conventional architectures,” Gartner wrote. “It completes batch processes that would otherwise take hours in minutes or even seconds, enabling processes to be delivered to clients, suppliers, citizens or patients as real-time cloud services. IMC can also be used to detect correlations and patterns to help pinpoint emerging opportunities and threats as things happen, across millions of events, in the blink of an eye.”
Again, the immediate impact of in-memory computing on logistics won’t be so profound. For one thing, it will take time and desire for companies to migrate from existing database architectures to IMC ones.
But once they do, the management of big data (another buzzword!) becomes more manageable, the creation of planning scenarios becomes easier, and execution of those plans becomes more straightforward.
In time, the hard disk will become to logistics computing what the compact disc has become to music — largely obsolete.