In-Memory Analytics: A Faster Computing Method

“In-memory computing will have a long term, disruptive impact by radically changing users’ expectations, application design principles, products’ architecture and vendors’ strategy”

Donald Feinberg, Gartner VP and Distinguished Analyst

“In-memory is just a hype spread out by SAP”

Introduction:

Be it the support for the increase in bets from 12000 per second to 15000 per second for the online gaming company Bwin.party, the near-real-time insight into product availability for customers by the retailer Edgenet or be it the response by ConAgra, an $18 billion-a-year consumer package good company to the fluctuating costs of 40,000 raw material, the in-memory analytics made it possible.  Before going to the business aspect of in-memory analytics let’s have a look into what is in-memory analytics and how it is helping business to grow faster?

In-memory analytics facilitates querying of data from Random Access Memory (RAM) instead of physical disk. Data can be loaded from multiple source into the system memory directly. This helps in faster processing of data and faster business decision. Following figure depicts the general difference between the Traditional Computing and In-Memory computing.

Types:

There are 5 types of In-Memory analytics and they are:

Major Industry Drivers:

The figure bellow shows powerful drivers push organizations towards In-memory analytics.

The first and foremost driver for In-memory analytics is Big Data. To handle wide volume, velocity and variety of data faster computing technology is required and In-Memory is the need of the time. Real-time analytics is the second and an important driver. To make faster, accurate business decisions, organizations are relying on the real-time analytics and In-memory analytics makes it easy. Another driver for this new technology is the need of scalability and flexibility that required for the business. In the other hand some factors are there standing as hindrance for this technology to adopt. Among them lack of knowledge and development, security challenges are main hinders.

Value Creation:

The four dimensions of the value created by In-Memory Analytics are:

1. Performance: The time taken for data analysis comes down drastically from hours to seconds. In-Memory Analysis allowed levering the most recent data for the purpose of analysis and informed decisions.

2. Process innovation: The performance gain gives space for innovation in various application and it leads to competitive advantage for the organization.

3. Simplification: Due to reduction in layer the complexity of data models can be reduced significantly. This simpler architecture reduces the sources of potential error.

4. Flexibility: During analysis new data source can easily be plugged in as an additional source of information and this provide flexibility to the data analysis.

Adaptation Rate:

As per a survey done by Deloitte (German CIOs) 52% of respondents said that they are still evaluating the IMA technology, where as 22% CIOs said that they are on the plot and only 4% said they have adopted this technology. The bellow graph shows the break down:

Conclusion:

IMA is the future of computing but requires a clear strategy for all steps from evaluation to implementation. This includes, but is not limited to, the identification and evaluation of opportunities for the utilization of IMA, business case development, management of the implementation as well as learning and change management for pilot scheme and roll out.    

References:

·         http://www.webopedia.com/TERM/I/in_memory_analytics.html

·         http://searchbusinessanalytics.techtarget.com/definition/in-memory-analytics

·         http://timoelliott.com/blog/2013/04/why-in-memory-computing-is-cheaper-and-changes-everything.html

·         https://www.yellowfinbi.com/Document.i4?DocumentId=104879

·         http://www.csi-india.org/c/document_library/get_file?uuid=2b5d426d-f36c-4591-a7a3-22159cb3ccb4

·         http://www2.deloitte.com/content/dam/Deloitte/de/Documents/technology-media-telecommunications/TMT_Studie_In_Memory_Computing.pdf