CIOs’ ears perked up a couple of months ago when Gartner declared that in-memory computing (IMC) is racing towards mainstream adoption, and that CIOs must reskill their teams if they want to fully exploit it.
Shifting away from the norm of electromagnetic disks, IMC stores application data in a computer’s main DRAM without compromising its availability, consistency and integrity.
This allows organisations to develop applications that run advanced queries or perform complex transactions on very large datasets much faster than when using conventional architectures.
The result is an opportunity to change the way in which organisations support business requirements. Batch processes that normally take hours are completed in minutes or even seconds, enabling processes to be delivered to clients, suppliers, citizens or patients as real-time cloud services.
From a security perspective, IMC can also be used to quickly detect correlations and patterns to help pinpoint emerging opportunities and threats as things happen, across millions of events.
According to Gartner, a rapid penetration of IMC is likely to lead to it being adopted by at least 35 percent of midsize and large organisations by 2015. Furthermore, the in-memory data grid (IMDG) market — a key IMC segment — is forecasted by Gartner to reach $1 billion by 2016.
But what part does the Middle East play in this global forecast?
The US and EMEA regions probably account for 45 to 50 percent and 35 to 40 percent of the installed-base, respectively, says Massimo Pezzini, VP & Gartner Fellow, Gartner. “The rest of the world accounts for the remaining 10-20 percent,” he adds.
One worthy insight comes from SAP, whose in-memory offering, SAP HANA, has been the most successful product launch in the company’s history.
“Regionally there is extremely high interest and the market is beginning to understand the potential of this technology,” says Sam Taha, Principal In-Memory Architect, SAP MENA.
“As a result, we expect that adoption in this region will soon catch up with other parts of the world such as Europe and America.”
Holding back that adoption, for the time-being at least, appears to be the factor of awareness, amongst others.
“There may be a lack of understanding of what IMC actually does,” says Nassir Nauthoa, GM, GCC Intel Corporation. “But, also, there’s a lack of standards and a generally unfamiliar architecture to most organisations’ IT departments. In time these points will get addressed to enable wider adoption.”
As a result of this, many IT organisations are still investing in traditional application infrastructures, Taha says.
“Also, many businesses have already invested heavily in legacy and home-grown applications, which have likely grown in undocumented complexity over time, making it difficult to migrate to a new environment,” he adds.
So, the key step towards increasing Middle East adoption is increasing the awareness and understanding on its capabilities.
This comes only through educating customers, and the customer’s willingness to adopt new technology, according to Nauthoa.
“It’s about sailing in unchartered waters sometimes and it is incumbent on suppliers to accelerate know-how around IMC in order to grow adoption,” he says.
Once that understanding is there, the burden falls onto the CIOs to ensure their teams have the right skills to exploit in-memory computing.
“IMC skills are scarce and rare,” Pezzini warns. “So they are not easy to find in the market.”
The best way to start building these skills internally, he says, is by proof of concepts.
“Another way,” he adds, “is through pilot projects that target circumscribed, but not trivial, projects targeting grow-the-business types of requirements.”
However, Taha, perhaps understandably, disagrees. He believes most existing database and application development skills are directly transferrable to in-memory computing platforms, particularly when these platforms adhere to industry standards.
He does add, however, that in-memory platforms also introduce new paradigms in the way data is managed, modelled, and presented.
“For example, logical views and virtual models are powerful new concepts that should be learned and exploited,” he says. “The new skills required will be focused on exploiting the in-memory platform for increased productivity, agility and flexibility, rather than on tuning and administration as with traditional systems.”
Nauthoa adds it is vital for vendors to offer IMS training. “We encourage CIOs to make sure their teams take up all opportunities to learn about this new technology to capitalise on sales opportunities for IMC enabled systems,” he says.
An important consideration when planning an in-memory deployment is the amount of re-engineering that may be required to make the most out of the new platform.
According to Pezzini, in many cases, the re-engineering efforts are not that extensive, although testing the application for the new environment can require substantial efforts.
“In order to get the most out of IMC, such as introducing real-time analytics to transactional processes, significant application engineering may be needed,” he says. “Often, this re-engineering can be done incrementally, rather than through a “big bang” approach.”
Considering no one application is built the same as another, Nauthoa adds that it really depends on the applications in question.
“Some may require a minimal investment to re-engineer in order to maximise the in-memory platform,” he says.
“Once the industry regulates by creating standards — APIs, tools and protocols — then it will become much easier for developers to create applications that are in-memory computing efficient, and in turn much more cost effective to engineer or re-engineer applications.”
Perhaps more vitally, implementing IMC brings with it a requisite for CIOs to rethink their application architectures.
IMC opens new, partially unexplored opportunities that may imply dramatic changes in organisations’ application architectures.
However, many vendors are laying out approaches CIOs can take to deploy IMC in a minimally disruptive way, Pezzini explains.
But before that is done, CIOs must of course face the task of selling the project to senior management and thus securing the necessary project budget.
According to Taha, with IMC, ROI can be based on various metrics.
“These could be time and cost saved running a month-end closing cycle for financials, which has an immediate impact on bottom line,” he says. “Additionally, intangible benefits have to be taken into consideration, such as the cost of an IT employee’s time to maintain outdated database techniques.
Pezzini says that, although it is possible that IMC helps reducing certain types of IT costs, IMC-adoption ROI can only be measured in business terms by looking at things like customer retention, revenue growth and greater product quality.
“For example, IMC enables ecommerce Web sites to run faster, and this turns into revenue growth,” he says. “But entire industries, such as online gaming, financial trading, online advertisement, travel reservation, SaaS and social networking, could not exist — at least not as we know them today — without IMC.”
Thus, the original question remains, when will adoption of in-memory computing reach a mainstream level in the Middle East?
This will depend on things such as expanding messages on the benefits, investing in enabling the sales channels to seek-out opportunities, and being able to demonstrate tangible gains in ROI to CIOs for adopting IMC systems, Nauthoa says.
“All this is readily achievable and the more that such things are done the quicker the technology will gain traction,” he adds.
Taha says that SAP is seeing the growth of in-memory at a “very fast pace” in its MENA customers. “And lots of interest from businesses that aren’t SAP customers but are extremely excited by the prospect of holding all their data in an SAP HANA database.”
Pezzini concludes that, given the significant SAP penetration in large Middle East organisations, he expects a “pretty rapid uptake” in the region, at least on the part of the most advanced IT users.
“However, at a global level we think it will take at least another three to five years, possibly longer in some cases, for IMC to be adopted on a mainstream basis,” he adds. “The Middle East is likely to follow a similar path.”