Avoiding Vendor Lock-in – ITC Shows the Way

Summary:

Vendor lock-in is a treacherous shackle that restrains most enterprises’ IT systems, thereby restricting business growth. Read on to know how the ITC Hotel Group found a way to break this shackle.

The USP of the luxury hotel group ITC has always been offering traditional Indian hospitality to the judicious traveler. With over 90 hotels in nearly 70 destinations across the country, the group amalgamates time-honored tradition with globally-benchmarked services and sustainable business practices.

But when tradition begins to choke business development, it’s always a good idea to tip the scale towards change. And that’s exactly what Bhujay Bhatta, operation manager- IT shared services, ITC Hotels, did.

Until last year, ITC Hotel’s business-critical systems ran on a RISC-based platform, solely dependent on a limited choice of vendors for support. “Applications like ERP and CRM were traditionally run on proprietary RISC-UNIX environment. In all of this, managing a UNIX platform was extremely expensive,” says Bhatta. This made maintaining features like high performance, scalability, and RAS (reliability, availability and serviceability) incredibly expensive. “But application and hardware vendor support was superior to its counter part—the CISC-based platform. And that kept us going,” adds Bhatta.

However, the RISC-based server industry received major jolts last year when big players like Oracle stopped developing software for the platform (the battle between Oracle-HP was one of the major reasons). Many enterprises that still employed RISC servers in their infrastructure could now rely on only a limited choice of vendors for support.

Handling aging RISC infrastructure with leases inching towards expiry dates, longer payback duration for a significant capex, high power consumption, datacenter space, and cooling costs became challenging. By putting backbone applications like ERP at stake, the business was risking a downtime that could last up to months. Left with no choice, Bhatta perceived a huge risk—in the long run—to support the company’s infrastructure.

“We began to sketch an alternate platform strategy. Our leadership team along with the internal IT team studied future trends in server hardware and OS development,” says Bhatta. The research had to be meticulous and in-depth as the new platform would have to match or be better than the performance and reliability that the RISC platform provided. “It had to give us cloud-like benefits like elasticity and agility. But most importantly, it had to have a low cost of ownership. We finally zeroed in on Open Source,” says Bhatta.

His internal IT team developed an innovative infrastructure design built on open source with a combination of different infrastructure components. “We focused on details such as server hardware type, suitable Linux variant, type of virtualization and clustering to deliver a superior infrastructure platform that would reduce IT cost,” he adds.

Also, this time Bhatta was careful not to get locked in by completely eliminating dependency on server hardware. “We chose SUSE Linux for SAP and Redhat for other important apps. Existing SAP servers which are due for technology upgrade will also be migrated to Linux-Xeon-based environments. Apart from a direct benefit on account of the low cost of Xeon-Linux Environment, this would also free up space, power and cooling capacity from our datacenter,” Bhatta says. The entire landscape has been implemented on Xeon-Linux servers with 400000 (SAPS) or 250 traditional RISC-UNIX equivalent CPUs.

The options to move towards change now or later are always questioned, but never for the necessity to change. As Bhatta says, “It’s all in the game of taking smart risks when the stakes are high.”

Shweta Rao is a correspondent at CIO & Computerworld, India.For feedback, mail her at Shweta_rao@idgindia.com.Follow her on Twitter at @Lazzylass

We chose SUSE Linux for SAP and Redhat for other important apps. Existing SAP servers which are due for technology upgrade will also be migrated to Linux-Xeon-based environments. Apart from a direct be

We focused on details such as server hardware type, suitable Linux variant, type of virtualization and clustering to deliver a superior infrastructure platform that would reduce IT cost