While no one disputes the extra power and performance that dual-core processors provide, there are still looming questions about what this means for enterprise software costs. Specifically, will the gains from multi-core processors be negated by additional software licensing costs?
Dual-core technology refers to the presence of two physical processor cores — the unit within the actual processor itself that performs computing calculations. As Intel Corp., Sun Microsystems Inc., IBM Corp., and Hewlett-Packard Co. come on board with multi-core offerings, the issue is how will software vendors charge for the software that runs on top?
The answer is still not clear. One reason for the confusion is that software vendors have not regarded chips with simultaneous multithreading (SMT) or Intel’s Hyper-Threading technology, which make single processors work as though they were two processors, as more than a single processor. With chipmaker Advanced Micro Devices (AMD) preparing to launch dual-core Opteron processors by mid-2005, the company has stated that software vendors charge per physical socket, essentially treating the multicore processor as a single unit.
HP recently unveiled the dual core “Mako” PA-8800 RISC processor for its Unix servers. And Intel plans to have more than half of all Intel’s processors as dual-core by 2006, moving to multiple-core chips on a single piece of silicon.
What’s certain is there’s still a lot of uncertainty in the market, said Steve Shaw, business development manager for enterprise servers for Hewlett-Packard (Canada) Co. in Mississauga, Ont. HP licenses its own software by the server. As technology evolves, HP has taken the position to license by the socket, Shaw said, adding that this eliminates any discrepancy.
“That way as innovation comes with the chip technologies customers are still going to get an appropriate amount of software licensing to go along with it.” Industry observers note that enterprises plan IT budgets by negotiating software licences by counting a single-chip device as one processor. The intended aim of enterprises for using multi-core processors is to cut down the system power requirements and heat generation and get the most performance out of a single chip; this is cancelled out if software vendors start imposing a penalty on buyers.
The direction of the market, both on the hardware and software side, however, is definitely leaning toward counting a multi-core/dual core unit as a single CPU, said Warren Shiau, a software analyst for IDC Canada Ltd. in Toronto.
That’s why the hardware vendors say multi-core units should either simply be counted as a single unit, or counted by socket — which makes them count as a single unit — or that software licensing should go on a per-user basis. The current reality is that users have been paying a premium to certain software vendors who consider a dual core unit to be two CPUs “and if users are willing to buy into that then you can’t call it anything but fair,” Shiau said.
While most software vendors still are feeling their way around the multi-processor issue, Oracle Corp. recently made changes in its contract language to ensure that there is no confusion. In an e-mailed statement to ComputerWorld Canada, Jacqueline Woods, vice-president, global pricing and licensing strategy for Oracle, said Oracle would continue to use existing licence models and offer new models to account for environments such as multi-core processors, soft partitioning and enhanced utility models. In April, the vendor changed its licence agreement to redefine its processor metric to explicitly state: “For the purposes of counting the number of processors which require licensing, a multi-core chip with ‘n’ processors shall be counted as ‘n’ processors.”
W.L. Gore & Associates Inc., best known for its Gore-tex fabric, negotiated a software licence with Oracle late last year and then began updating its Sun hardware. In the process of bringing in servers based on Sun’s first dual-core chip, UltraSparc IV, the Newark, Del., company discovered that Oracle considered each dual-core UltraSparc IV to be two processors for licensing purposes. That meant W.L. Gore was faced with paying double what it originally expected to pay for licensing, amounting to US$100,000 per server in additional costs.
“When we first heard this, we weren’t happy, of course,” said Richard Sun, network systems engineer at W.L. Gore. “In all of our initial conversations with Oracle, before we purchased the licensing, we had said we were going to buy the (Sun) hardware when we were ready to implement, so we felt going to UltraSparc IV was no issue.” Eric Kuzmack, IT architect at newspaper conglomerate Gannett Co. Inc. in Silver Spring, Md., says he deploys dual-processor systems in order to have hardware redundancy.
“If the software vendors choose to charge us for two processors when we move to dual-core, it makes it more difficult for us to afford to put in two processors because then I’ll be buying two physical packages of chips, but I’ll have to pay for four processors worth of licenses, not two,” he says.
Dan Kaberon, director of computer resource management at human resources service firm Hewitt Associates in Lincolnshire, Ill., agrees that today’s software licensing is lagging behind advances in hardware and systems management.
“I would just like to see a rational alignment of licensing so that if we’re able to collapse multiple workloads onto a smaller number of servers that our licensing reflect that new environment and not pretend as if we were still taking advantage of all of our dedicated machines,” he said. — with files from