Comparing Top Ten Lists: IBM Vs. BCC Technologies.

The two companies discuss BCC's internal disks-publicly.

When David Letterman first presented his Top Ten lists, the world was listening. Since then, we've seen lists ranging from the Top Ten Signs You Have a Bad HMO (#1 is "You get a 10% discount on your co-pay if you do the surgery yourself') to the Top Ten Signs You've Read Too Many Top Ten Lists (#2 "You wonder who submitted the 10 Commandments.")

Not to be outdone, IBM and BCC Technologies have written Top Ten lists themselves about BCC's Extender Internal AS/400 disks. IBM got off the first volley in March 2000 with their paper, "Top Ten Considerations When Evaluating BCC Technologies' AS/400 Internal Disks." BCC shot back in April with its own paper bearing the same title. The papers are not as entertaining as the Top Ten Reasons Why The World Wide Web Is Out Of Hand, but they're a good deal more useful to resellers, VARs, and IT professionals in charge of evaluating internal disks for IBM's AS/400.

The issue began over claims BCC made for its Extender Internal AS/400 disks, which utilize IBM's own Ultrastar 18LZX with different microcode, electronics, casing, and power regulators and which began shipping in summer 1999. The claims boiled down to plug-compatible, better performance, and lower prices for the BCC drives. So in response to those claims, IBM issued its Top Ten Considerations in March 2000, regarding BCC's claims. The timing is likely due to IBM's March shipment of its own Ultrastar 18LZX drives, even though its Information Technology Group had been providing the drives to hardware vendors such as BCC since May 1999. (BCC is now shipping drives based on IBM's Ultrastar 36LZX.)

Technology

IBM begins by pointing out that BCC drives are neither comparable nor identical to IBM drives because the microcode is different and there are additional hardware components. IBM's microcode fixes or updates are available for download via the IBM Web site, but BCC disks cannot use IBM microcode. Further, IBM reported that "the BCC microcode and additional hardware were slower than with genuine IBM AS/400 disks." There is also the concern that BCC disks are controller (IOP) and/or system-dependent.

IBM makes a statement that "IBM has ... tested some BCC disk drives. We found they did not meet the same level of performance as 'equivalent' genuine AS/400 IBM disk drives."

There were no test results included, a fact that BCC questioned. But Ian Jarman, AS/400 product manager says, "We're talking about a system that's very integrated where all these elements work together. We're coming from a system view and want to bring it back to what's more relevant to users."

IBM tested not only the disk drive itself but also the entire throughput path to arrive their conclusions. IBM counters BCC's claims for FAST technology. The technology uses only half the potential physical disk capacity order to speed disk arm movement (seek time). The disk can increase capacity but loses its performance gain. IBM points out that customers using IBM disk drives and only filling them to half capacity will realize the same performance gains.

BCC counters by saying that it does not expect customers to watch for themselves when a disk is at half capacity and that the FAST technology reports 90%-95% utilization which is when performance begins to suffer.

IBM makes a second point, that "disk level performance and disk subsystem performance are different. BCC's claims are made at the disk level." They make the distinction between disk seek time at half that of IBM disks and disk access time--the two are not the same. IBM's Jarman adds that system-wide elements are important, as well. Seek time is important, but other components such as the I/O processor and caching are involved in judging total throughput and performance levels.

Pricing

IBM claims that "BCC disk drive units can be priced lower than IBM OR potentially faster (under specific conditions), but not with the same disk drives." IBM refers to the premium price for BCC's FAST technology as it goes on to warn the customer to consider additional costs for installation and maintenance.

BCC provides a chart comparing list prices between similar disks without the addition of FAST. For example, the IBM 8.58GB list price is $2,000 for a price per megabyte of 23[cts.] while BCC's 8.58GB list price is $1,484 for a price per megabyte of 17[cts.] Street prices for the 17.54GB drives run about $3,600 for IBM's and $2,500 for BCC's.

IBM does not charge for installation and first year maintenance, while BCC does. A typical charge for a 30-drive installation is $929 for installation and $2,700 for first year maintenance. After the first year, IBM will charge yearly maintenance fees at about $3,600 a year for a similar installation.

In fact, BCC claims that a comparable disk setup from BCC instead of IBM is roughly half price for the customer. As BCC stated in its document White Paper: Operations Management "Significantly lower cost of ownership than current AS/400 drives shipped by IBM is achieved because as little as half the number of Extender drives are needed for similar performance levels." BM takes issue with this, as yell they might. IBM reports that BCC makes the "half as many drives" claim based on 115OPS (Operations Per Second) while IBM's drives only deliver up to 62OPS. IBM does not believe that BCC can achieve this rate in real world customer environments and questions the utilization level BCC is using when they claim 115OPS.

BCC answered by asking to see IBM's test data and insisting that their drives can too achieve 115OPS. They did not detail the level of utilization used in their own tests. The sniping continues around the comparison models BCC used for its white paper with IBM insisting that BCC's lab setup included IBM units that were "older/slower, more expensive, and are not configured in the most cost-effective manner."

Support

Of course, pricing is only part of the story.

Barry Parker, president of AS/400 integrator D.S.P. Inc. said, "A SCSI drive is a SCSI drive, but IBM does offer really good support and some customers will pay a premium for that support."

So IBM, justly known for its customer support, takes BCC to task for its warranty policies. The Basic Warranty is for a five-year period and requires the customer to ship disk drives back to BCC for repair within two business days from receipt. For additional fees, the customer may purchase Extended Warranty (additional one-year warranty) and On-Site IBM Warranty Service (BCC dispatches an IBM service engineer to a customer's site or does repair within one business day at its facility).

Another flurry of claims centers around BCC customer repair needs and IBM Global Services. Although IBM admits that BCC has contracted with IGS to provide "screwdriver for hire" services in replacing BCC disk drives, it says the customer must place the service call to BCC, not IBM. This is true, although the initial service call will most likely be to IBM, since an AS/400 failure may be due to any number of failing parts. If the IBM CE determines that the failing part is a BCC device, the customer, then, calls BCC ServiceOne, who will also dispatch an IBM CE. (In most cases, this will be the CE already on site, though there is no guarantee that this will happen.) From then on, BCC and the CE will work together to solve the problem, including using replacement parts arriving from BCC's warehouses or the customer's on-site spares.

In the end, both companies make high capacity, reliable disks based on the same Ultrastar 18LZX disk drives. The points of comparison are speed based on seek time, price, and customer service. Customers will-and should- put different weights on different points according to their environment, needs, and comfort levels. Thus, IBM makes a good point: customers should carefully evaluate all of BCC's claims, making sure that they understand the criteria involved and their own environment's needs. BCC concurs.