In January, IBM declared war on shamplex users, companies that had purchase coupling facilities for their mainframes to qualify for the hefty software discounts available through IBM’s Parallel Sysplex License Charging (PSLC) initiative but had no intention of making the next step to sharing workloads between nodes in the mainframe. The company stated that, after a grace period, users would have to show that at least one application or system function running actively-coupled across all capable sysplex images and supporting at least 50% of each image workload. Otherwise, they would no longer qualify for PSLC, lose the discount (which amounted to as much as 30% off the IBM software bill), and would have to revert to Graduated Monthly License Charge (GMLC) (CI No 3,331).
By David Johnson
At the announcement of its G5 mainframes in May, the company announced the grace period would end on September 30th. Mainframe users, already battling withY2K projects, balked at the prospect of their IBM software bills leaping 30%, and since then IBM has been looking for an exit strategy that will let it avoid nasty confrontations with some of its largest and most profitable customers. As reported yesterday (CI No 3,501), IBM will unveil a new pricing structure next Wednesday, September 30th, that will scrap all other pricing models than PSLC and will shift to usage- based pricing. Most importantly, it will solve the problem with the shamplex user, since there will be no longer be any GMLC charges to revert to. The new pricing model will use PSLC as its basis with a series of discounts on top according to software usage. This is by no means IBM’s first attempt at usage-based pricing. In 1994, chairman Gerstner met 200 mainframe users in Chantilly, Virginia who – as he put it: shook me warmly by the throat over the issue of his company’s software pricing. He charged two of his deputies to come up with a fairer system within 24 hours. The result: removing the two highest GMLC software groups and Measured Usage License Charge (MULC). But MULC failed to attract either users or ISVs. It was too complicated, requiring hundreds of changes to IBM system software, and widely unpredictable for IT managers. Charges were based on usage on the fourth highest day of the month and a small change in usage levels could easily catapult users into the next software bracket. As one mainframe analyst puts it, You could barely fill a taxi with UK MULC users. The take-up was minimal. And this was the case globally. Next week’s pricing model is expected to have ironed out many of these problems: usage will be calculated on the average of a whole month over a six month period and ultimately the MULC software brackets will be abolished altogether. Instead, users will be charged strictly according to MSUs (millions of service units) a piece of software uses. As well as getting round the problem it created with its shamplex crusade, IBM hopes that usage-based pricing will make S/390 a reasonable option as a consolidation platform and for new applications. And it was partly for this reason that IBM said that it was considering a move to usage-based pricing at the G5 launch. Many mainframe users have applications running on Unix and NT servers that they would consider migrating to the mainframe. But under GMLC, it was cost-prohibitive. The software might require only a couple of MIPS but would be charged according to the size of the whole mainframe. This is said to have hurt mainframe Lotus Domino sales. One UK supermarket chain wanted to run Domino on its mainframe until it looked at the effect it would have on its software budget, even though Domino is sold purely on a usage basis. Domino would have required another 200 MIPS of processing power in the company’s mainframe complex. Under IBM’s existing mainframe software pricing model, this would have pushed up the software charges of every other piece of IBM software. The supermarket walked away from the deal.
Content from our partners
