ORLANDO -- Dramatic advances in computer hardware, particularly multi-core CPUs, over the next 12 months will in...
turn engender a new approach to building enterprise applications, according to Hasso Plattner, chairman of the supervisory board and co-founder of SAP.
At Sapphire on Wednesday, Plattner shared his vision of the future of software engineering in a keynote speech titled "The Power of Speed."
Many companies are working on ways to make software cheaper, with less maintenance cost. "And all that is good," Plattner said. "But we want to make everything tremendously faster and change the way we work."
For instance, companies today can collect "unbelievable amounts of data," he said, noting that the average SAP customer has seven to 10 years' worth of data on disk. But, he noted, "how we digest that data is slow, and it's getting slower because of the increased sizes of databases."
What Plattner wants to see is the ability to extract information from databases instantaneously. "For a company the size of SAP, with $15 billion in revenue, I want to answer every question you have with [Microsoft] Excel or [SAP] Business Explorer in less than one second."
The ability to do this will require cheap and fast main memory, multi-core architecture CPUs, and relational database management systems that rely on columnar storage, in which data is stored by fields instead of records-based storage. It will also require that, in Plattner's words, enterprise software engineers understand how computers work.
"Gaming software developers are exploiting this technology," Plattner said. "Now enterprise software manufacturers have to follow."
"We will soon have, on one board, eight CPUs, and each of those CPUs will have 16 cores," he predicted. "This will mean 128 computing units on one board, called a blade. Today a blade can hold 144 GB of memory and costs about $6,300. Now, if you take 100 blades, you'll have 12,800 computing units for less than $1 million. That is not much money, relatively speaking, for all that computing power."
Columnar storage is much faster than records-based storage, according to Plattner, since it allows for 10 times more compression and because database updates will be done as insert-only operations, i.e., simply adding fields to strings.
Extracting information from columnar storage databases is be much faster, Plattner said, "because the area we have to access is much, much smaller than in a database with records storage." He predicted that with columnar storage, companies will be able to run through 4 million sales records in 700 msec. "And when we go to the new multi-core boards," he continued, "a user will be able to get a response to any database query in under 100 msec."
If Plattner is right, it could be "the beginning of the end of the relational database as the mainstay of enterprise computing," according to industry analyst Josh Greenbaum, principal of Enterprise Application Computing. Asked whether that could occur within the next 12 months, as Plattner predicted, Greenbaum said it was possible, but he called the schedule "more of a marketing-driven time frame than technical-driven time frame."