Is storage headed for irrelevance?

Fig 1 THE growing pervasiveness of mobile products like MP3 players and personal media players (PMPs) has stirred speculation that within a decade the electronics industry could produce a portable device that will have sufficient storage capacity to hold all the music, movies, and television shows ever made. If this comes to pass, it raises an important question: When everyone can carry such a device, will storage technology become irrelevant? Of course, the need for storage won’t go away, but if it becomes as cheap and readily available as tap water, will it lose relevance as a technology market?

This potential scenario has important implications for the Flash memory, HDD, and networkbased storage industries, because their growth is predicated on the anticipated need for ever-increasing amounts of storage at ever-decreasing costs. If everyone already carries such a limitless library of media, what need will there be for new HDD or NAND products with more gigabytes of storage every year?

iSuppli Corp. analysts recently debated the issue and came to the conclusion that it’s too early to predict that storage, and the continuing development of the storage industry itself, will someday become moot. In fact, iSuppli analysts foresee an intensifying battle between HDDs, NAND flash memory and networkbased storage to serve the rising storage needs of consumers and corporations around the world. However, they also noted there are signs that the global appetite for more storage is not insatiable.

Much of the skepticism over predictions of storage’s demise arise from the poor record for such prognostications in the past. Krishna Chander, senior analyst, storage systems for iSuppli, and a former senior engineer at IBM, recalls hearing warnings in 2000 that the need for increased storage capacities was soon to exhaust itself.

“Six years ago, when we were introducing a 20Gbyte 3.5-inch HDD at IBM, some of my colleagues fought against the product, arguing that users had stopped buying higher-capacity drives once they had reached the 4.3Gbyte level,” Chander recalled. “Well, the sweet spot for desktop- PC HDD capacities now has risen to 80Gbytes and it’s still increasing. In fact, drive suppliers don’t even make 20Gbyte HDDs for desktops anymore.”

Chander noted that observers persist in making similar arguments regarding MP3 players, but storage requirements continue to increase, driven by the need to carry new media types, such as video. Mark Kirstein, vice president, multimedia content and services for iSuppli, agreed with Chander’s assessment.

“Every statement in the past that there is enough storage to meet all needs has proven to be folly,” Kirstein said. “In the past, the content was text, then it was sound, then games and now video. Video consumes a lot of storage, so it may take more than 10 years before storage capacity exceeds demand generated by it.” Kirstein anticipated that a new driver would arise for increased storage after video.

Need for NAND
Nam Hyung Kim, Director and Principal Analyst of memory ICs/storage systems for iSuppli, said there is some concern among chip suppliers that current content types and storage applications will not drive sufficient demand growth for NAND-type Flash memory in the future. However, he noted that the memory makers are working to mitigate this concern by finding new uses for NAND.

“A major challenge for NAND suppliers is to entice users to spend more money to buy more bits of storage,” Kim said. “Unless they can find some killer applications to drive demand for more NAND bits, the market will be risky for them. This is why the NAND suppliers, Samsung Electronics Co. Ltd and Toshiba Corp., are working furiously to develop new markets for their products.”

The most obvious path to expanding the use of NAND is price reduction. Although recent declines in NAND pricing have diminished profitability at memory suppliers, the long-term impact of these reductions is beneficial from a strategic perspective, Kim said.

NAND vs HDDs
Kim said the storage battle between NAND and HDDs will continue, with flash memory making inroads into markets traditionally served by rotatingmedia drives.

The cost gap between NAND and HDD is closing, making Flash a more attractive alternative for some applications. In 2003, 1Gbyte of NAND was nearly 100 times more expensive than an equivalent quantity of HDD storage, according to iSuppli. By 2009, that price gap will dwindle to only a factor of 13.9 times, as illustrated in Table 1.

Meanwhile, Kim believes that the increase in storage requirements for some mobile PCs will slow in the coming years, opening an opportunity for NAND Flash memory in computers. By 2010, at least 15 percent of mobile PCs shipped will use NAND Flashbased solid-state drives (SSDs) in place of traditional rotating-media HDDs, he predicted.

Flash-based SSDs not only will provide users with faster boot times and quicker data access, they also save tremendous amounts of system power and offer superb drop-shock resistance. Because of these factors, Flash eventually will eliminate traditional rotating HDDs at lower densities in PCs, Kim said.

However, one possibility could stop SSDs in their tracks: If users don’t need more than 100Gbytes of storage. If they do, the prospects for SSDs in mobile PCs will be severely curtailed.

Network option
Beyond HDD and Flash is another approach to stockpiling user data: network-based storage, Kirstein said. The premise is that as more bandwidth becomes available, consumers and businesses will opt to put their data “in the cloud,” i.e., on the network. This has some appeal, particularly as it allows access to the data from nearly anywhere.

Companies like Streamload Inc. are championing this vision. However, there are challenges to the networkstorage scenario, Kirstein warned. Among them, the size of content continues growing, keeping ahead of bandwidth increases.

Secondly , there’s the challenge with copyrighted content, like music and video. Media companies don’t want their content pushed up to third-party servers, and it’s not clear that “fair-use” laws would support it.

A third challenge is privacy.

What’s in store?
In the past, predictions that storage capacity needs will reach a plateau have turned out to be false. It’s likely that these predictions will fall flat again. However, a slowing in the increase of storage requirements could open opportunities for NAND Flash and network-based storage as they compete with HDD in applications like mobile PCs.