STORAGE - Never-ending storage
Tape storage is dead. Or so the pundits of optical technology would have you believe. However, this storage veteran has shrugged off more than one challenge in its time.
With data protection increasingly high on the corporate agenda, and have you believe. However, this storage veteran has shrugged off more than one challenge in its time. phrases such as storage area networks Sans and DVD becoming the buzzwords of the latter-day computing generation, the late 1990s have seen the storage market enjoy something of a renaissance. Capacities are up, prices are down and storage vendors are falling over themselves to ensure they have a bigger slice of the action than the next guy.
High capacity, reliable, cheap and freely available, tape has spent the best part of half a century fighting the magnetic corner of the storage market - and winning, even in the face of newer, more glamorous competition. Now, however, pitted against the younger and ever more ambitious, emerging contenders, tape is, more than ever, the old man of the storage market. And what's more, its age is beginning to show.
While the long-term prospects are decidedly ropey for tape storage against its optical contenders, it's still very much a subject of debate as to when tape will be superseded by a cheaper and better alternative. In light of industry speculation that tape is dead, how much longer can it hold out against the pretenders to its storage throne? According to some, it might be a lot longer than its rivals would like.
Steve Georgis, director of technology and business development at Exabyte, says the potential longevity of tape technology has been consistently underestimated over the years. 'I have been working in the tape industry since 1978. Tape was pronounced dead that year by supposed industry visionaries and the end of its reign has been the source of constant speculation ever since. However, even as other storage technologies have continued to advance - as does tape - nothing has really ever come close to replacing it, except perhaps on the desktop where requirements are far lower.'
This may be viewed by some as the defiant vitriol of a dying technology, but not according to Mark Reeves, European vice president of Tecmar Technologies.
He agrees with Georgis and adds: 'Many experts have been predicting the demise of tape for 15 years or more, but the fact is, there's still no other technology that can offer either the comparable price per gigabyte or the same capacity point for a removable storage medium.'
Georgis says that even with the cost of hard disk-based storage falling, tape continues to hold its own comfortably, as hard disk drive (HDD) storage is still comparatively expensive at about the $20 per gigabyte mark. 'Compared with HDD pricing, the cost of tape sits in the lowly $2-$4 per 1Gb ballpark, even for the most high-performance drives. Moreover, we expect the price of tape storage to drop even further - to about $1-2 per Gb - over the next two years or so.' Such a wide difference in price makes all the difference in applications where high capacities are needed, particularly where budget is a motivator, he adds.
However, Bob Peyton, senior storage analyst at IDC, says that although tape storage price points are low and falling, and tape's cost per gigabyte is still much lower than that of hard disk-based storage, tape is ultimately still under threat in the long term. 'Hard disk cost per gigabyte is now falling at a rate of about 40 per cent each year. It's not going to be many more years before HDD prices fall to levels where they will present a significant threat to tape - even to the newer technologies such as libraries.'
This seems to suggest that it is not a question of if, but when tape will finally die. Once again though, the vendors paint a very different picture. According to Mel Taylor, business development manager at StorageTek, HDD-based storage might not be catching up as fast as some in the industry would have us believe.
'While the projected average cost per megabyte for disk drives may sound appealing at first glance, some estimates can be misleading.
Enterprise-class storage needs are exponentially higher than a single megabyte. Additionally, the published cost for disk drives alone does not always reflect the cost of a total system. It is only when one considers the actual amount of data being stored and the actual costs of the medium that the most effective alternative becomes clear.'
Most tape markets are actually growing, according to IDC. Paul Sangster, sales and marketing director at Hammer Distribution, says this is with good reason. 'First of all, back-up applications - which need to be cheap - are still a important part of the IT mix. There isn't a suitable alternative to tape and unless a new technology emerges, tape will continue to be the only viable option here. Secondly, tape has such a huge installed base that no one is going to ditch the technology lightly. In addition, many of the key storage players are still heavily invested in tape technology and for that reason alone its future is pre-determined.'
DVD, or a hybrid thereof, may take the place of tape eventually, but at the moment it simply can't compete on capacity, so is not a viable option. Conversely, HDD storage can compete on capacity but not on price.
Tape's future prosperity looks dependent on the gap between the two.
According to Sangster, tape's familiarity will also help it maintain market position. 'The tape market is much more established than some of its immediate competitors in terms of market penetration, with systems freely available for entry level, mid range and high-end usage. In the longer term, other technologies may threaten its dominance but, for now, tape sales are increasing. At Hammer we are seeing our tape business continue to grow month on month. There is still a great deal of margin in it for everyone, resellers included.'
David Huntingdon, sales and marketing director at M4 Data, agrees that tape is providing good systems for customers and good margins for resellers.
'The way I see it, there is little chance that this market will erode,' he says. 'Tape storage will continue to offer the lowest cost, most reliable method of backing up large amounts of data. And by using proven tape technologies with newer, more sophisticated library automation and software techniques, tape will continue to give users the best value for money.'
Indeed, several companies have joined the tape market in the past 10 years. These include Ecrix, with its 8mm technology, Benchmark (DLT) and OnStream (DC2000). The arrival of these new players, coupled with the emergence of standards such as Linear Tape Open (LTO) and the continuing presence of the vendors which developed it, such as Hewlett Packard, IBM and Seagate, seems to indicate there is life in the old dog yet.
Mike Hill, vice president of European marketing at OnStream, comments: 'What other form of storage media will feasibly, within a couple of years, deliver half a terabyte of removable storage on a single cartridge about the size of two audio cassettes? How many Zip disks or DVDs or CDs would it take to store that amount of data - and at what cost?
'Additionally, the storage element of a server now makes up as much as half the build cost. Under those auspices, tape suddenly makes a lot of sense. OK, you can't run core business applications straight from tape, but tape has never really tried to compete in that sector and has never claimed to be able to.'
Trevor Duplock, sales and marketing director at Samsung - a company which claims to have grown its storage business from a standing start to more than $500 million - says that although there is still room for tape technology as a mass storage medium for networks in the short term - potentially two or three years - the desktop represents a very different story. 'We should expect to see cheaper, more reliable DVD Ram products beginning to have an impact on desktop back-up by the end of the year,' he says.
'Whichever way you look at it, tape technology is a dinosaur and tape manufacturers have now realised they have no future on the desktop - tape is too slow and expensive in that area. They (the tape vendors) are all focusing on high-end library functions, but in time even these will be replaced by DVD in jukebox-type devices.'
OnStream's Hill disagrees in part, believing that tape technology is going to make a concerted move back towards the desktop over the next few years - particularly at the low end. 'Until recently, tape technology was always marginalised and uncompetitive at the desktop because it priced itself out of that market. The average desktop user had no use for the capacities on offer and therefore refused to pay the price being asked.
'However, with the advent of more storage-intensive software and applications, desktop capacities have increased and so has the need for online storage.
As this need grows in corporates across the world, buyers will begin to see the benefit of tape once again - in terms of cost if nothing else.
The demand for storage has grown exponentially.'
Hill explains: 'The other day, I saw an advertisement from a key PC vendor for a desktop with a 17Gb hard disk drive and a 100Mb Zip drive. How will the user of that machine back up his data? The 100Mb Zip is no use. In the end, what purpose does that Zip drive serve? What the user needs - if he isn't backing up to a server - is a 30Gb tape unit. Even if he is backing up to a server, the server will itself need to be backed up at some point.'
There is another key reason for this turnaround, according to Hill. If the individual user is to be charged with the task of backing up the machine himself at the desktop, a tape-based system makes far more sense. Changing a hard disk drive cannot be easily entrusted to the user; a tape cartridge, on the other hand, can be replaced by the user within seconds.
Hill maintains that the transfer rates of today's tape technologies mean users can deploy applications such as full-motion video playback without suffering any loss of quality. 'With some of our products you can play full-motion video directly from the tape, rather than having to move it to disk first. This cuts the time to desktop considerably.
'With the kind of capacities that tape is now delivering, you're soon going to be able to fit 100 full-length motion pictures onto a single cassette. Once again, what other media can offer that kind of capacity at the price?'
He recognises that the lack of random access to data discourages many purchasers from buying into tape technology, but suggests that this argument may now be a red herring. 'In light of the newer, hgher-functioning tape technologies now coming to the fore, the random access issue is no longer a problem for tape,' Hill says.
'Along with the technology's image, the lack of random access to data has always been the biggest stumbling block for tape-based storage in the past. But it's no longer necessarily a problem. The speeds of tape drives have increased significantly in recent years and will continue to do so. In addition, performance can now be further enhanced using software.
Going back to the movie analogy - any one of those 100 films can be found sequentially and started within just 20 milliseconds using the new advanced digital recording (ADR) technology.
'While it still isn't really genuine random access to data, the user doesn't really care as long as the information he needs is there quickly.
In this way, if you can at least give the appearance of random access, if not the real thing, then the whole performance argument becomes irrelevant,' Hill claims.
Jim Dawson, vice president and general manager of European operations at Data General's Clariion, says the needs of users are beginning to change and tape will have to change along with them. 'In future, companies will have to look at their storage system in terms of how important it is to their core business. Can they scale their storage capacity over time?
Is it open? What's the technology roadmap? This is how many industry pundits view the advantages of storage area networking (San),' Dawson says.
Some believe that tape does have a role to play in the upcoming San area.
Peter Coleman, technical manager at storage area networking specialist Gadzoox Networks, argues there is still room in the storage market for tape. He says that in addition to its natural resilience in disaster situations, this is due mainly to the very technological advancements that are threatening it elsewhere. Fibre channel technology, for instance, allows added flexibility to tape functionality, such as remote access to tape libraries.
M4 Data's Huntingdon agrees: 'Back-up is going to remain a key requirement for all systems in the future - not just to protect against disk failure, but also against accidental data erasure and attack from viruses. Put simply, tape is, and will continue to be, far and away the best and lowest cost system for backing up data.
'Beyond back-up, tape is the ideal system for very large-scale data storage and, with technologies such as San - in which tape has a key role - there are huge new applications for tape-based storage just around the corner.'
Georgis of Exabyte believes that the immature nature of the storage media that threaten tape are precisely what will ensure that tape continues to be an integral part of the storage hierarchy - specifically for capacity, performance and cost reasons. He concedes, however, that tape's role may change as IT moves closer to the broader network architectures such as San.
'Disk-based storage will continue to be the primary storage medium in the data centre, along with some back-up functions. In this light, tape will likely become more of an archival-type storage technology. Tape's removeability, long shelf life and continuing low cost per gigabyte will make it the preferred media for archiving applications.
'I expect that someday, there will be a breakthrough in optical storage or in other recording technologies. There are still people working on holographic cubes for storage applications, but none of those technologies are close to being competitive or even real for the foreseeable future.
I don't think that anyone can accurately predict what will replace tape or when,' Georgis says.
StorageTek's Taylor sums up: 'The flexibility and cost-effectiveness of tape storage still make it one of the best competitive advantages an enterprise can have. Clearly, tape-based data storage is not dead, but only beginning to realise its true potential.'
So despite the debate over tape's future and the last rites being read by the optical camp, it looks as if tape is hanging on by more than just the skin of its teeth.
WHAT THE FUTURE HOLDS FOR TAPE
PC Dealer - Realistically, is there a long-term future for tape technology?
Tony Rush, worldwide marketing communications manager for tape storage at Hewlett Packard - Yes. Tape offers an unbeatable combination of two key features - it's low cost and it's removable. Together, these two features are not matched by any other technology.
PC Dealer - How long term is that future likely to be?
Rush - When investigating our investment in Linear Tape-Open (LTO), we looked forward between five and 10 years to see if tape was still a viable technology to invest in. We could not see a strong competitor for tape within that timeframe.
One may appear after that period, but we have customers with real needs today.
PC Dealer - What tape technologies are on the horizon to extend its lifespan?
Rush - The most significant new tape technology is LTO, and in particular the Ultrium format of LTO.
We have worked with IBM and Seagate on Ultrium to design a format that incorporates the sum total of our three companies' knowledge of magnetic recording. The format will deliver 100Gb per cartridge - typically 200Gb with data compression - in its first generation. We have defined and published a four-generation roadmap for Ultrium that doubles the capacity and transfer rate with each generation to arrive at 1.6Tb per cartridge by generation four.