Sony sticks to its AIT knitting
Can servers keep up with D2D?
Sony is steaming ahead at full speed in the tape market and not deviating from its 'tape is best for backup' policy. Will Trotman, Sony's senior product manager discussed Sony's view of the backup market and its place within it. That place is characterised by growth currently and forecasted growth too.
This is against a background of the backup market being affected by its biggest challenge for over a decade; disk. Disk-to-disk (D2D) backup is encroaching into the backup market in several ways. At the low-end external USB-connected drives, such as Maxtor's One Touch, are convenient for desktop backup. Microsoft has introduced its DPM, which validated the D2D idea for small and medium enterprises (SME) with a specific server and its disks acing as a backup server. Thirdly, virtual tape libraries (VTLs) are drive arrays looking like a tape library to backup software and, like all these D2D products, offering disk-level access and write speeds to shorten backup windows.
A fourth front in the disk attack on the backup market has opened with removable hard drives such as the recent Imation Odyssey and ProStor's RDX line.
Low end and mid-range tape markets
According to Quantum the low end tape market is in negative growth. Yet Sony is content to stay in the low-end tape market. Trotman says it is because Sony is growing its market share with its AIT range which spans the range from 20GB (native) cartridges at the low-end to 80GB and 150GB ones at the top part of the low-end market (defined as 100GB or smaller cartridges).
Trotman reminds us that customers have upgrade rungs in the AIT product ladder which, means that a customer can start at the 20GB AIT-1 level and go rung by rung to the AIT-3Ex level with he ability to use his previous format at each upgrade point.
That is not true for DDS/DAT-72 which currently stops at DAT-72 (36GB capacity native). The market is waiting for DAT-160 (80GB native) but it seems to be late and, anyway, is the end-point of the DAT line with no upgrade path.
Quantum's DLT/VS is similarly limited in its low-end market spread as it starts at 40GB native with DLTVS80 and stops at DLTV4 (160GB native capacity) whereas AIT carries on up to 400GB capacity with AIT-5.
LTO doesn't cover the low-end space, starting at 200GB native with LTO2 half height drives. Trotman says the recent quarter-on-quarter sales trends for AIT in the low-end have been upwards. We are looking at a Sony strategy to grow its share in a declining low-end tape market.
D2D or D2D2T and disk vs tape backup speed
The mid-range tape market occupies a 150GB - 400GB cartridge capacity range and is growing, with Sony's AIT sales growing with it. Disk-based backup product incursions are less prominent here.
All in all AIT has a 14-15 percent share of the European tape market and is growing. Quantum's DLTVS has a similar share but its trend has been one of falling share.
Sony's view of D2D is that misses the point that disk doesn't have the unique low-cost, long-term, high capacity and reliability of tape-based backup strategies. You need tape as the archive of last resort as it were; D2D to tape (D2D2T) is the best course to take which combines the speed of disk with the longevity and cost advantages of tape.
Another point that Trotman makes is that although disk offers potentially a faster backup process the actuality may be different. Most customers in the low-end and mid-range space don't have top-notch servers. It's quite likely that the average SME server can't keep up with a disk, with waits to refill a data buffer, whereas it can happily stream data to tape without stopping. In this situation we might view it as the backup tape tortoise arriving at the completion point before the backup disk hare because it has not had to stop and wait en route.
A notable aspect of the AIT-5 introduction as that the product had the same transfer rate as AIT-4, at 24MB/sec. Sony found that doubling its transfer rate was not needed by its customers whose servers wouldn't be able to take advantage of a 48MB/sec transfer speed. That must surely be true of mid-range tape users in general.
Trotman observed in passing that many LTO users don't get the full, rated LTO speed because their servers can't keep up.
Removable hard drives, according to Trotman, are unlikely to have the reliability and longevity of tape. They are more likely to be used for traditional grandfather/father/son backup cycles than for archiving data for the longer term, which is where tape shines.
In the automation sphere tape has its, in Sony's case, StorStation autoloaders and small libraries. There is no disk equivalent, saving Iomega's, capacity-limited, REV autoloader. Sony automation products comfortably beat its capacity points.
Beyond current AIT and Super AIT
At capacity points above AIT LTO is on a roll and seemingly all-conquering, witness Quantum's buying-in to the LTO community with its Certance acquisition. Sony's own super-tape format, S-AIT, is now prominent in the broadcast vertical market and features in Sony's PetaSite libraries for this market.
There is a further AIT generation to come; AIT-6 with its 800GB native capacity. The format could go further. Trotman says that, for example, the read/write head technology currently in use can be extended to possible 1600GB and 3200GB capacities, the potential AIT-7 and AIT-8 product points.
There is no plan to introduce these extended AIT formats yet. A large part of the consideration about deciding whether or not to introduce extended AIT formats is asking if there is a viable business model. Would customers in the installed base of AIT-5 and future AIT-6 users be likely to upgrade to a future AIT-7 product? If that business is viable then an AIT-7 drive and format might well arrive.
Sony expects that customers in the large AIT-3 installed base may well upgrade direct to AIT-5, as well as AIT-4 customers upgrading to 5 as well. There are well over 700,000 AIT drives installed world-wide. It's probable that AIT-5's arrival will grow that to and past 800,000 installed drives, possibly on to 900,000. It's from this pool that potential AIT-7 customers would have to come.
Obviously the upgrade pool needs to be of a certain size and upgrade rate for the AIT-7 business case to be strong enough to cause product commitment. That size and rate weren't revealed or discussed.
We might see a go-faster/hold-more AIT-4 product in the AIT-2 Turbo/AIR-3Ex mould but Trotman said nothing about that.
Currently Sony is considering whether to add new interfaces to AIT drives. Candidate ones are external SATA (eSATA) and SAS. He said that not many server vendors are adding eSATA ports although most are using SATA now. SAS seems expensive for the low end; it's more of a mid-range interconnect. Sony doesn't know whether SAS will replace SCSI in the mi-range or not.
As with eSATA it has to wait for the server vendors to jump. But Sony will probably not provide both SAS and SATA for its AIT drives, preferring to use one or the other. Low-end AIT has SATA already. There is an LTO SAS option and Quantum has both SATA and eSATA interfaces.
Sony thinks that there is a lot of life in the tape market yet. It can continue to grow its share in the sub-150GB market through its more complete range and better upgrade story than DDS/DAT, DLTVS, or LTO. It will continue to grow in the mid-range space as the market grows too. AIT's future beyond AIT-6 depends upon the size of the pool of customers likely to upgrade to a potential AIT-7. That sizing is currently unknown.
AIT has delivered its five generations pretty much as regular as clockwork, one every two years. Satisfied AIT users, of whom there must be many hundreds of thousands, will no doubt wish that this regularity continues. Much will depend upon how they and the market generally responds to the continuing encroachment of disk-based backup and archive products.
Contact Sony Storage by e-mail to: firstname.lastname@example.org