Two thousand and six is almost upon us and I have been asked to break out my crystal ball and predict a few things about storage in the new year. This is usually a challenge, but not this time around.
Prediction 1: Stand by for more privacy legislation
First, I expect 2006 to produce federal legislation extending the provisions of Graham Leach Bliley, which became law back in 1999, to include some of the more painful provisions of state laws, like California's SB1386. In addition to making a blanket mea culpa press release on Business Wire as required by GLBA, the new law will require companies that accidentally release private information to make personal contact with each victim to let them know that bad things happened to their information and that the company is very, very sorry.
There is a lot of momentum behind the six bills already in the legislature and scary statistics (not to mention an almost daily litany of disclosure events) are bolstering the case. In October, The Ponemon Institute stated that about 11% of Americans indicated they had received notice from a firm revealing their personal information had been lost in the last 12 months. About 40% of them almost missed the notice, the researchers claimed, because they believed it was either junk mail or a telemarketing call. Sixty percent of those surveyed said they had "discontinued" their relationship with the company involved or were thinking about doing so.
The analyst bean counters
Leading to prediction 2: More tape is dead foo
Considering the strain that encryption will place on already stressed backup processes (and especially restore, where it introduces about 40% more inefficiency into the write-back operation). We can expect the "tape is dead" crowd to become even more vociferous -- pushing disk-as-tape-replacement as though it is the only solution to a new regulatory mandate. The content addressable storage (CAS) system vendors have been hitching their lock-in products to Sarbanes-Oxley for the last year, and will now try to extend their messaging to include GLBA II, or whatever it is ultimately called.
Look for cooler heads to prevail by end of year when folks start to realize that the real solution to encrypted tape is not more technology, but better data management. I am already seeing smart companies beginning to segregate their data into logical classes for better management. Only after this is done can encryption be reasonably and judiciously applied to those data classes that actually need it.
Prediction 3: Stovepipe CAS is dead
While we are on the subject of CAS, I expect to see an announcement close to the beginning of the year of a hardware-agnostic CAS software product that will undo the value proposition of the hardware products shamelessly described as "sticky" by vendors. If you are considering a hardware CAS today, don't buy it. You deserve a product for storing and tracking data for a long time that does not require you and three generations of your children to pick up the tab. I know of at least one software-based product that works with any hardware you own that is getting ready to ship. There will doubtless be others.
Zetera's MicroSAN, branded by NetGear as StorageCentral, continues to charm my socks off. We have been beating our boxes to death in our labs and can't get them to break. With their technology, drives are connected directly to an IP net and an IP standard service, and multicasting is used to stripe across nodes and create an array. The current product aims at the home user and provides a TB or so of REAL network storage for $129 plus the cost of whatever drives you decide to buy.
In the coming year, expect Zetera technology on disk arrays aimed at the SMBs. This may also trigger companies like LeftHand Networks and Adaptec to blow the dust off of their UDP-based block storage protocols in order to deliver lower cost alternatives to companies that don't have the deep pockets of the Fortune 500.
Prediction 5: SAS/SATA II woes in the offing
Speaking of lower price, I do not expect SAS/SATA II technology roll-outs to go smoothly. Individual products we are seeing in our labs from Promise Technologies and others are top notch. But, interoperability problems may begin cropping up because of (1) the balkanization of the SATA II standard, and (2) the efforts of different elements within the storage industry to kill SAS in its cradle.
I had a great chat with a fellow at SiliconStor the other day who tells me that his company's Interposer chip enables you to get the same benefits from a SATA array that SAS promises (turns single-ported disk into dual-ported disk for all intents and purposes) and similar results are possible from dual-backplane board designs from AMCC, I'm told. Both products are said to be cheaper, despite the doubling of disk drives, than a SAS array.
Meanwhile, the FC guys are not terribly happy with rival SAS, despite the carefully coded messages from SAS manufacturers that they do not intend to go head to head with their FC brethren. Both give you the same value inside the array, but SAS is cheaper. What no one is saying is that SAS could be retrofitted into a very FC-comparable fabric outside the box and without a lot of work. Trouble for FC SANs? Maybe.
Also, iSCSI is trying to get a toe-hold, but that will require vendors to stop setting all their sites on the large enterprise. Remember when all you needed for iSCSI was an $80 per port GigE switch and a NIC? We seem to have strayed from that configuration quite a bit. If iSCSI has any hopes of becoming SAN for the rest of us (though I fail to see why the rest of us need a SAN), they'd better get back to their roots.
Well, that's my take for now.
About the author: Jon William Toigo is a managing partner for Toigo Productions. Jon has over 20 years of experience in IT and storage.