Storage Effect

Because Disk looks more and more like Tape

Chris Evans points out the challenges of disk drive capacity outgrowing performance.  Sound familiar?  Until recently, these complaints were perennially aimed at linear-access tape technology, with random-access disk in the role of savior.

As disk now faces the same “can’t get the data out fast enough” problem, the solution is clear: SSDs.  Flash is the new access king in storage.  Or will be, because early products are struggling with reliability, write performance, and other capabilities that disk drive device and interface technology have fine-tuned over the decades. You can’t expect a breakthrough technology to hit the ground running.

Yesterday I found this Microsoft presentation in the blogosphere on the topic from 2006: flash_is_good

I guess I wasn’t original with my “Disk is the New Tape” thing.  Makes the idea all the more interesting! 

Next question:  Is there room in the flash future for SSD, Disk and  Tape?

Dave Hitz from NetApp on data retention, flash, forks

I read an interesting profile of Dave Hitz, founder and EVP of NetApp in the latest ComputerWorld magazine.  (yes, the actual magazine, made from real paper products!).  Great perspectives on the growth of information and storage.  I recommend you take a look, even electronically.

 Takeaways:

• Deleting data is more difficult for most companies than finding room to save it. A customer CIO quote: “There are two kinds of data: information deleted within a week and data kept forever.”
• Data rarely ever disappears.  Deleting an old email doesn’t protect one legally if the receiver has a copy.
• Flash and SSD will start to have an impact in the enterprise, even though many supposed storage alternatives to disk never made it.  Flash is different because it’s a classic come-from-below disrupting technology.
• Flash will also find a place displacing/expanding RAM cache.  It’s 10X disk price, but 100X random read speeds vs. disk.
• His favorite foods are those you can eat without a fork.

1 TB in 10 years is a yawner, unless price/performance is truly unique

The latest announcement on the future of storage tech comes from Nanochip, Inc: An array-based flash memory that scales to 1 TB, is cheaper and easier to manufacture than today’s flash, and is supported by Intel and others.  Exciting!

Before you get too enthralled, note that the techology is 10 years from production.   Two problems with this:

1. Flash will have replaced disk in many applications by then, but spinning disk will still be around and will be heading towards a petabyte a disk per drive by then.  This is not a cure-all device that replaces all storage technologies.
2. The technology has yet to face the scrutiny that comes with actual existence.  Reliability, cost and functionality in a real-world machine are the true tests that tend to morph such visions as they go from “apparate” to “operate”.  Ask the guys with flash products out today about this.

Keep filling the pipe with fun technology!  But don’t take your eye of the ball on your current storage solution plans for this one.

Tenfold notebook failure rates for SSDs vs. disk drives 

Engadget reports a 10X failure rate for SSDs in notebook PCs compared to hard drives.   Can you say “early adopter”? SSDs are an exciting future, but a future still for mainstream markets.

Reliability is relative, especially in terms of the technology being replaced.  The same goes for pricing - see my Flash tags posts.

Any flash early adopters out there?  What’s your experience with returns or failures, either personally or through solutions you’ve sold to someone?

Update: Yet another issue with flash as a near-term disk drive replacement: potential patent infringement issues. 

Moving from exciting technology to real-world product is not trivial 

Robin Harris posted yesterday and last week on flash - both are interesting reading.  My takeaway is that the more flash is applied, the more real-world wrinkles bubble to the surface.  That’s exactly as it should be - new technology buzz always begins with what’s possible, then moves to “OK, now how exactly will that work in my solution today?” 

Borrowing from disk drives to make flash work

Also noticed how most of the issues and workarounds are things that have already been addressed with today’s disk drives.  Just goes to show that storage devices are more than the media.  Seagate’s in a great position as they enter the flash solutions world because of their depth in experience in making storage devices out of storage media.

What do you think? 

Comments welcome!  How do you think you will use your first flash storage (beyond thumb drives)?  Have you already?

There are 3 ways to integrate flash in your customers’ solutions today. 

Option 2: The ”Prius” approach 

• Use off-the-shelf hybrid disk drives (flash + disk) to upgrade notebook PCs or create differentiated whitebook PCs
• Value: faster application boot and longer battery life
• Focus on Vista OS notebooks today, since hybrid drives rely on Vista to achieve their performance benefits

Hybrid drives are aptly named.  Just like hybrid cars, these drives straddle the fence between new (flash memory) and mature (spinning disk storage) technologies to provide the advantages of both.  They’re priced more like disk drives, and perform more like flash. 

Hybrid drives will expand beyond notebook PCs 

In the future, hybrid drives will be available for a broader set of applications, and their performance advantage will not depend on the operating system.  Today, without Vista, a hybrid drive performs just like a normal drive.

Seagate offers a hybrid drive today, the Momentus 5400 PSD.

Using flash storage today, part 1

There are 3 ways to integrate flash in your customers’ solutions today. 

 Option 1: just do it.

• Buy off-the-shelf solid state drives (SSDs) and substitute for disk drives
• Or, incorporate thumb drives in your solution or process to solve a problem (security codes or data, for example)
• Identify high-end applications where SSDs will make a real difference in user experience or performance

Flash storage means a 5X to 10X cost multiplier per gigabyte for you.  This is justifiable and even profitable if you can find applications that justify a commensurate price increase or savings in your development/support costs. 

Not a lot of market for this today, but it will grow fast.  This is a great chance to get up the learning curve early.

Let’s hear how you’re doing this today.  What’s the best current application for flash?

Yet another data point from TrendFocus on flash.  Flash has a great future, but disk drives continue to outscale and outsell them. 

Dream about flash, but rely on disk - at least for a while.

SSD is kinda cool in MacBook Air, but not worth today’s price

Ars Technica’s review of the MacBook Air provides a thorough user-level evaluation of the value of SSD in a notebook. 

The conclusion:

The $1,300 question is whether the SSD is worth the extra cash. The answer seems to be no. I experienced only moderate gains in battery life and not very noticeable speed differences. The one major benefit of the SSD model is that it doesn’t cause the same types of slowdowns as the HDD model during times of high disk activity, and that’s certainly a huge plus. Speedy read times are great, too, but they are balanced out by pokey write times.

Still, even if it’s more usable, it’s hard to justify the huge price difference for the SSD model. If you’ve got an extra $1,300 to blow and, for some reason, haven’t just bought a second computer with it, perhaps the SSD model is for you. For anyone else looking to buy an Air, the HDD model appears to provide the most bang for the buck.

You get what you pay for.  For SSDs, the ‘gets’ aren’t worth $1300 bucks.  The big surprise to me was that in some ways the SSD notebook was actually worse than the disk drive version. 

Don’t worry, the premium will drop and SSD’s will improve.  But it’s going to be years, not months, before most of us choose SSD.

Solid state drives are a winning solution - but when and where? 

Bill Watkins, CEO of Seagate, talked to X-bit labs about flash and SSDs.  Flash is exciting technology - just don’t expect it to replace tried and true disk drives in a big way any time soon. 

Despite all the noise about flash in notebook PCs, Bill sees the first true sweet spot for flash being in enterprise applications. For notebooks, much of the power, performance and reliability benefits of pure flash are either overstated or can be gained today for a much lower cost with hybrid drives. 

Hybrid drives combine flash with a disk drive and use the flash as a uber-cache, giving you the benefits of both technologies in one solution.

I’ve posted many time on this topic - here and here and here.

What say you?  Will you be plunking down the big bucks for a solid state notebook this year?