SSD

Size matters – in more ways than you think

HP-laptop-LP-laptop-lgBigger is better in the disk drive world.  But you might be surprised at how much better.

Robin Harris wrote a great piece on why high capacity drives offer much more than just more space for your stuff.  The bigger the drive, the faster it works.  Significantly faster.

So much so that Robin believes a bigger hard drive is the best upgrade besides RAM you can give your PC.

But not as fast as SSD, right?  Wrong.

According to Robin, A laptop PC with a Seagate Momentus 500GB drive will work faster with large files than an Intel X-25M G2 SSD.

That’s more speed and six times the capacity for a drive that costs less.

Sometimes it pays to read the fine print.

That’s a faster drive with 6 times the capacity for less money.

Sometimes it pays to read the fine print.

How SSD will emerge from the Trough of Disillusionment

The Gartner Hype CycleThe SSD angst continues around STEC’s recent announcement, EMC’s reduced demand forecast and the enterprise SSD market in general.  Beth Parisseau interviewed Jeff Boles of Taneja Group on this.  He sees a lot of hesitation from customers considering SSD product in their arrays.  What’s the solution?  He’s on the right track:

“What the market needs is a good round of commodization.”

Customers are not buying enterprise SSD in droves yet because the uncertainties and risks outweigh the benefits: 

  • Will this product/controller/architecture be supported next year? 
  • What’s the long term roadmap for this product line? 
  • Will this vendor be around to support me in two years?
  • What’s a real-world annualized failure rate for this product? 
  • How does this product stack up with alternatives?
  • Can I count on the same performance from this device/system as time goes by?

You might ask why more people haven’t predicted these somewhat obvious questions.  I postulate that they have been relying on the maturity and consistency of hard drives for so long that it was easy to take such things for granted. 

The good news is that the industry has not forgotten about these needs.  Seagate for one has been investing for years in transforming Solid State Storage (SSS) technology into enterprise-ready SSD products, working with other vendors and associations to develop the necessary standards, as well as product capabilities and specifications that are predictable and dependable. 

Seagate is well-qualified for this role.  They have been making enterprise-class storage devices for decades.  The real magic of hard drives aren’t the spinning media, but the dependable ‘total package’ that works seamlessly within dynamic storage infrastructures.

That’s why Seagate’s entry into the SSD market last month was viewed by some as a much-needed step on the way to a viable enterprise SSD market, one that mainstream customers are ready to invest in.

The death of the hard drive has been greatly exaggerated

twain3SSD is all the rage in the storage industry.  Very exciting technology for sure and just now going Prime Time in the enterprise. But the rise of SSDs does not correlate with a fall for hard drives.  

In fact, it looks like hard drives have quite a bit of life left in them. 

A new study by Dr. Mark Kryder and Chang Soo Kim at Carnegie Mellon on the state of storage technology in 2020 yields some surprising findings:

  • A 2.5″ disk drive will likely store 14 TB for about $40
  • Hard drives look to remain considerably less expensive than any competing technology
  • Flash memory will be the next best technology, but will be battling technology limits at about that time
  • Two other technologies to watch: phase change random access memory (PCRAM) and spin transfer torque random access memory (STTRAM)

The question to ask is not “Which technologies will replace hard drives?”, but “Which technologies will complement hard drives?”

What are your thoughts?  Agree or disagree?

The painful path to SSD adoption in netbooks

netbookBirth is painful. 

This universal truth is proving to be true for the nascent SSD netbook market.

Flash manufacturers have been struggling with losses due to depressed prices for a while now.

Prices are up!  Good news, right?  Not really, as Ars Technica reports. Netbooks are incredibly price sensitive.  At the same time, disk drives remain a popular storage option as netbooks experience feature creep.

Someday this may be a market that brings lots of joy and profits to those involved.  But it will take some strenuous effort and time up front to come to fruition.

Photo source: newverhost.com

SSD and deduplication: turbocharger and trash compactor

272499_datadomain_logoIn StorageMojo’s analysis of the EMC bid to take Data Domain from NetApp, Robin Harris quoted Chuck Hollis of EMC on why the deal makes sense:

From a storage perspective, the real action is at both ends of the storage media spectrum: making storage capacity go really fast (think enterprise flash drives) – and making storage capacity really cheap (think data deduplication, spin down, etc.).

SSD and deduplication exemplify all that’s valuable in storage these days. IT has to do more with less;  SSD both accomplish this, but in entirely different ways. 

SSD is all about leverage.  A little flash turbocharges a much larger disk-based storage investment.

SSD overcomes the long-standing imbalance between capacity growth and I/O speed on disk drives. Capacity has grown a million-fold over a few decades.  I/O speed: not so much.   The rise of SSD is a repeat of what happened when linear-access tape was replaced by random-access disk as King of the Storage Media a couple of decades back.    

Deduplication has emerged as the most efficient and implementable data compression advance in a decade.  It’s the trash compactor of data, but not just for trashy data; it’s a equal opportunity opportunity for companies to greatly reduce their data storage without throwing away data.  (IT hates to throw away data.)

It’s no surprise to see Data Domain in play.  And the SSD story has only just begun! Just wait till more mature, enterprise-ready SSD devices hit the market later this year.

SSD and Amara’s Law

amaraI worked for StorageTek at the turn of the millenium, during the Tape Wars.  EMC was saying “Tape is dead.” StorageTek claimed otherwise, and their steady tape business pretty much disproved it.  But everyone knew that disk was the future. The question was when. 

Fast forward ten years.  Tape is still not dead,  but its role in the industry has changed in unexpected ways. Content growth has transformed the storage media landscape.  There’s more to store for all media types, enabled largely by affordable, ubiquitous disk.

SSD is now talked about as the future of storage.  But when is that future and what will it look like?

Amara’s Law states that “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”  SSD, like disk storage before it, will follow Roy Amara’s law to the letter.

SSD is going to transform storage device technology far beyond simply swapping out disk drives for solid state devices at the top of the storage media chain.  SSD changes the relationship between processors, memory and storage.  It will redefine the disk drive and it’s role and expand the market for storage.

Eventually.

Meanwhile, disk drives continue to drive storage solutions more than any other storage medium.  SSDs are gradually approaching prime time.  It’s hard, slow work that involves incremental product improvement, standards development and industry coordination and cooperation.  It also will include a few steps backward in between the bigger steps forward. 

Give SSD time to fulfill its destiny.

 Photo courtesy of boingboing.net

Is SSD ready? 3PAR says no

3parDave Raffo says 3PAR didn’t add SSD in their latest product refresh because they didn’t need it – at least not yet, not in its current form. 

In the same article, Evaluator Group’s Russ Fellows says SSD is still over-priced for mid-range systems anyway. 

I’ve heard from someone involved with another storage vendor that their recent SSD addition was launched primarily to satisfy their customers’ infatuation with SDD.  This vendor doesn’t yet see enough performance improvement to justify the added cost.

The SSD dynamic today reminds me a bit of those “ask your doctor” commercials on TV: customers know how to ask for SSD, but not necessarily whether or not it’s right for them.

I’ve also heard of a vendor struggling with an SSD supplier that doesn’t get what true “enterprise-class” storage devices are all about – especially what it takes to fully test and support them in real-world applications.

Seagate is treating SSD as an emerging class of enterprise-class storage devices, with the commensurate development, testing and systems compatibility work that goes with that mantle. 

When SSD is ready for prime time, count on Seagate to have an enterprise-ready product to offer.

$24,000 in SSDs in a single PC

Samsung has some fun with 24 SSDs in a single PC.  A gamer’s dream!  If they’re $1,000 per drive, it’s a rich gamer, but fun nontheless.

This demo hints at the game changing nature of SSD once price per GB comes down over the next few years.  Expect the first major inroads in the high-end server market where an SSD can take the place of multiple ’abused’ hard drives, as Objective Analysis calls short-stroked drives. 

In actuality, desktop PCs are the last place we’re likely to see them for good reason.

Microsoft’s low-power data center ideas

Chris Pirillo caught an intriguing concept demonstration from Microsoft’s Data Center Futures team on video. 

pirilloThe Atom processor-based servers are less than $100 not counting storage, and run with very little power.  Moreover, the servers can be put into an almost powerless sleep mode, unlike conventional servers that use 70% of their peak power when idle.

The concept is that many very low power/cost servers are more manageable and power efficient than fewer meatier servers. 

How does this jive with server virtualization, which is based on fewer, meatier physical servers?

SSDs still have a power problem

Also notable: the SSDs in their experiment are not only more expensive than the disk drives, but consume more power.  Both specs will change over time in SSD’s favor.  But it’s a reminder that SSDs aren’t magic pixie dust – they need time to mature just like every other disruptive technology.

Just-in-time SSD

Seagate’s new CEO Steve Luczo talked with InfoWorld about reducing data center power, storage and the economy, small businesses and cloud steve-luczostorage, and slower-than expected market adoption of SSD. 

SSD will be accepted by the market, no doubt!  And Seagate will be part of that market. But it looks as though those predicting the disk drive’s rapid demise at the hands of SSD were mistaken.  Not that surprising, actually; the tape-to-disk transition began decades ago, and is still underway. 

With the help of increasing powerful data management across media types, SSD, disk and tape will co-exist for some time

Watch for SSD products from Seagate when the time is right.