Today the value of Bitcoin and other digital currencies is more volatile than Americans are used to. But with Amazon and eBay looking at accepting them, that could change.

If that seems unlikely, recall that much of what you use now as "money" is simply electronic transfers: credit cards; debit cards; PayPal. Now, grasshopper, are those more "real" than Bitcoin?

The bad old days

Digital currencies are like how the US currency system used to work, before we had a national dollar.

Local banks issued its own currency supported - in theory - by the deposits of customers. To redeem the currency you'd go to the bank and exchange the notes for coin. When business crashed people would "run" to the bank to exchange their notes for specie - gold and silver coins.

Since banks borrow short term and loan long term, they would often run out of coinage and close - often costing depositors their life savings - which was Very Bad for business. That's why the US has a national currency, a Federal Reserve Bank and insures bank deposits (FDIC).

Digital currencies have none of these protections. But maybe that won't matter if the utility of them is greater than the fear that they'll become worthless.

How would that work?

"Gold bugs" advocate going back on the gold standard rather than letting the dollar "float" against other currencies. After all, advocates contend, without gold the dollar isn't backed by anything at all.

And yet the dollar remains the worldwide currency of choice, not only for B2B but as a store of value and convertibility as hard cash. Proof: most US currency circulates outside the US - Americans prefer credit cards.

Since the US dollar isn't backed by gold, and since the Fed can loan as much money as it wants to banks that can use it as reserves against loans - if only they were making loans! - why do we ascribe value to the dollar? Global acceptance and ready convertibility are two major reasons.

Which is where the value proposition for digital currencies makes the most sense. So can a digital currency replace - or at least supplement - national currencies? Yes.

The Storage Bits take

That isn't much different from what we used not so long ago - or what we use today. Digital currency is the new frontier in more ways than one: the 19th century dressed up in 21st century tech.

I hope to have more to say about it because I'm attending a Bitcoin conference that starts today. If you see me, say hi!

Comments welcome, as always.  Would you use a virtual currency? Would you put your life savings into it?

Note: Another version of this post appeared earlier on StorageMojo.

Not for me. I was happy with the performance of the Mac Pro, although I did upgrade it. I craved more flexibility, not performance, which is why geeks are mostly surprised by the popularity of tablets — and I'm not.

Case in point: Video editing. The reason I had the big Mac Pro was to edit video, although its stability — an Intel workstation motherboard, ECC RAM, ample cooling — haven't been equalled by any newer Mac I've used.

While I hire a professional for video editing now, I still do some at home. And I really can't tell the difference between editing on the Mac Pro and editing on the MacBook Air.

That's progress.

Specsmanship

The Mac Pro had 2 dual-core 2.66GHz Xeon processors, an upgraded video card, a 10K RPM WD velociraptor disk, a two-disk RAID, and 10GB of ECC RAM. The new MacBook Air has a dual-core i7 processor running at 2GHz, Intel's HD 4000 graphics, a 500GB SSD, and an external four-drive Thunderbolt array.

The MacBook Air also has a 2,560x1,440 Thunderbolt display and USB 3.0 ports. And when it's time to travel, nothing beats slipping the MacBook Air into a briefcase — no syncing!

Geekbench scores

The Mac Pro with four cores achieved a Geekbench score of 5,873 in 64-bit mode. The dual core i7 MacBook Air achieved 7,644, a 30 percent increase over it it's hulking predecessor.

While the MacBook Air is only dual core, Intel's Hyperthreading gives it two more virtual cores. Looking at my system's test results demonstrates that for many benchmarks it really works, giving the system the performance of more than two physical cores.

Other MacBook Air features play a role. The system bus speed is significantly higher, the L1 cache — which the MacPro didn't have — is infinitely larger, the L2 cache is the same size, but it's on-chip. Main memory is two times faster, the standard VRAM is larger, and the total system bandwidth is significantly greater.

I have to make do with 8GB of RAM on the MacBook Air. But the usually — though not always — snappy 500GB SSD reduces paging overhead, so the lesser capacity is rarely noticeable.

If you compare the size of the motherboards on the two systems, well, there is no comparison. The MacBook Air motherboard is tiny,smaller than an iPhone. There's a lot less fan noise as well.

All this in a system that weighs 8 percent of the original. And costs less.

The Storage Bits take

It took me a while to believe that my tiny MacBook Air could actually be faster than my workhorse Mac Pro. Was my memory playing tricks on me?

But when I sat down and compared specs, it became clear that in the last six years, the pace of technology has given me my Mac Pro in a slim, stylish, 3-pound case with a display, the keyboard, and a very functional trackpad to boot.

Our perception of what we need in a computer changes much more slowly than the actual technology. That's why tablets are popular: We're able to put the technology that people need and want into lightweight, portable, and functional systems. They have more power, ease of use, and battery life than most notebooks did 10 years ago. No wonder people love 'em!

While desktops and notebooks aren't going away, the tablet will be the leading platform for the next eight to 10 years. Perhaps in 2018, they'll even be fast enough for me.

Comments welcome. Have you switched to a smaller but faster system lately? Fun fact (for me, anyway): my #1 YouTube video has over 320,000 views.

The big issue is that big data is a couple of orders of magnitude greater scale than anything we've dealt with before. Add to that fact that we are dealing with streaming data — data that is coming in real time — that we intend to act on. This is not your father's data mining application.

Big data is often looking at what is trending. Whether it is flu or the latest on Taylor Swift, streaming data tells us where we are going, not where we've been.

It is the predictive aspect of big data that calls for actual science — the making and testing of hypotheses — so we can understand which trends are meaningful and which are spurious. If that nut can be cracked then data scientists will have earned their titles and their pay.

Check out the debate for more from both Andrew and I.

The Storage Bits take

Scaling always breaks something. Maybe not right away, but scale is consistently one of the toughest problems in computer science, as well as life.

The advent of massive storage — driven by the vastness of the internet's content — has taken our ability to store and manipulate data beyond our current ability to analyze it for actionable information. We have a lot to learn and much to gain if we can master the information that our technology now allows us to gather.

Comments welcome, of course. My guess is that big data is about where computers were in 1960. Agree?

The Electronic Frontier Foundation (EFF) published its third annual report (pdf) on online privacy and transparency on government access to your data. The EFF looked at 18 online companies and their policies across six criteria:

• Require a warrant for content of communications

• Tell users about government data requests

• Publish transparency reports

• Publish law enforcement guidelines

• Fight for users' privacy rights in courts

• Fight for users' privacy in Congress.

Not everyone has to defend a user before a judge, but the EFF wanted to acknowledge those that do. Requiring a warrant is a new category, but warrants aren't a legal requirement in much of America.

Let's go to the EFF's summary graphic:

The Storage Bits take

Human beings are imperfect, especially when it comes to wielding power over others. That's why constitutional protections are important.

Major online service providers can afford to, if they wish, go toe-to-toe with the government to protect their users from abuse of governmental power. Individuals — as the case of Aaron Swartz reminds us — don't have the resources of corporate "persons" to defend against governmental overreach.

Let's hear it for companies like Twitter, which actively work to protect us. And let's encourage other tech powerhouses, like Amazon and Apple, to support their customers. It's good for their business and good for America.

Comments welcome, of course.

NAS

The OneBlox system supports file data accessed through either NFS or SMB. Each OneBlox 2U appliance has 8 drive bays for a current max capacity of 32TB.

Thanks to cloud-based management, a naïve user can get a new share up in less than five minutes. Adding new nodes is even simpler: they auto-discover each other when they are on the same subnet.

It will take longer to unpack the box than to start using the storage.

Cool technology

While it looks like standard NAS file storage under the covers, it is a much more modern design. The system breaks files into hashed 32K objects to ensure data integrity.

The index that keeps track of these objects is kept on a fast SSD. The objects are triple replicated across other nodes to ensure data availability.

The system uses a consistent hashing algorithm to minimize internode communication, a problem that limits the scalability of some older clustered storage designs. The interconnect is gigabit Ethernet with trunking of up to four links.

All objects are AES 256 encrypted for security at the rest of and across the network. The OneSystem cloud management is designed for non-techies and makes it simple to configure shares or set up regular snapshot copies.

Money

But all the technical goodness in the world won't attract customers unless it is fairly priced. Unlike enterprise clustered storage, which typically goes for $3-$5 per gigabyte, the Exablox products are $.30-$.50 per gigabyte.

This is the kind of pricing needed to compete with Amazon Web Services and other cloud-based storage. Why go to the cloud when local is cheaper and faster?

The Storage Bits take

Traditional RAID systems — where all disks are the same size and rebuilds take forever — are dead. There are better options and Exablox looks to have one of them. 

Storage is so important that it needs to be as easy and reliable as a book on a shelf. Exablox is a step in that direction. Let's hope the reality (they ship in June) matches the promise.

Comments welcome, of course.  

But as I discovered on the show floor, Intel's list was not complete. Here's some more new Thunderbolt kit, fresh from the 2013 NAB Show:

• Blackmagic Design showed their new Production Camera 4k that, using Thunderbolt, can monitor video waveforms on a notebook. At only $4,000, that will get a lot of attention, assuming they can fix their delivery problems

• One Stop Systems showed their family of well-priced Thunderbolt PCIe expansion enclosures with room for up to 8 PCIe cards. Want to edit 4k video on a notebook? This is your box. One Stop has been focussed on PCIe expansion, but noted that mobile pros prefer compact Thunderbolt to clunky multi-lane PCIe connectors

• Ciphertex showed their range of encrypted portable RAID arrays with Thunderbolt interfaces. Great for protecting digital assets on the go. Not yet on their website, though

• Corning showed — but did not announce — optical Thunderbolt and USB 3.0 cables in lengths up to 50M/165ft. Optical cables are thinner, lighter and have excellent noise immunity. Sumitomo Electric also showed shorter optical Thunderbolt cables.

The Storage Bits take 

Thunderbolt is — and will remain — a tool for professionals who need its performance and reliability, not a mass market USB-killer. With the mass market migrating to tablets and phones, though, expect Thunderbolt to become a more attractive option to PC vendors looking to differentiate their products.

Intel's faster and lower-cost Thunderbolt options — due next year — will help adoption. USB 3.0 is great for casual users, but as I was reminded yet again this morning, it is not nearly as robust as Thunderbolt, or even FireWire 800.

There's a new Thunderbolt technology community website that has a fairly up-to-date product listing.

Comments are welcome, of course.  

That means the cross-sectional bandwidth, the total theoretical bandwidth with each channel operating at full capacity, is 40Gbps. No bits move at 40Gbps, but if all four channels were fully loaded, 40Gbits of data would move.

20Gbps Thunderbolt

It wasn't easy to pin the Intel guy down about the new controllers, but after repeated questions I finally got the story straight. The key differences are:

• Dedicated DisplayPort channel goes away: DisplayPort is still supported, but if you aren't using it — say you're connecting to a very fast storage device instead — the entire bandwidth of the Thunderbolt link is available for that one application.

• Still 40Gbps bandwidth: Since there is only one channel (shareable across DisplayPort and PCIe protocols) the number of bits across the cable — the maximum cross-sectional bandwidth in other words — remains at 40Gbps.

• Faster: But those 40Gbits will move at the higher signaling rate of 20Gbps.

The Storage Bits take

Leave it to Intel marketing to fumble good news. The reality is that the 20Gbps Falcon Ridge Thunderbolt controllers due next year will offer more usable bandwidth than the current parts.

Why? Because the DisplayPort's bi-directional bandwidth is never used in practice. Displays only receive data and send nothing back. A fully loaded Thunderbolt cable today will max out at 30Gbps, despite the theoretical 40Gbps cross-sectional bandwidth.

But a single high-performance SSD array could max out next year's Thunderbolt in both directions concurrently, actually moving 40Gbps — 20 in and 20 out — on the cable. And, of course, the maximum data rate is doubled: a 20GB file would transfer in 10 seconds rather than today's 20 seconds.

Bottom line: the 20Gbps Thunderbolt will give creative professionals more of what they want — speed — if they need it. Launching Thunderbolt into the headwinds of a global PC market meltdown is challenging, but will help differentiate pro products going forward.

Comments welcome, of course. I'll have more on Thunderbolt at NAB '13 next week.

The culprit  

What is killing the film market is not that more directors are shooting digitally. It's the fact that movie theaters are converting from film to digital projection.

While a major Hollywood film might shoot a few hundred thousand feet of film stock, the real payoff for film vendors is when several thousand copies are made for physical distribution to movie theaters around the world.

Fujifilm will continue to make some recording film for uses such as long-term archiving, but they have discontinued their color positive film, color negative film, black-and-white positive and negative film, intermediate film, sound recording film, high contrast panchromatic films and, in Japan only, the chemicals needed to produce and develop these films.

What about cameras?  

Film cameras are also going out of production, but this is a much less serious problem. Film cameras are built like tanks and last for decades. 

The most costly item in pro shops is neither camera or film, but lenses. They don't wear out and can be adapted to digital, so the cost of moving to digital is lower than you might think.

The Storage Bits take  

It may be that someone will stockpile the manufacturing equipment and chemicals required to make motion picture film, as they have for Polaroid film. That equipment will be very cheap, and as long as the chemicals are still available, probably not that difficult to produce.

But we are seeing the end of an era: The first modern storage medium to become obsolete. Digital cameras will only continue to improve, and there will be less and less reason to use film going forward.

Film-loving purists will complain for decades to come, but for the rest of us, it is the story, not the technology, that counts. Digital makes 3D video much simpler than film, as well. 

Comments welcome, as always. 

Why? Because an accelerating sales decline in PCs is meeting a rising tide of disk drive production.

The bad news? Vendors can afford to do this because they're cutting back on hard drive innovation.

There are two reasons for this. First, hard drives capacities have outstripped the needs of most consumers and system builders. Second, vendors have decided it isn't worth investing heavily in the technology required for areal density increases.

Yes, hard drives will continue to get cheaper, but at a slower rate. Say goodbye to the 40 percent annual increases in areal density and drive capacity that we got used to in the first decade of the 21st century.

The Storage Bits take  

As I noted last fall,"... with PC sales declining and the threat of SSDs, drive vendors can't afford to get greedy on pricing". They got a welcome burst of profitability thanks to the floods, but that was a one time event.

Your best deals in hard drives are external hard drives (preferably with a USB 3.0 interface) rather than raw drives. I'd love to know how they compute internal transfer pricing to make external drives cheaper than internal drives, but I guess that's for finance people to know and the rest of us to wonder at.

If you have a choice, I would wait another three months before buying any 4 terabyte drives. It usually takes 3 to 6 months for a new drive capacity to get the production bugs out, yields up, and marginal costs down.

Today, the sweet spot is in the 2 to 3 TB drives. Yet 4 terabyte drives are also coming down in price as availability improves, and should hit the sub-3¢/GB level by the second half of 2013.

You can also expect to see Western Digital and Toshiba follow Seagate's lead into hybrid drives. There is substantial room for performance improvements and creative engineering to make hybrid drives higher-margin.

Comments welcome, as always. Find any good deals lately?

Back when disk drives were young, dozens of vendors fought for storage dollars. Over 220 companies manufactured disk drives in the last 50 years, but today we are down to three.

Seagate and WD have moved into the low-end subsystem market, as well. Not just internal drives, but drives in USB enclosures and four- to six-drive RAID arrays.

PC as appliance

Likewise, we no longer need a dozen PC vendors. Most people buy notebooks today while the even more appliance-like tablets continue to win converts.

We don't need dozens of different kinds of desktops. We don't even need dozens of different notebooks or all-in-one systems.

Read this

Dell's directors forecast a grim future for the PC industry

• Read more

The increasing consolidation of the industry — with HP and Lenovo taking market share away from struggling Dell — tells us what we need to know. With two CPU vendors and a few display vendors, there really aren't that many combinations possible.

As appliance PCs proliferate, we also no longer need software that can be adapted to dozens of systems. The Windows OEM model is broken. Microsoft's tablets won't be the last hardware from Redmond, Washington.

Good news

Big, deep-pocketed players bring substantial advantages to consumers. Tighter integration will mean higher quality, greater reliability, and better security. Of course, lower hardware margins will reduce Microsoft's and Google's profits, but investors will adjust.

Apple has invested billions of dollars in equipment, forward purchase contracts, and innovative small firms to ensure a large supply of high-quality products. No independent hardware vendor could compete with them, which is why they dominate the high end — $1,000+ — market.

Just think of what Microsoft and Google could afford to do, if they put their minds to it.

The Storage Bits take

Enthusiasts will keep building their own systems. Specialized requirements for gaming, video, CAD, and 3D visualiztion will drive a custom PC market for the next several decades.

But the PC market has stopped growing. That means PCs are no longer an exciting market for investors. Consolidation — such as in the hard drive industry — is the order of the day in PCs.

The good news is that we've still only scratched the surface of what is possible with smart machines. The locus of investment has shifted, but the computer industry's future is still bright.

Comments welcome, as always. Will you miss having dozens of choices in PCs, or will it simplify your life?

Michael Dell has been pushing his namesake company in the direction of enterprise and cloud data center products and services. He's bought several interesting storage companies: EqualLogic, Compellent and AppAssure, Force10 (a high-end networking company) and Perot Systems in the last 5 years.

Especially in the cloud, the market margins are thin but the volumes are huge. That's a market where Dell's operational excellence can win.

While the company's balance sheet is strong (with about $3.50 per-share of net cash), turning Dell into a growth company is going to require major changes.

These include:

Shrink the PC business. The benefits of scale aren't important to Dell in higher margin data center markets. They'll give up on the low-end consumer space to focus on more profitable notebooks for road warriors.

Make a bold move in storage. In the new world of Big Data storage is the critical piece, both in cost and capability. If Dell could harness its low-cost expertise to scale-out storage they could have a hit. They need something, since their [dumb] deal with storage giant EMC went south.

Recast the corporate culture. Dell is a classic stack 'em high, sell 'em cheap hardware company. They do not have the DNA to be an enterprise vendor today. Getting there will require the corporate equivalent of a heart, lung and liver transplant.

Goodbye Mr. Dell. Mr. Dell has irritated many shareholders with his greedy lowball offer for the company. Nor has he had great success refashioning the company since he took control in 2007. The new owners will ax Mr. Dell. Unless he's part of the winning bid.

The Storage Bits take. With over 110,000 employees and more than $50 billion in revenue, Dell is a substantial enterprise despite its lackluster results over the last 5 years. Let's hope that whoever ends up running the company gets it right.

Comments welcome, as always.  Alas, Michael Dell is no Steve Jobs.

The problem is so widespread that there is a storage company mdash; X-IO — whose products can do a factory-level drive format on in-place drives to recover from transient drive failures. As a result, X-IO guarantees their storage performance, capacity, and availability for 5 years at no extra charge.

Obviously there was, and is, a disconnect somewhere. But where?

It's in the stack

Most industry people assume that the transient failure problem lies in the hardware and software stack. Disk drives have powerful microcontrollers running hundreds of thousands of lines of code and are expected to connect to thousands of versions of operating systems and file systems.

Every system vendor maintains drive qualification groups whose sole job is to ensure that a particular version of drive firmware works reliably with that vendor's OS and I/O stack. Once qualified, the vendor will insist that all drives continue to use that particular version of the drive firmware.

Is it any wonder, the thinking goes, that this causes problems?

A different view

Yesterday, I came across an excellent blog post about false disk failures from a senior LSI technologist, Robert Ober, that had a different view. Ober is a processor and system architect at LSI and holds dozens of patents.

In relating the experience of a large internet datacenter, he wrote:

... about 40 percent of the time with SAS and about 50 percent of the time with SATA, the drive didn't actually fail. It just lost its marbles for a while. When they pull the drive out and put it into a test jig, everything is just fine. And more interesting, when they put the drive back into service, it is no more statistically likely to fail again than any other drive in the datacenter. Why?

He went on to relate his experience working on engine controllers, which is, he said:

... a very paranoid business. If something goes wrong and someone crashes, you have a lawsuit on your hands. If a controller needs a recall, that's millions of units to replace, with a multi-hundred dollar module, and hundreds of dollars in labor for each one replaced...

So we designed very carefully to handle soft errors in memory and registers. We incorporated ECC like servers use, background code checksums and scrubbing, and all sorts of proprietary techniques, including watchdogs and super-fast self-resets that could get [the controller] operational again in less than a full revolution of the engine.

Disk drives don't include such protections. But maybe they should.

The Storage Bits take

The cost of transient drive failures is huge for both drive vendors and customers. Pulling out and returning a "failed" drive is expensive in time, money, and lost compute time. I'm sure vendors don't like the hassle either.

If a vendor could reduce their transient failure rate from 50 percent to 5 percent — still way worse than engine controllers — they could save themselves and their customers millions of dollars every year. The competitive advantage would be huge.

So why don't they? Do they not know any better? Or is the task of rewriting all that code to incorporate failsafe technologies just too daunting?

If I were running Seagate or WD, I'd want to know what the engineers could do to make transient failures much more rare. And while it would be a 3-5 year slog to make it happen, I'd do it.

Storage users deserve the best vendors can do. And right now, it doesn't look like we're getting it.

Comments are welcome, as always.

Robin Harris is currently doing some work with LSI and Robert Ober, which led to Harris reading Ober's post. Harris also previously worked with the CTO of X-IO 20 years ago.

The over-$1,000 market is not a volume Windows market. Apple owns over 70 percent of the over $1,000 notebook market, and as PC sales start declining, Apple owns even more of that market.

That's why Ed's question is so astonishing. Windows 8 cannot be saved with notebooks that cost over $1,000 because there simply aren't enough of them sold, or enough of a market to "save" W8.

Winning over consumers

For Windows to remain a viable consumer platform — not only a business tool — vendors have to find a way to meet the requirements of the market they have created. And that market wants cheap hardware.

But there's hope in the form of the HP Envy X2. While the machine is slow and under-configured, it does a couple of things very well: terrific battery life and it only costs $700. If Windows has a future as a consumer platform, it will be with machines like the HP Envy.

The problem is that, today, even not too demanding consumers will likely see that the machine is slow and unresponsive. They've been spoiled by fast iPad and Android tablets.

PCs — including Macs — overshot consumer needs about 10 years ago, which explains why XP is still common. PCs are too hard to use and manage for most people — which is why tablets are so popular.

The Storage Bits take

I'm encouraged by the creativity of the industrial design in the units Ed chose to review. While none of these machines will be very successful, Moore's Law will help their successors.

However, Intel's more powerful processors may not be enough. The PC's ever-expanding volumes drove the lower prices. But as volumes decline, the ability of vendors to invest to lower costs even more will also decline.

We're looking at the long-term stagnation of the PC market. Thank goodness we have smartphones and tablets driving innovation today.

Stagnation isn't the worst fate for PCs, but the PC market going forward will be a lot different than what we've been used to for the last 30 years.

Comments are welcome, of course.

Despite billions of dollars spent on backup power batteries and generators, power failures at major datacenters are not uncommon — just ask Netflix — so this is a real issue. Given proprietary Flash Translation Layers (FTL), there's no easy way to understand SSD behavior without testing.

In Understanding the Robustness of SSDs under Power Fault (PDF), researchers Mai Zheng and Feng Qin of Ohio State and Mark Lillibridge and Joseph Tucek of HP Labs look at how power faults affect flash-based SSDs. Short answer: It's not pretty.

The research

The team developed hardware to inject power faults and software to stress devices and check post-fault consistency. These were used to check 15 different SSDs and two hard drives.

The authors looked for several types of errors, including bit corruption, shorn writes, metadata corruption, and dead (bricked) devices. Write data was configured to enable detection of these and other errors.

Three workloads — concurrent random writes, concurrent sequential writes, and single-threaded sequential writes — maximized the SSD's internal workloads. SSDs have several background tasks, such as garbage collection, running constantly to keep the SSD ready and organized.

Tested SSDs

15 different SSDs — 10 different models from five vendors — were tested. Prices ranged from 63¢/GB to $6.50/GB using both MLC and SLC flash. Two hard drives, one low end and one high end, were also tested.

Vendor names were not revealed.

Results

The good news: Of six expected failures, only five were observed; and two of the devices behaved as expected. The bad news: 13 of the devices had poor failure behavior.

Every failed device lost some amount of data or became massively corrupted under power faults.

Bit corruption hit three devices; three had shorn writes; eight had serializability errors; one device lost one third of its data; and one SSD bricked. The low-end hard drive had some unserializable writes, while the high-end drive had no power fault failures.

The two SSDs that had no failures? Both were MLC 2012 model years with a mid-range — $1.17/GB — price.

The Storage Bits take

Because it is persistent, storage is the hardest part of IT infrastructure. There are myriad ways data gets scrambled.

This paper reminds us that SSDs are very new technology, with idiosyncrasies still being engineered around. We're still five years away from the average enterprise SSD being as reliable as the average enterprise hard drive is today.

Home and small office SSD users would be wise to have a battery backup on critical servers and desktops. Notebooks, of course, already have a battery backup.

Comments welcome, as always. The paper was presented at FAST 13. Have you seen any power-related SSD problems?

Researchers Peng Li and David J Lilja of the University of Minnesota, and James Hughes and John Plocher of FutureWei Technologies reported on SATA port multiplier behavior in a poster (PDF) presented at FAST '13. They conclude that port multipliers work well when the disks are working well — but not so well when a drive fails.

Inducing disk drive failure

Their first problem was figuring out how to induce a disk drive failure that would look like a normal disk drive failure. Simply disconnecting or powering down a disk drive happens too quickly.

Their solution was to remove the cover from the disk drive while it was under load. This typically resulted in the drive's failure within 3 to 4 minutes. They tested both Seagate and Western Digital hard drives, in both enterprise and consumer versions.

The researchers tested drive failures on a system running Linux with two SATA controllers. In the first testbed, there was one drive connected to each SATA controller. In the second testbed, there was one drive connected to one SATA controller and a port multiplier with two drives on the other SATA controller.

Results

In the first setup, with no port multiplier, the failure of one drive had no impact on the other drive on the system. The test workload, the fio program, always completed.

In stark contrast, when a drive was failed on the port multiplier, the second drive on the port multiplier would also fail without completing the fio workload. This was true on both Seagate enterprise and Western Digital consumer drives.

The Storage Bits take

This research is not conclusive, and the authors hope to do more. Only a small number of drives on a single Linux platform were tested.

But it suggests that caution is in order. Using RAID software across a port multiplier array may result in an unrecoverable failure when a single drive fails.

It is possible, using advanced erasure coding or a high-end file system like Gluster, to use a large number of disk drives on port multipliers in such a way that even several failures will not compromise data integrity or availability. But this is not something the average SOHO user could implement.

Because disks are marvels of engineering and precision manufacturing, many people will have a port multiplier where no drive fails for years. But when one does, it could be brutal.

This points to a larger issue in IT: We have few independent sources of underlying technology evaluation. We are all guinea pigs.

Comments welcome, as always. Have you experienced a disk failure on a port multiplier? Please share what you learned.

Update: Below is a video of a running drive being taken apart. A rough process, but how else can you create a head crash on demand?

First up: Write Policies for Host-side Flash Caches by Leonardo Marmol, Raju Rangaswami and Ming Zhao of Florida International U., Swaminathan Sundararaman and Nisha Talagala of Fusion-io and Ricardo Koller of FIU and VMware.

Why write caching? Write-through caching is safe because all writes are committed to disk before being acknowledged. But write-through is expensive in time and I/O traffic since the server has to wait for the network storage to complete the write.

Write-back caches - where the write is cached locally before being written to disk - reduces I/O latency, improves server performance and uses network storage bandwidth more efficiently. Better write caching is needed because read caching is not the win it used to be. Thanks to large server memory read-only DRAM caches, read-only flash caching isn't a major performance booster on most servers.

If a cache access takes 50µs and a network storage access takes 2ms, then I/O rates double when cache hits go from 95 percent to 99 percent. Doubling server I/O performance by changing caching strategy is a Very Good Thing.

Options for write caching. NAND's non-volatility enables novel write-back cache strategies that preserve data integrity while improving performance. The paper explores two write-back caching strategies, ordered and journaled.

Ordered write-back is the simpler, preserving the original data block updating order when writing to the network storage. Journaled write-back allows coalescing writes in the cache, with atomic journal updates to network storage for recovery if needed.

The authors implemented a Linux test bed with the ext3 file system. They found that both strategies are major improvements over write-through caching. The journaled write-back cache is more complex, but is also usually faster than ordered cache.

The Storage Bits take Using flash only for reads mean ignoring half - or more - of the I/O problem. PCIe NAND flash is expensive compared to disk - and so are servers - so maximizing the economic benefit is a worthy goal.

Industry politics may scuttle this approach though. The network storage vendors don't want servers emitting steady streams of updates: they want bursty traffic that forces over-configuration and larger sales.

The technology is there. What about the will to use it?

Comments welcome, as always. How much would you say a PCIe flash storage card is worth if it would double your system performance?

What's new? There's a long list of improvements, including:

• 180TB raw capacity. 45 4TB will do that, but you might save money with 3TB drives today.
• Better vibration control. Drives are locked down, reducing retries and improving reliability.
• New mobo. Faster CPU and PCIe raise the performance bar.
• RAID1 boot drive option. Ensures you can boot the pod.
• Better cooling. End the heartbreak of data melt.
• Lower price. The chassis parts - not including drives - cost 1.9% less than the old version.

Lessons learned. The Backblaze blog post goes into great detail about what they've learned from prior versions. Nothing that hardware engineers don't already know, but may be news to to the DIY crowd: watch firmware versions; stick to the design; be careful about +5v power; don't mix components.

Backblaze has over 450 Storage Pods in use, and there are hundreds at other sites. They aren't designed for heavy duty database transaction processing, nor do they have much redundancy, so RAIDing or mirrroring across Pods is important.

Drive prices. The 2011 Thai floods knocked out about a quarter of the industry's production capacity, driving prices sky-high as vendors competed for the available supply. Your best budget bet is often to buy USB drives, toss the case, and use the drive, a strategy Backblaze used during the worst of the drive crisis.

The good news: I'm seeing more evidence of competition for consumer dollars and expect to see significant price drops over the next 2-3 months. On a per GB basis we are already back to where we were in October 2011, but expect to see  3¢/GB soon.

The Storage Bits take. If you need hundreds of TB of bulk storage, the Storage Pod 3.0 is the cheapest game in town. Yes, it won't be as bulletproof as an array costing 10x, but if you know how to configure them you can have highly reliable storage for a fraction of the cost of most large arrays.

Comments welcome, of course. What would you use 180TB for?

But Microsoft isn't calling it an Ultrabook, but a "laptop in tablet form." And it shares the same root name - Surface - with the Surface RT, which has been out for months and is marketed only as a tablet.

The Surface Pro can also be used as a tablet, albeit a thick, heavy tablet with lousy battery life. So, is it a tablet or an Ultrabook?

Perceptions matter A marketing truism is "perception is reality." Not in the Fox News sense of making stuff up, but in the mind of the customer: if they believe it, that is the reality the seller has to deal with.

Thus it is possible - and all too common in tech - for companies to be honest and correct and still run afoul of customer perceptions. Which is just what Microsoft did here. Their marketing messages are leading to confusion that the capacity fracas highlights.

Where Microsoft went wrong With the Surface Pro Microsoft is pioneering a new PC form factor: the tablet PC. Yes, there have been other "tablet" PCs, but this is the first that attempts to take the best of the iPad - form factor, app store, keyboardless use, vendor integrated - with a touch-optimized full Windows OS.

Here's how Microsoft can start to fix their confused marketing:

• Call the Surface Pro something that acknowledges that it is a new category of computer. UltraTablet perhaps?
• Focus the Surface brand only on iPad competitors, not UltraTablets too. Why? Think it through: in 3-5 years UltraTablets - if successful - will probably be the fastest growing part of a shrinking PC market, while Surface tablets will be the consumer platform of choice for people who don't want to manage a PC. Touch will be everywhere and the Surface brand will obscure important differences between Microsoft's touch-enabled platforms. 
• Give up your non-standard binary measurement of storage as Apple did several years ago  adopt the IEEE/SI/IEC standard meaning of gigabyte as 1 billion bytes. Consumers and your marketing people will be less confused, your apparent capacities will be larger and you'll be supporting standards.

The Storage Bits take The Surface Pro is a worthy v.1 product, much more usable than the first MacBook Air, but it will need a couple of year's advances in CPU, storage and battery technology to become an Ultrabook competitor. But if Microsoft marketing continues down its current confused path the Surface Pro may never make it to v.2.

And for all of us who benefit by robust competition, that would be a real shame. I hope that Microsoft will up its marketing game to match the fine engineering of the Surface products.

Comments welcome, of course. And please, no rants about base 2 capacity metrics being the "standard." They aren't, so move on before the problem gets worse.

Obviously, this could be used in an exploit to sabotage Macs all over the world. And that's bad.

But the real question: what the hell are the system software guys doing in Cupertino? It's past time to put down your lattes - they make fine ones at the Infinite Loop campus cafeteria - and stop embarrassing yourselves.

OS X is turning into a steaming mess.

Time Machine system rot A recent technical comment on an Apple developer mailing list documents more OS X stupidity. Thomas Templemann, author of the app Find Any File had some telling comments.

Mr. Templemann is a ". . . big fan of Time Machine, at least on the technical level." He went on to write:

Well, today, a friend showed me the horrors of Time Capsule: Saving a few 100MB of changed files to the Capsule can easily take one to two hours. Which is incredible, isn't it?

And why is that? . . .

. . . accesses to the SAME item are repeated 10-20 times. Each of them issuing a fresh network call. . . .

. . . Isn't anyone at Apple noticing that [Time Capsule] or any network-based [Time Machine] backup is so incredibly slow?"

Sure they are. They just don't care. Mac sales keep climbing anyway.

Too many network accesses aren't the only problem with Time Machine. I stopped using it years ago because if you have a lot of email TM has to break and remake thousands of symbolic links, an expensive process, after email deletions.

Time Machine slowed my Mac Pro to a crawl. TM just isn't intended for serious users.

And about HFS+ Some Mac fans pooh-pooh assertions about the low quality of HFS plus despite the fact that it's been documented in independent research. Apple did nothing. Microsoft hired the researcher and has implemented major improvements.

But you can see for yourself, if you're comfortable with the Terminal, that you can corrupt a volume and OS X's Disk Utility won't see it or repair it. Still think OS X is safe for important data?

The Storage Bits take I don't know what it will take for true believers to see that there is a serious problem with Mac system software quality. Maybe a Google search on Mac file corruption? Or worse, personal experience of massive data corruption?

What is worrying to me is that these issues get worse with each release of OS X. That says that the OS X software group needs adult supervision, much as Microsoft did when they brought in uber-engineer Dave Cutler to drive Windows 7 after the Vista fiasco.

It's time for Tim Cook to step up and make sure that whoever is responsible for system software for the Mac has the same commitment to quality and innovation that Steve imbued in the rest of Apple.

Comments welcome, of course. Still using my Mac, but I'm starting to look at Linux distro's.

Geoff Barrall, a brilliant and creative serial entrepreneur, has come up with a product that I believe will be his most successful yet: the Transporter. Connected Data is the 5th company he's founded, with NAS vendor BlueArc and Drobo as earlier startups.

The Transporter takes a page out of the Steve Jobs playbook: make something as simple as possible and no simpler. It combines file synchronization, back up and file sharing in one simple appliance that isn't much bigger than a disk drive.

How simple? Buy the product, link to your router, register it at the company website, specify what folders you want to have and who can access them, and the system automatically takes care of the rest.

The data transporter will be available in three configurations. One with no storage – which needs a 2.5" disk drive – or with a 1TB or 2TB hard drive. They'll range in price from $199 to $399.

How is it different? As a user of Drop Box, iCloud - which I find pretty useless - and SimpleNote, I have some sense of the current market. The Transporter has several advantages:

 One time fee. Buy the product and you're done.

 Privacy. All data is stored on your disks, not in the cloud. All data is encrypted for transport. Safe for lawyers and health care providers.

 Shareable. You can share data with any Transporter in the world.

For more info, check out their data sheet.

The Storage Bits take  I'm planning to buy a couple for testing, so I'll let you know how the reality measures up to the promise. But I know Geoff and he's got a good track record of delivery.

The Transporter should remind us all that we are still early in the Internet age - about where automobile transportation was in the 1930s - and we still need many improvements. The Transporter looks to be one of them.

Comments welcome, of course. How would you use a pair of Transporters?