Storage Bits

July 24th, 2008

Apple’s coming notebook transition

Posted by Robin Harris @ 8:14 am

Categories: Infrastructure, Marketing

Tags: Notebook, Apple Macintosh, Apple Inc., Product Transition, Portable Growth, Desktops, Hardware, Robin Harris

Apple is readying a major product transition. We know because they told us so. In the quarterly financial crime call chief financial Officer Peter Oppenheimer referred several times to a “…future product transition which I can’t discuss today”.

That was interesting for two reasons. First, the CFO was trying to explain why they were forecasting a large drop in Apple’s gross margins for the coming quarter, from 34.8% to 31.5%. Secondly, of course because the famously secretive company does everything it can to maintain the element of surprise.

But this transition’s costs are too big to bury. So it is going to be big.

Running the numbers
Managing Wall Street means managing Wall Street’s expectations. Successful companies - companies that deliver above expectations - have healthy stock prices. So the key is to talk down one’s prospects and consistently do better.

Press reports of the call have tended to emphasize the difference between Apple’s reported gross margin of 34.8% and the forecast gross margin 31.5%. But that isn’t the right way to look at it.

The real difference is between what they forecast for the June quarter and what they are forecasting the September quarter. The gross margin forecast for the June quarter was a 33%. Thus the drop from the June quarter to the September quarter is a 150 basis points drop.

Starting with that and depending on how you forecast growth, Apple is looking at an investment in the current quarter’s Back-To-School promotion and the “product transition” of from $120 million to as much as $400 million. The average iPod goes at the door at $150 - so if they sell 1 million BTS Macs, the promotion will cost (price less GM) on the order of $100 million, probably much less since they only sold 2.5 million Macs in the June quarter.

So the product transition has a war chest of anywhere from $60 million to - on the high side - $300 million. Let’s split the difference on the conservative side and say $150 million for the transition.

What product transition could cost $150 million?
While iPod and iPhone sales are big and getting bigger, Macs are still over 60% of Apple’s revenue. The iPhone has just been refreshed and so have most of the iPods.

The Mac unit split is 60/40 ‘Books/desktops. The iMac desktop line has just been refreshed and the Mac Pro line, while the packaging is getting old, has state of the art workstation innards.

Other than the MacBook Air though, the ‘Book line hasn’t had a major design refresh in years. That is where the transition will be. It has the volume and the aging designs to require a change.

What will be in the new ‘Books
Apple doesn’t play with pricing very much. They intro a product at a set price and it generally stays there until the next refresh. They use declining commodity prices to increase their margins.

When they do a refresh they add in features, such as faster processors, larger hard drives or improved backlighting. These things are all more expensive at the beginning of the refresh, but add to the perceived value.

There are several expensive features in the Mac book line that are due for significant upgrades.

• Quad core processors. The differentiation between the Mac book pros and the vanilla MacBooks has been declining for several quarters. The easiest way to reestablish their high-end credentials is through the recently announced Intel Quad core processors.
• Power management. Intel and others have announced a slew of more power efficient chips. Not only processors but multimedia decoding chips, networking chips, graphics, DRAM and hard drives. Maybe even solid state disks. Taken singly none of these warrant a redesign, but together a significantly more power efficient notebook can be built - either lighter or with a longer battery life.
• Blu-ray optical drives. Macs in general are simply not with the Blu-ray program, which is odd since Apple and Pixar have been longtime supporters of the blue ray program. Apple codecs do not support it and there are no factory installed Blu-ray capable drives available. That will change.
• Design. The MacBook Air points to several design themes that we can expect to see in the new Mac pros. These include larger track pads for multi-touch use and beveled edges to make the system look and feel slimmer.
• Motherboard flexibility. Intel’s current chipset architectures are reaching their limits. They have a number of new initiatives in I/O, memory, and system interconnect that Apple, with its focus on high-end notebooks, will incorporate.

The Storage Bits take
Given those size of the investment, this is no ordinary product transition. Some people are forecasting a whole new product line for Apple, such as a netbook entry. But that would be out of character. Let other people establish the market. Look how long they waited to enter the smart phone market.

There is also the chance that Apple will lower prices. Portable growth has been trailing, in percentage terms, desktop growth. Since portables are more profitable than desktops Apple clearly has incentive to goose up ‘book sales.

Plus they have to stay cool.

Comments welcome, of course.

July 17th, 2008

Super flash, hyper flash

Posted by Robin Harris @ 5:44 am

Categories: Solid State Disk, Infrastructure, Marketing

Tags: Sun Microsystems Inc., Storage Pyramid, Marketing Research, Storage, Marketing, Hardware, Robin Harris

Samsung has announced a new SLC flash part with 5x the write/erase cycle spec over standard single-level cell (SLC) NAND flash. They call it server grade flash. Is this the flash promised land or marketing hype?

5x what?
The press release doesn’t mention any numbers, but most SLC flash is spec’ed at 100,000 write/erase cycles. 500,000 write/erase cycles is a significant improvement.

But is it a significant technology?

Announced with Sun
More than one source has told me that current SLC flash is capable of close to 1 million write/erase cycles. Yet no vendor has been willing, until now, to claim a higher spec.

This announcement is a marketing enhancement rather than a technical one. Here’s a quote from the press release:

The ultra-endurance server-grade memory has been developed in close cooperation with Sun over the past several months.

Months? For a technology as complex as flash? I don’t think so.

The Storage Bits take
The announcement is smart marketing. If Samsung had just increased the spec without tying it to some real or imagined customer benefit they would have missed an opportunity to differentiate their products - and increase their margins.

There may be other issues at work here too. For example the added endurance may enable designers to implement more aggressive garbage collection and cleaning algorithms to improve write performance for OLTP use.

There are a lot of moving parts in SSD engineering. Giving designers more flexibility is a Good Thing.

Sun’s involvement points to a larger trend: massive memory servers that provide very high performance without the cost and management headaches of storage arrays. The traditional storage pyramid is being re-architected in real time.

Comments welcome, of course.

July 16th, 2008

The edge-centric Internet

Posted by Robin Harris @ 6:46 pm

Categories: Infrastructure

Tags: Consumer, Internet Data Center, Data Center, Internet, Data Centers, Internet Service Providers (ISPs), Storage, Hardware, Data Management, Robin Harris

Why isn’t the edge of the network the center? More computes and storage gets shipped to consumers than to Internet data centers, so consumers should help power the Internet. And get paid for it.

P2P vs Internet data centers
A new project, spearheaded by National ICT Australia, is now part of a €50 billion program to investigate innovative technologies. The project is called Nano Data Centers (NADA) and is part of the future Internet initiative.

The goal: learn how to deliver data from the edge of the network instead of from costly, power gobbling Internet data centers.

Consumer driven to consumer powered
Google is famous for using consumer grade technology to build the world’s largest and most powerful data centers. This project asks “why don’t we use the technology that consumers are already buying to build virtual data centers?”

We already do. The Folding@home project already uses the power of PS3’s and PCs to perform computationally costly research on game consoles. P2P file sharing already spreads bandwidth among many servers.

The technology is there.

The business model
ISPs could package up unused storage and CPU cycles from their subscribers and sell them to major content and service providers. The ISP could charge a third less than what those resources would cost in a data center, keep half the money for themselves and be give the other half back to us.

At my house we currently have close to a terabyte of unused disk and 7 cores that are typically no more busy than any other PC. And that is without a PS3.

Amazon charges about $.15 a gigabyte for storage space. If my ISP sold it for $.10 per gigabyte and gave me a nickel of that, I’d be getting $50 a month just for storage. That would just about cover my broadband connection.

Is it green?
Google noted in their paper Powering The Warehouse Size Computer that provisioning power cost as much as 10 years of power. The power distribution units, the diesel generators, the power monitoring and isolation equipment, all cost a lot of money. And then there is the air-conditioning equipment.

Spread that across 150 million American homes and that power provisioning cost goes away. We’d still use the power but we wouldn’t have to build and transport all that equipment to a data center.

The Storage Bits take
America probably won’t be able to take advantage of this concept due to the slow build out of our broadband infrastructure. But the basic idea is a good one.

Why build what consumers already own?

Comments welcome, of course.

July 14th, 2008

FTP: untrustworthy file transfer

Posted by Robin Harris @ 4:20 pm

Categories: Infrastructure, Security, Software

Tags: FTP, SFTP, Robin Harris

FTP - file transfer protocol - is the most commonly used method for moving files around Web. Now Steve Frank, a founder and developer for Mac software company Panic, has come out and recommended that people stop using FTP.

I wrote about this (see If hackers don’t get you, maybe Google will) after my other blog, StorageMojo, was hacked. I’m glad to see a vendor of FTP software - I use their fine product Transmit - jump on board with a strong recommendation.

Why? Here are a couple of the best reasons he gives.

• Unless totaled over a secure socket, FTP is 100% insecure. Your password, and the contents of all your files are sent in the clear, free to be examined or captured by any network hop between you and your server. . . .
• FTP is not friendly with firewalls. Because it constantly needs to establish new connections, this has led us to “passive mode” which might as well be black magic as far as most people are concerned. Briefly, passive mode means the client initiates data connections to the server, rather than the default where the server makes connections to the client (yes, really). Worse still, data connections occur on varying high port numbers (usually 49152 - 65335) which means since Edmonds would have to open over 16,000 ports in the firewall, almost defeating the purpose of a firewall in the first place. It’s a mess, and it’s really hard to understand.

If not FTP, what?
As noted in my blog post two months ago as FTP - secure FTP - is an excellent alternative.

To quote Steve Frank again:

It’s secure, it’s consistently implemented, and its machine-readable. That all adds up to a more reliable future proof transfer client for you.

I’ve talked to a lot of people who didn’t even realize their host supported SFTP. If your hosting service supports SFTP, you usually don’t have to change anything except for switching your client protocol from FTP to as FTP period if it doesn’t work, you should ask your host if there is anything else you have to do (such as use a different port number)….

FTP has served us well but it’s time to move on. You wouldn’t use a 23-year-old computer to do your work, so don’t use a protocol from the same vintage. Demand of modern transfer protocols from your host.

Amen, brother.

The Storage Bits take
Just one thing.

Steve, if this is such a great idea - and it is - why isn’t SFTP the default FTP option in your company’s product?

You are the expert. You have the knowledge. Your users typically have no more idea how FTP works than they do about ray tracing.

As computers become more pervasive the technical expertise of the average user continues to drop. Smart vendors will make an effort to meet their customers more than half way.

Comments welcome, of course.

July 9th, 2008

Attacking phishing at the source

Posted by Robin Harris @ 8:08 am

Categories: Security, Public policy

Tags: Bank, E-mail Authentication, PayPal, eBay Inc., E-mail, Phishing, Financial Services, Online Communications, Security, Spam And Phishing

Good news on the phishing front. Google reports that they are now authenticating all e-mail that purports to come from eBay and PayPal.

Using e-mail authentication standards including DomainKeys and Domain Keys identified Mail, and working with PayPal and eBay, Google’s Gmail now verifies every e-mail that claims to come from PayPal or eBay. If it doesn’t verify you’ll never see it.

Angry eBay buyer
The first time I got one of those angry eBay buyer emails - “where’s my stuff, you thief!” - it got me going for a second. But now if you use Gmail, those e-mails won’t even make it to your trash folder. They’re just gone.

EBay and PayPal had to undertake the effort to ensure that all of their e-mails used the domain Keys and domain keys identified Mail authentication protocols. They have blazed a trail for all companies who want their customers trust.

Accountability for banks
In even better news a German court has ruled that banks are liable for phishing attacks on customers. Criminals compromised a German couple’s PC and it used their bank information to pay out €4000 on a fraudulent eBay transaction.

The court held that the payment request demonstrably did not come from the customer. “The bank bears the forgery risk of the transfer order” the judge said. Yay!

Financial institutions love getting customers to use the Internet for payments, statements and applications because they save money. But when there is a problem it is the customer who sorts it out.

The Storage Bits take
Kudos to Google, eBay and PayPal for implementing e-mail authentication. It points the way to a future where all financial institutions use authentication to ensure that e-mails bearing their names are genuine.

Kudos also to the German court for holding banker’s feet to the fire. For too long banks and other financial institutions have underspent on security, leaving the burden of enforcement to individuals and local police who are ill-equipped to handle sophisticated online thievery.

Sadly, in the United States, the U.S. Congress is a wholly owned subsidiary of the American Banking Association. Do not expect any relief for consumers here anytime soon. Do not look to the courts either: the current rage for judicial inactivism means that justice, or even simple fairness, will be denied to netizens.

While it moves us in the right direction, there is a downside to partial protection. As e-mail becomes a more trustworthy, some people will become more vulnerable to scams. But the ultimate goal is to squeeze the profits out of online crime so the criminals find something else to do.

Comments welcome, of course.

July 6th, 2008

Linux for housewives. XP for geeks.

Posted by Robin Harris @ 10:52 pm

Categories: Software, Marketing

Tags: ASUS, Computing, Microsoft Research, Linux, Microsoft Windows XP, UNIX, Operating Systems, Open Source, Software, Microsoft Windows

The computer proletariat is rising up - and computing will never be the same. Tiny, sub-$500 “netbooks” like the Asus Eee are the hottest thing going in notebooks today. And some surprising things are happening. Like housewives on Linux.

Asus is forecasting worldwide shipments of 10 million 7 to 10 inch screen netbooks this year! And a billion in 2018.

Appliance computing
In an article in the Asian business publication Tech-on reporter Tomohiro Otsuki writes:

Retailers and contract manufacturers in Taiwan say that novice PC users there, like students and housewives, tend to buy the Linux version of the Eee PC701, while geeks go for Windows XP.

Does that sound backwards?

Yet a quick look at Amazon shows that Asus Eee’s with XP roughly $35-$100 more than their Linux brethern. Housewives know a bargain when they see one.

Microsoft Research’s Gordon Bell noted that every 10 years a new form of computing emerges thanks to Moore’s Law and the declining cost/increasing performance of ICs. Looks like the netbook is this decade’s new form: a minimalist computer for Internet, email, chatting, video and light application use.

The new market leaders
A big surprise is that inventors and leaders in this new segment are the Taiwanese firms that build, for other people, most of the world’s notebooks. The contract manufacturers, who mostly assemble to spec, are enjoying the freedom to build their own products using their own sense of what the market wants.

Taiwan is the wild East. Expect some crazy experiments - and some revolutionary products.

The Storage Bits take
This is the chance Linux partisans have been waiting for - and it’s coming faster than I’d expected. Microsoft is reportedly charging $60 for netbook XP - a big chunk of a $200 computer’s cost. As Netscape discovered it is hard to compete with “free.”

If Taiwanese housewives are buying Linux the guys in Redmond need to sharpen their pencils. Housewives don’t need Office and Exchange. What do they need?

And Apple will be late to this party as they’ve got their hands full with the iPhone and Snow Leopard. Apple has a history of missing these big shifts - if they haven’t invented them - as they did with towers replacing desktops in the mid-90s.

The netbook space promises to be a lot of fun.

Comments welcome, of course. Maybe Taiwanese housewives are really really smart?

July 4th, 2008

July 4th, 2008: A trip down memory lane

Posted by Robin Harris @ 11:29 am

Categories: Security, Public policy

Tags: U.S., President, Storage, Hardware, Robin Harris

A 1 question civics quiz
Today American’s celebrate the Declaration of Independence, the legal foundation of the USA and, after the failure of the Articles of Confederation, the Constitution of the United States.

I revere the Constitution as a magnificent piece of political engineering. But it takes people who remember what it says to make it work.

Here’s a simple question: what is the President’s highest duty under the Constitution?

To protect the American people?
I’ve heard both Presidential candidates and the current officeholder claim that their number one duty is to “protect” the American people.

Baloney!

The Constitution gives the President an even higher duty.

That duty is embodied in the oath of office for President of the United States in Article II, Section 1 of the Constitution

“I do solemnly swear (or affirm) that I will faithfully execute the office of President of the United States, and will to the best of my ability, preserve, protect and defend the Constitution of the United States.”

That’s it: “. . . preserve, protect and defend the Constitution of the United States.”

The security of the United States is based on the Constitution and the system it embodies, where Government operates under the rule of law, not men.

The Storage Bits take
Storage is of little use if we don’t access it - and that includes historical documents.

I urge all to re-read the Declaration of Independence and the Constitution this 4th of July.

Then enjoy your day off.

Comments welcome, of course.

July 2nd, 2008

Notebook SSDs REDUCE battery life at 50x the cost

Posted by Robin Harris @ 7:04 pm

Categories: Solid State Disk, Disk drives, Marketing

Tags: Disk, Battery, USB Flash Drive, Tom, Notebooks, Engineering, Flash Memory, Hardware, Notebooks & Tablets, Robin Harris

Tom’s Hardware is reporting test results that show that costly notebook flash drives actually use MORE power than hard drives - shrinking battery life by up to an hour. How could this be?

The test
Testing 4 SSDs against a 7200 RPM 2.5″ drive on a Dell Latitude D630 notebook they found

. . . the power savings aren’t there: in fact, battery runtimes actually decrease if you use a flash SSD.

TH stacked the deck using a 7200 RPM drive since the common 5400 RPM drives require less power. 1.8″ drives are even more efficient. A smaller form factor doesn’t help flash drives since they aren’t mechanical.

Marketing from the twilight zone
How could this be? Flash vendors have been touting power savings for over a year and - gulp - they don’t exist?

This is as bad a case of high-tech marketing groupthink as I’ve seen in years.

You can buy better but you can’t pay more!
TH went to great lengths to document their results since the outcome was unexpected. Here’s some of what they found:

• A Crucial SSD - costing $25/GB - used more power - 1.6 W at idle - than any 2.5″ notebook drive requires.
• A Memoright 32 GB drive used a full 2 W at idle
• An Mtron 32 GB flash drive reduced battery life by almost an hour.
• The slowest drive - a year old Sandisk SSD 5000 - almost equaled the Hitachi 7200 RPM Travelstar’s energy use. But the SSD offers fewer IOPS than the hard drive!
• They tested against a 200 GB Hitachi Travelstar 7k200, but other 2.5″ 7200 RPM drives have similar power envelopes.

What’s going on?
TH and I have several theories for the added power use:

• Limited power management TH saw that flash drives have just 2 states: low power and full power. Hard drives have dozens of power saving strategies and features.
• The flash control logic - disk translation layer - requires computing for wear leveling, garbage collection, buffer and cache management, flash mux/demux and the SATA interface - with frequent background operations even when the drive is idle.
• Hard drives are well-optimized and only reach peak power when they have a steady stream of single reads and writes. Elevator algorithms minimize head movement and power with a deep I/O queue.
• It takes 20 volts to write a flash cell - and since you have to write them all to erase a block and then write it with the data - that adds up.

The Storage Bits take
Initially enthusiastic about notebook flash drives (see A flash drive in your future? from 15 months ago), I started getting skeptical after running some numbers (see Power, notebooks and solid state disks).

But not skeptical enough. The power trade-off to get flash performance is grim. Some is tweakable, given enough time and money. Yet the cost/benefit ratio may not give vendors enough of either.

But the real idiots in this are the product marketing people who claimed “longer battery life” to help justify their high prices. Didn’t anyone bother to check?

The notebook SSD market is cratering before our eyes
Notebooks SSDs will come nowhere near the 25% market share Toshiba fantasized a few months ago. Server and array SSDs will be popular for the right reasons, and low-end notebooks - Eee-style - that can’t afford disks will also use flash to cut cost.

Comments welcome, of course. This won’t affect the media player market. They don’t need the performance notebooks do.

July 1st, 2008

Why computers fail

Posted by Robin Harris @ 1:25 am

Categories: Infrastructure, Clusters

Tags: Checkpoint, Multiprocessor, Failure, Computer, Desktop Computer, LANL, Desktops, Processors, Hardware, Semiconductors

Good failure data for PCs is hard to find: who knows how many times PC users are told to reinstall Windows? But in a recent paper, Bianca Schroeder and Garth Gibson of CMU found some surprising results in 10 years of large scale cluster system failures at Los Alamos National Labs.

Among the surprises: new hardware isn’t any more reliable than the old stuff. And even wicked smart LANL physicists can’t figure out the cause for every failure.

Special problems of petascale computing
Despite the incredible performance of Roadrunner, LANL’s new petaflop computer, the jobs it runs often take months to complete. With 3,000 nodes failures are inevitable.

What to do?
LANL’s strategy is stop the job and checkpoint. When a node fails they can roll the job back to the last checkpoint and restart, preserving the work already done - but losing the work done after the checkpoint.

Even using massively parallel high-performance storage the checkpoints take time away from getting the answer. Understanding Failures in Petascale Computers uses LANL’s data to better manage the tradeoffs and to suggest new strategies.

But its the failure data itself - and what it suggests about our own computers - that I found most interesting.

Failure etiology
Hardware accounts for over 50% of all LANL failures - with software about 20%. Given all the PhD’s at LANL you’d hope human error would be low on the list - and it is.

Here’s the graph:

Is reliability improving?
Nope. LANL hasn’t seen any improvement over the years - even with hardware from a decade ago.

The key metric
The research showed that

. . . the failure rate of a system grows proportional to the number of processor chips in the system.

Which is a big problem for massive multi-processor systems.

The Storage Bits take
Extrapolating these results to our desktop systems is straightforward - with one big caveat: most desktop system crashes are software, not hardware.

Otherwise the Blue Screen of Death would be the No Screen of Death.

The biggest finding is that we shouldn’t expect our system hardware to get more reliable. Improvements get balanced out by increased complexity.

Those of us with multi-processor systems can expect to see less reliability - though with just a few systems you won’t see any trends. It’s a classic “glass half full” situation: our systems won’t get better, but al least they won’t get worse.

Comments welcome, of course.

June 25th, 2008

Microsoft’s manycore strategy

Posted by Robin Harris @ 7:18 am

Categories: Infrastructure, Software

Tags: Strategy, Microsoft Corp., Multi-core, MR, Robin Harris

The post on Apple’s Grand Central multicore support in OS X.vi had some commenters claiming that Microsoft had already solved the multicore software problem. Far from it.

And Microsoft knows it
Microsoft has a multi-pronged campaign to get in front of multicore. They’re hiring smart people and spending real money. They will have a response to Grand Central.

Only 2 questions: when and how good? Microsoft’s forte is software that enables the moderately talented to produce functional software. Enabling Visual Basic programmers to exploit multicore chips is no mean feat.

Update: The word “multicore” - and the fact that we’re successfully using dual-core processors today - has confused some readers. Yes, limited forms of parallelization are built into most OS’s and many applications today. But these don’t scale well even to quad-core processors, let alone to 8-core and higher, except on some specialized and tuned apps.

The focus of Microsoft Research’s multicore effort is figuring out how to effectively use “manycore” processors in single systems as well as aggregating many systems for scale-out performance. Intel will be able to produce 32-core x86 processors in a few years. It sure would be nice to get 32x performance out of them.

The industry should start using the term “manycore” to describe the technology. I changed the title above. End update

Supercomputer guru
Why the focus on multicore? As Dan Reed, MR’s multicore research leader said recently in his blog:

. . . we are now approaching some fundamental physical limits. The bounds on CMOS junction voltages are near, and the number of atoms on a transistor junction is countable. Parallelism, some new technologies or radical new approaches (quantum computing or biological computing), will be required. All of this is driving the multicore revolution.

MR has an aggressive agenda:

• $20 million to support 2 Universal Parallel Computing Research Centers
• Dan Reed’s dedicated research group inside Microsoft Research
• A Safe and Scalable Multicore Computing external research program
• And more, such as the collaboration with the Barcelona Supercomputing Center

Microsoft Research takes the multicore problem seriously.

The Storage Bits take
Despite their obvious commitment to multicore research, Microsoft Research isn’t in the product development business. There is no direct link to what we’ll see in Windows 7. Time and chance happen to all features.

Reading the tea leaves it appears that Microsoft and Apple have - surprise! - different philosophies in exploiting multicore.

Microsoft - the developers, not the researchers - are focusing on using additional cores to add features to the OS and Office. High quality voice recognition and predictive workflow modeling, both items that will improve the man-machine interface. Spawning additional tasks also simplifies using more cores since the new features can be loosely coupled processes.

Apple - much more dependent on 3rd party developers - is focused on platform support for multicore. Like their successful Core Services strategy Apple will present APIs that enable developers to quickly leverage multicore performance.

The success of each approach remains to be seen. What’s important is that we, the users, will see some real competition. That’s good for all of us.

Comments welcome, of course. Update 2: Some have taken exception to my characterization of Microsoft’s forte above. My point was statistical, not personal.

When you own 90% of the desktop market and a good chunk of the server market the target market for your development tools will be the “average” developer, not the compsci PhD. To stress that I struck out the reference to programmers to put the focus on the tools, not the users. HTH. End update.