What does the MBA’s $1,000 SSD give you?
According to the Wall Street Journal’s Walt Mossberg: not much. Exactly as I predicted 3 weeks ago.

Battery life
Walt says:

The SSD MacBook gave me just five more minutes of battery life. Apple says this is because its hard-drive model already uses a very low-power drive.

I said:

Flash drives make exciting copy, but in today’s power-hungry notebooks they don’t make much of difference in battery life

You’d expect a 1.8″, 4200 RPM drive to sip power and it does. But the performance should be terrible, right?

Performance
Walt says:

The SSD version of the Apple booted up from a cold start, and rebooted with several programs running, about 40% faster. But the gain isn’t as impressive as it seems because even the hard-drive versions of the MacBook Air booted up in under a minute and rebooted in just a little over a minute.

I said:

The Air should boot faster with the flash drive, but an ultra-portable like this will mostly awake from sleep mode, where the difference will be minimal. Once up and running few users will be able to tell the difference.

What about work?
Walt says:

I also tested launching Microsoft Word and Excel, and opening a couple of hefty PDF files . . . . The SSD versions were faster. But in most cases, the gains were just a few seconds or even fractions of a second.

I said:

Nor did [Apple] offer any specific performance claims for the flash [drive] . . . .

The Storage Bits take
I’ve been following the flash follies closely for over a year. I would have been surprised if the SSD results had been different.

Kudos to Apple for not hyping the SSD. It is more durable than a disk - but not by much - and if you drop your MBA you’re more likely to break the display than the hard drive anyway.

More kudos to Apple for an excellent job of engineering the MBA for maximum battery life and performance with hard drive technology. As Mossberg noted, the ultra-portable Toshiba Portege R500 had significant battery life gains with their SSD, since they’d used a more power-hungry 2.5″ drive.

Comments welcome, of course. I’m thinking about getting an MBA - but not with the SSD.

Not only has the White House lost 5 million emails - or not, they aren’t sure - vital White House emails have for years been run through an insecure 12-man ISP in Chattanooga Tennessee. Why? Because our laws around Presidential records preservation are at odds with other laws against partisan political work by government employees.

The Internet is insecure? Who knew?
Not the first MBA president. Or his security-mad VP.

The VP’s mania for secrecy includes a man-sized safe in his office and stamping “Top Secret” on press releases. Evidently it never occurred to him that White House emails could be easily read by foreign powers or terrorist groups.

Another oddity: the White House migrated from an enterprise-class Lotus Notes email system to Microsoft Exchange during the build-up to the Iraq war. The White House has suggested that the emails were lost during that migration. Which is as likely as the “accidental” 18 and a half minute gap in the Nixon tapes.

Read the whole story
David Gewirtz, author of the book Where Have All The Emails Gone? has pioneered a new form of investigative reporting by taking a technical look at the White House email infrastructure. You don’t find screenshots of WHOIS output and Mail eXchange lookups in the Washington Post.

According to his bio on Zatz.com,

Gewirtz is a former professor of computer science, has lectured at Princeton, Berkeley, UCLA, and Stanford, and has been awarded the prestigious Sigma Xi Research Award in Engineering.

His company publishes books and newsletters about enterprise email systems.

It isn’t about lost emails
Gewirtz started looking at the White House email infrastructure - he publishes newsletters on Outlook and Lotus Domino - to find out how 5 million emails might get lost. He found something far more dangerous: an insecure White House email infrastructure.

The White House is legally required to preserve all presidential records, a law that Presidents don’t like. Ever since document shredding by Ollie North and Admiral John Poindexter - a Cal Tech PhD, by the way - during Iran/Contra was documented in “deleted” emails, the executive branch has had an uneasy relationship with email. They use it, but they fear it.

Under the Hatch Act, passed in 1939, Federal employees are prohibited from engaging in partisan political activity at work. But for senior presidential political advisors some of its prohibitions are outdated and dangerous to national security.

As Mr. Gewirtz puts it:

This isn’t about whether you George W. Bush or not, or whether you like Bill Clinton on not. It’s about email usage, management and oversight in the Executive Office of the President - not just Goerge W. Bush’s EOP, but email usage for any administration.

GWB43.com
When the Bush administration came in they decided to use external servers for “political” email that, legally, could not be sent through internal government servers. These “political” email accounts were set up on the domain “GWB43.com” hosted by the small ISP in Tennessee which also hosted numerous GOP-related sites.

The security lapses
Gewirtz identifies several problems with this arrangement.

1. He estimates that, at a minimum, over 100,000,000 emails from White House staffers have been sent since 9-11-01, completely in the open for anyone to read.
2. Many emails likely contained information useful to our enemies.
3. No archiving systems kept copies of these official documents.

The convoluted White House email system almost seems designed to lose email.

The Storage Bits take
I recommend Gewirtz’ book to anyone who wants a non-partisan understanding of the White House email mess. While I don’t agree with his analysis at every point he makes some important recommendations, including:

• The Hatch Act should be amended to require that all EOP emails, political or not, go through secure government servers.
• A professional civil-service IT team is needed to manage and protect White House email systems and archiving across administrations, just as the Secret Service protects the President.
• The primacy of Presidential records preservation must be established once and for all. Presidents have secrets they’d like to keep, but succeeding Presidents need to be able to know what their predecessors did. That means all records, all the time. No exceptions.

Comments welcome, as always.

Take 1 flailing Internet company, add a 2nd flailing Internet company, and what do you get? A bigger flailing Internet company.

Accentuate the positive
The proposal isn’t all bad. :

• It is an OK deal for Yahoo shareholders. Yahoo was trading above $30 just 3 months ago, so today’s big premium is where the stock was much of last year. It makes up for the damage a dithering Jerry Yang has done.
• Jerry Yang can quit pretending to be a CEO. There are a lot of rich people in Silicon Valley. Most of the them are lucky rich - people who just happened to be in the right place when big money arrived. Then there is a much smaller group of smart rich - people who figured out how to generate big money. Jerry is lucky rich, just like 99% of Googlers.
• The industry needs to consolidate. Companies need to merge to develop economies of scale. This will be Microsoft’s excuse to shut down major chunks of money-losing MSN.

Yahoo has a strong base in the email market and other web portal services. They’re profitable - no small thing on the Internet - and a well-known, if not well-respected, brand. All real assets.

The downside
Where to start?

• Both companies are clueless on Internet strategy. To be fair everyone else is clueless too. The Internet is new and still evolving. The only certainty is that old models don’t work. And the avatars of the old models are least likely to come up with new ones.
• This looks like AOL plus Time Warner all over again. Why Microsoft thinks it needs to be in the Internet business is beyond me. They are a software company with no demonstrated Internet smarts. How will buying Yahoo change that? It is a Gates/Ballmer ego thing.
• Neither company is cost-competitive with Google. Integration costs will be huge.
• Internet advertising isn’t turning out to be quite the money machine everyone imagined - not bad, but not change the world either. As a content provider I can tell you that Google’s pay-per-click model sucks. As new media networks with a pay-per-impression model develop, Google will be left selling ads on content that no one cares about.
• Paid web-services are the future. Google’s non-economic model for most of its services has depressed prices. But as services like SmugMug demonstrate, the best customers are willing to pay for quality. But you have to be really good. Microsoft is mediocre at best and Yahoo not much better.

Ballmer’s swan song
Pity poor billionaire Steve Ballmer. His entire career has been spent in the shadow of Bill Gates. For years he was the world’s richest man who worked for someone else. Since his elevation to CEO he’s done nothing to fix Microsoft, instead pursuing one ill-conceived venture after another. This is his biggest yet.

If the Yahoo deal goes through, Steve will have one last chance to show that he is an executive on par with Chambers, Jobs, Hurd or Palmisano. He is, sadly, too smart. Lucky smart.

Comments welcome, of course.

Create the ultimate gaming supercomputer?
You’ve overclocked, water cooled, matched DIMMs, added 10k drives and the latest 1 GB video card. But so have all your friends. What now? How about a 10 Gig home network for the ultimate gaming supercomputer?

In a pricing breakthrough you can now buy an 8-port 10 Gig switch, 2 PCI-Express 10 Gig adapters and cables for under $1100. It is the fastest network available for the dollar. Update:
By comparison the cheapest 10 gigE NIC at Newegg is almost $900.

One word, my friend: Infiniband
No, this isn’t 10 Gig Ethernet. An average 10 GigE switch port costs over $2500 today and the overhead of TCP/IP will bog down even hefty systems unless you buy a costly TOE (TCP/IP Offload Engine) adapter. No, this is Infiniband, a high-speed, low-latency, low-overhead network widely used in supercomputers, high-end storage and clustered computing.

Originally spec’d in 1999 by Intel, Microsoft and Sun (ngio) and Compaq, IBM and HP (Future I/O) to replace PCI, Infiniband has evolved into a general-purpose high-performance interconnect. As volumes have grown, prices have dropped, but this latest price-cutting iteration took me by surprise.

Drivers are available for Linux, Windows XP and OS X - though serious gamers aren’t likely to be using the latter. The kit is available from Colfax Direct, a new e-store subsidiary of 20 year-old Colfax International.

Some pricing from their web site:

• PCI-Express 10 Gbit adapter: $125
• 8-port unmanaged switch: $750
• Cables: range from $35 to over $900 for plenum-rated 100 M length

The Storage Bits take
Networks and storage can often substitute for each other. With a 10 Gig low-latency network you can configure diskless workstations that really scream. While today’s Infiniband networks are practical only for serious gear heads, early adopters will help point the way to a not-to-distant future when we all have 10 Gig home networks.

Commments welcome, of course. Disclosure: I have no relationship, financial or otherwise, with Colfax. I worked with Colfax’s chip provider, Mellanox, at a previous company and found them a pleasure to deal with.

Corporate desktops average single-digit CPU utilization and less than 20% storage utilization. Can this unused capacity be put to work for the data center? Revstor says “yes!”

I spoke to Russ Felker, the founder and CEO of RevStor to learn more about the product.

Average desktop PCs today come with disks ranging in size from 120 GB to 500. But unless you are editing video, producing presentations or crunching big data sets you probably use less than 50 GB. RevStor’s SANware backup product puts this unused space to productive use.

Recycling in place
Power! Cooling! Capacity! Data centers worldwide are bumping into infrastructure limits due to increasing server and storage density. Yet some estimate that 2/3rds of corporate computing and electrical power is used on the desktop. Why not offload the glass house data centers?

Revstor offers a product called SANware, a backup product that includes what they call a Distributed File System. DFS compresses, encrypts and splits files across desktops to ensure data security and minimizing impact on individual systems. Think of it as corporate BitTorrent.

Data availability
You can select how many copies of each piece are stored, so desktops can be offline without compromising data access. As desktops go offline SANware ensures that your required number of copies is maintained. A system goal is not to maintain more than 1/3rd of any file on any single desktop.

SANware also includes a de-duplication feature like other disk2disk backup appliances to reduce traffic. If you’ve got multiple offices you can specify remote locations for true disaster tolerance.

Performance
SANware is designed to be invisible to users. It doesn’t reduce the desktop’s available space - if you need to suddenly import a 100 GB video you can. SANware simply replicates from remaining copies to maintain your desired fault tolerance.

SANware only uses resources when the system is idle. It won’t interfere with your fantasy baseball league and other important business functions. It is so light weight that it is usable on a 500 MHz Pentium 2.

Security
All data has 256-bit encryption before it leaves the desktop. Even if someone swiped the desktop they’d have to find the invisible files on the disk, decrypt them and then find the other 2 desktops to get the complete file.

Management
Once the agent is installed and the initial settings completed, the system is designed to run without further management intervention. DFS is fully distributed so any node or group of nodes can host the file system metadata. If any of the metadata nodes go offline another node will be promoted to replace it without human intervention.

Who should use this?
Russ recommends that SANware makes the most sense for companies with 500-1000 Windows or Linux systems spread over 2-3 locations. You’ll have plenty of capacity and the additional locations give added availability.

The Storage Bits take
Revstor, along with companies like Seanodes and Cleversafe, is looking at how to use resources already in place for data storage. Most data doesn’t need the performance of a million dollar storage array, but it does need security and availability.

SANware is a neat middle step between local dedicated backup systems and remote data storage like Mozy or S3. It uses storage and a fast LAN you already have to create a self-managing backup cluster infrastructure. At $2500 per TB it is an affordable choice for companies who can’t add more raised floor space, power, cooling and maintenance contracts without busting their budgets.

Comments welcome, of course.

Steve Ballmer may be the worst CEO among large tech companies - now that Kevin Rollins got booted from Dell and Sanjay Kumar of CA is in jail. Put him in a room with Steve Jobs of Apple, John Chambers of Cisco and Mark Hurd of HP and he’d look like the bouncer, not a peer. He just isn’t in their league and Microsoft is suffering for it.

Ballmer became CEO in 2000 after 20 years as a top business manager at the company. He’s been intimately involved in company policy for decades. And like many hand-picked successors - like the 2 above - he hasn’t measured up.

What hath Steve wrought?
If he wasn’t Bill’s freshman roommate at Harvard - and Bill wasn’t majority owner of Microsoft - any other board would have booted Ballmer years ago. He’s cost Microsoft billions in profits while the stock price stagnated.

Just to pick some of the most obvious fiascos on his watch:

• The Netscape anti-trust fiasco. Microsoft broke the law. The many follow-on suits by states, competitors like Sun and resellers like Gateway have cost the company over $4 billion, innumerable hours of executive time and the continued attention of European anti-trust regulators. And for what? IE is losing market share and is no more critical to MS success than it was 10 years ago.
• The Google fiasco. Insisting on multiple frontal assaults against Google is pointless and costly. MSN is a money-loser and Microsoft will never catch Google’s ad business. But does Ballmer look ahead to the Next Big Thing that changes the game? No, why start now?
• All the other product and market fiascos. Virtualization: late to the game. The Xbox - losing market share to the Wii - will never pay back its investment despite a recent Halo-fueled lurch into profitability. Vista’s 5 year development cycle. The continuing security mess. The smart phone failure - a market they should have owned. Major resellers, like Dell, offering Linux. The steadily shrinking cash horde. Continuing anti-trust troubles. The current OOXML debacle.

Any other CEO would have been shown the door years ago.

Missing Google was huge - and avoidable.
Google made no secret of the fact 5 years ago that it was hiring the best and the brightest Microsofties. Hundreds left Redmond for Kirkland and Seattle.

That’s the early warning signal whose significance Ballmer missed. Those Googlers are rich, while the loyalists are wondering why they stayed. Despite the incredible technologists in Microsoft Research, Ballmer could never get Microsoft on the technical leading edge in new markets.

The Storage Bits take
Microsoft is loaded with smart, passionate people who sincerely want to do the right thing. Yet the industry has moved on while Microsoft executives - starting with Ballmer - haven’t.

Steve, you’ve had a good run. You’re one of the wealthiest people in the world. You’ve been integral to one of the greatest business successes of all time. But you’ve lost your mojo. You aren’t the guy Microsoft needs.

Do yourself, Microsoft and Microsoft shareholders a huge favor: resign. Let someone else pick up the reins and, hopefully, take Microsoft to new heights.

Comments welcome, as always. How would you fix Microsoft? One thought: shut down MSN, sell off the assets, and use the billion or so to fund a few hundred software startups. A few are bound to hit if you leave them alone.

Update: True, Jerry Yang of Yahoo is clueless. But as I mentioned yesterday, Yahoo is going out of business because their infrastructure can’t compete with Google.

A particularly odd bit of goofiness has hit the infosphere: cloud/utility computing mania. Nick Carr has written a book. IBM has announced, for the umpteenth time, a variation on utility computing, now called cloud computing. Somebody at Sun is claiming they’ll get rid of all their data centers by 2015.

R-i-i-i-ght.

You know the flying car in your garage?
The syllogism is:

1. Google-style web-scale computing is really cheap
2. Networks are cheap and getting cheaper fast
3. Therefore we’re going to use really cheap computing over really cheap networks Real Soon Now

Can you spot the fallacies?

Fallacy #1: Google is Magick
The world’s largest Internet advertising agency does have the cheapest compute cycles and storage (see my StorageMojo article Killing With Kindness: Death By Big Iron for a comparison of Yahoo and Google’s computing costs). But they do nothing that the average enterprise data center couldn’t do if active cluster storage were productized.

Google built their infrastructure because they couldn’t buy it. They couldn’t buy it because no one had built it. But all Google did was package up ideas that academics had been working on, sometimes for decades. Google even hired many of the researchers to build the production systems. Happy multi-millionaire academics!

Blame vendor marketing myopia for missing that opportunity. But their eyes are wide open now. If your enterprise wants cluster computes or storage you can buy it. From Dell.

Fallacy #2: Networks are cheap
Or they will be Real Soon Now.

10 Mbit Ethernet from Intel, DEC and Xerox came out in 1983. A mere 25 years later we have 1000x Ethernet - 10 GigE - starting down the cost curve.

About the same time a first generation 5 MB Seagate disk cost $800. Today a 200,000x disk - 1 TB - costs 300 vastly cheaper dollars.

Also in 1983 the “hot box” - a VAX 11-780 - with a 5 MHz 32-bit processor and a honking 13.3 MByte/sec internal bus cost $150,000. Today a 64-bit, 3 GHz quad-core server - with specs too fabulous to compare - is $1300. Call it 1,000,000x.

Networks are the bottleneck, not the solution. Hey, Cisco! Get the lead out!

What’s really going on?
There are - currently - economies of scale, which Google is exploiting and MSN and Yahoo! aren’t. So the latter two are going out of business.

But when you look at the cost of going across the network compared to the rest of infrastructure you realize that local - what we used to call distributed - computing is the only way to go.

Ergo, cloud computing will remain in the clouds and real computing will remain local. Where you can kick the hardware and savor fan hum and blue LEDs.

Sure, some low data rate apps - like searching - can move to the web. But if you want a lot of data and you want it now, keep your processor close and your data closer.

Comments welcome, as always.

Bytes, not bits. Oh, and it’s fast, too.
Nanochip, a Silicon Valley-based fabless semiconductor firm, just received $14 million in funding to complete work on a 100 GB storage chip. Intel Capital, who should know something about chips, is an investor. The goal: “. . . allow Nanochip to complete development of its first prototypes later this year . . . .”

MEMSy were the borogoves . . .
The Nanochip design is a Micro-Electro-Mechanical System, or MEMS, device. A descendent of IBM’s Millipede device, it uses polarization instead of Millepede’s heat to store data.

An array of tiny probes - looking like phonograph needles, if any of you have ever seen one - less than 25 um in diameter, changes the state of the recording medium. The probes are movable - similar to the mirrors on DLP chips - so they can write more than one location. Since there are many thousands of probes, they have a lot of bandwidth.

Update: this is persistent storage, i.e. it retains data with the power off. The Millipede design proposed using a heat-sensitive polymer to create little CD-like pits to record data. They couldn’t get that to work so now they are using an electrical method.

Here’s a scanning electron microscope picture of a probe:

graphic courtesy of Nanochip

Double your fun
The chip is actually 2 chips bonded together. One chip has the array of probes and the other has the media. They are bonded, diced and then mounted in traditional plastic packaging.

The really cool thing is that they can use 10 year old, fully depreciated, fab equipment to build these chips. They don’t need deep UV technology or any of the other costly tricks chip makers use today to scrunch chip sizes. A 1 micron fab is fine. That should translate into much lower costs.

The design is scalable to 1 TB chips, according to Nanochip. How cool is that?

The Storage Bits take
The MEMS-based memory device idea has been around for over a decade, but no one has yet figured out how to produce reliable devices in volume. This funding round shows that the VCs are seeing light at the end of the tunnel.

I sure hope they are right.

Comments welcome, as always. Update: Hilarious take on the mathematics of a 100 GB chip. Cool! Realistically, I’m sure they’ll use some ECC to protect the data.

Wave good-bye to your 60% gross margins!
Michael Dell is on the warpath, seeking to revive Dell’s flagging fortunes. Storage, with its 60% gross margins, is ripe for the plucking through commoditization. And who knows more about commoditizing IT than Michael Dell?

Right, the $1.4 billion acquisition of EqualLogic. Anything else?
You probably know you can buy low-end EMC Clariion storage at Dell, along with Dell’s own PowerVault systems. Plus the EqualLogic systems, notable for providing almost every software feature you’re likely to ever need, all for one package price. But there’s more.

Build a Dell supercomputer?
For example, HPC clusters and storage. Stanford created a 1,696 core supercomputer cluster, for their Flow Physics lab. Dell’s Advanced Systems Group put it together.

Fitted with Infiniband and the Linux cluster Rocks+ distro, high-performance Panasas parallel file storage and a jazzy visualization wall, you’d never guess little old Dell could do this. Dell sells all the products including the Panasas gear.

Eat your hearts out, gamers.

More cluster storage
Dell also resells the little-known Ibrix cluster storage software. Slap it on a bunch of Dell servers and you can have a powerful NAS infrastructure for enterprise use.

But that’s not all. Rumor has it that Dell is working on a high-density storage box - like 48 drives in the chassis - to complement its cluster offerings. Mr. Dell wants to take on the current $10/GB storage array business.

It is about time. Storage is a fat target.

Update: The first commenter is skeptical about Dell’s ability to sell complex storage. But that is part of the change in high-end storage: products that are much easier to configure, install and manage. Especially those that use Ethernet interconnects and self-discovering cluster software. A new cluster can be up in a few minutes. That ease of translates into fewer support headaches and is perfect for Dell’s business model.

Comments welcome, of course. Disclosure: I’ve done consulting with Panasas.

Apple offers a flash drive on their new MacBook Air. And it ain’t cheap. Should road warriors bite?

$999
Replacing the standard 80 GB hard drive with the 64 GB flash drive costs $999. 64 GB is plenty for business use and way-too-small for personal use - as is the 80 GB. But this isn’t a desktop replacement. The Air is a road warrior’s status symbol.

2 odd things in the announcement
Apple didn’t make a specific claim for improved battery life with the flash drive, like X minutes more. They spec’d the Air at 5 hours battery life with either drive.

Nor did they offer any specific performance claims for the flash, which surprises me. The stock 4200 RPM 1.8″ drive is about as slow as they come for today’s notebooks. If a flash drive can be faster than a disk, this is an easy target.

But perhaps Apple made a marketing decision to ignore any differences. So few people will shell out a grand for the flash, so why denigrate the popular model?

The only specific Apple claim for the flash drive is that it is more durable than the disk drive. While the 1.8″ has a very good shock and vibe spec, the flash drive is better and will likely outlive the disk. Both will outlive the rest of the system.

The Storage Bits take
Flash drives make exciting copy, but in today’s power-hungry notebooks they don’t make much of difference in battery life. The LED-backlight on the Air contributes more to the battery life than the flash drive can.

The Air should boot faster with the flash drive, but an ultra-portable like this will mostly awake from sleep mode, where the difference will be minimal. Once up and running few users will be able to tell the difference.

Apple’s low-key announcement of the flash drive is a welcome change from the hype and spin offered by flash vendors. Don’t expect too much and perhaps you’ll be pleasantly surprised.

Comments welcome, of course. Is the flash drive worth it for you?