Millicomputing - Open Hardware by the milliWatt

LTE - Huawei shows off 1.2gbps wireless

2010-03-23T16:04:00.000-07:00

Standard setting in 2011, with no-one yet saying when they might implement it, but this shows that the end game is wireless internet as the default

Mobile Trends 2020

2010-01-15T21:10:00.001-08:00

Check out this SlideShare Presentation: lots of different takes on the mobile future, and some common themes and echoes of some what I've been saying here...

Mobile Trends 2020

View more documents from rudydw.

Intel getting closer to phones with LG GW990

2010-01-08T00:17:00.000-08:00

This looks like an over-sized iPhone, and is a cross between a tablet and an iPhone. It has a SIM slot, but it has an Intel Moorestown low power Atom CPU and a 4.8 inch screen. So it can run variants of windows and Linux.

I think that Intel needs at least one more generation of even lower power CPUs to get into direct competition with ARM, but this is the closest I've seen so far.

Here is the LG GW990 hands-on from Engadget Mobile at CES

Quad-core ARM and 2010 predictions

2010-01-06T12:26:00.000-08:00

More than 1GHz per core, and now ARM has two and four core options. This is tracking my predictions from a few years ago quite well. The google Nexus One has a 1GHz ARM CPU with 512MB RAM. I expect that the rumored Apple tablet or iSlate is more likely to be a big iPhone than a small MacBook, so it would make sense for it to use a multi-core ARM and a Gigabyte or so of memory. This gives vendors a nice way to stratify their product, for example, Apple could support multi-core and background/multiple running apps in iPhone OS 4, release a two-core tablet now, then upgrade the iPhone to have a two core high end option in the summer, and push the tablet to have a four core high end option. If Apple doesn't, then the Android vendors will.

My other prediction for 2010 (that I made in early 2008) was that as ARM goes up market, Intel will come down-market to lower power consumption. So later in 2010 we may see Android based tablets and phones that use Intel Atom variants going head to head with ARM, as well as running more generic laptop derived operating systems.

http://www.marvell.com/armada/quadruple_core_arm_instruction_set/release/1363/

My millicomputing update ignite talk wasn't showing any text on the slides on slideshare, so I just re-loaded it.

ignite slides

2009-12-20T08:56:00.000-08:00

http://www.slideshare.net/adrianco/millicomputing-ignite-talk #millicomputing updated predictions

4G LTE Wireless data at 43Mbit/s

2009-12-16T08:39:00.000-08:00

LTE trials have started in Sweden, and a laptop data card shows this data rate, with 5Mbit/s upload and 43Mbit/s download. At this point we are 1 to 2 years away from this in the USA.
http://www.engadget.com/2009/12/16/teliasoneras-43mbps-wireless-data-downloads/

Mobile Skype Video Calls via Fring on Nokia

2009-11-24T10:24:00.000-08:00

It was just a matter of time, now it will spread to other platforms...

http://www.engadget.com/2009/11/24/fring-adds-skype-video-support-on-s60-threatens-to-make-front-c/

ARM CPUs for power efficient Web Servers - James agrees...

2009-09-07T01:46:00.000-07:00

James Hamilton highlights the power efficiency of ARM for general purpose web servers. It's almost two years since I gave the first talk on Millicomputing at the HPTS Workshop (which is where I met James for the first time) so its great to see him talking up the principles. He also makes the important point that ARM uses error correcting (ECC) memory, while the Intel Atom doesn't, and thus the Atom is actually less suitable for configuring large numbers in low power enterprise server applications.

The systems at http://www.linux-arm.org/Main/LinuxArmOrg are relatively inefficient blades, they have archaic spinning rust storage attached which must dominate the power consumption. A flash based storage subsystem would make much more sense to me. Web content delivery workloads are very well suited to low cost read-mostly flash storage. They do have a 1.2GHz ARM CPU and 1.5GB of RAM per blade, which is the biggest and fastest ARM configuration I've seen so far.

Wireless HD Video at 60GHz

2009-09-03T22:54:00.000-07:00

GiGaOM has an update on wireless HD video It seems that the UWB standard stalled, but alternatives are on the way.

Windows Mobile fades to irrelevance...

2009-08-19T10:00:00.000-07:00

see this comment on GigaOM like I said almost two years ago....

Augmented reality for the iPhone

2009-08-06T07:19:00.000-07:00

from Metaio just a demo, but heading in an interesting direction...

TI Sees multi-core phones in 2011

2009-06-23T20:21:00.000-07:00

Like I said at Ignite #velocityconf yesterday - see ARM Cortex A9 - which is 4x the iPhone 3GS http://gigaom.com/2009/06/23/ti-sees-multicore-phones-coming-in-2011/

iPhone 3GS Graphics

2009-06-19T16:16:00.000-07:00

The new hardware is a huge leap forward, the 3GS is in Xbox territory according to Noel - details: http://www.mobileorchard.com/a-huge-leap-forward-graphics-on-the-iphone-3gs/ and RoughlyDrafted http://www.roughlydrafted.com/2009/06/10/a-closer-look-at-iphone-3g-s-cortex-a8-arm-and-powervr-chips/
My order is in :-)

Ignite talk at Velocity Conference

2009-06-17T16:37:00.000-07:00

I'm at the Velocity Conference next week in San Jose, Mon-Wed. On Monday evening there is an Ignite session where we get 5 minutes to talk about something using 20 slides that auto-advance every 15sec. I've been accepted to do an update on Millicomputing, so I have to figure out how to get a slide deck together. I'm going to look back at what I predicted in 2007 and 2008, see how accurate I was, and look forward a couple of years for what low power devices we should expect to have in our pockets and our datacenters in 2010 and 2011.

Qualcomm Snapdragon - 1.3GHz Arm Cortex A8

2009-06-01T07:14:00.000-07:00

The performance of ARM architecture chips continues to increase, this announcement bumps the clock rate up to 1.3GHz and has a long list of integrated features.

The new chipset supports multi-mode UMTS and CDMA 3G connectivity in the same 15 x 15-mm package as well as featuring enhanced 2D acceleration and 3D graphics core, integrated GPS, high-def video recording and playback, Bluetooth 2.1, WiFi, and support for WXGA (1280x800 pixels) displays pumping out your choice of MediaFLO, DVBH, or ISDB-T digital mobile television.

Wireless broadband speeds in practice

2009-05-26T11:35:00.000-07:00

Useful table of standards and speeds via GigaOM

Spacial audio

2009-05-11T07:05:00.000-07:00

Another example of an idea I mentioned in my Millicomputing talk turning up at NTT. This is something that could be programmed on an iPhone using OpenAL I think. It's really the idea of using stereo sound to carry conversations, which depends upon people using stereo headsets. I think there needs to be additional value in the headset such as video capture of what I'm looking at, and into-eye video display like the MyVue, then spacial audio is an add-on not the main feature.

Renesas Low power HD Video processor

2009-05-06T06:51:00.000-07:00

Sighted at engadget mobile, where they think it could turn up in mobile devices that do both input and output of HD video at 1080p, 30fps

Andy Bechtolsheim on the Solid State Storage Revolution

2009-05-05T22:35:00.000-07:00

At the MySQL conference Andy gives a talk that is up on YouTube.

He gives a good overview of the hardware changes, but doesn't talk about the software challenges, where many of the assumptions and algorithms that are built into all the common operating systems, filesystems and databases are simply wrong. There is a lot of work to be done here. My own work at Sun involved several attempts to point this problem out over the years, and if I had done a better job of getting traction for my ideas, there would be a lot more preparation and research in this area by now. The specific examples include trying to explain the benefits of embedding NVRAM products into servers during the 1990's, and doing extensive testing with solid state disks around 2002. The current size and cost benefits that we see were very predictable, and so are the problems. The SSD's at the time were very expensive, but they were perfectly capable of supporting research and development of new algorithms, and an opportunity was missed.

The next thing that is coming is non-volatile RAM (all the RAM in the system), in case anyone is listening. There need to be fundamental changes in operating systems (user level memory protection models etc.) and we then have the ability to build truly diskless database servers, with safely persistent in-memory databases, in user space, no device drivers.

Wireless Video standards war

2009-04-29T07:56:00.000-07:00

A bunch of references from GigaOM - it looks as if UltraWideBand 480Mbit USB is in trouble, and the alternatives are a dedicated Wireless HD video standard and Wifi based video streaming. UWB has some nice characteristics, but perhaps the ubiquity of WiFi wins again.... (There have been many cases over the years of technologies losing to Ethernet and IP based networks in the end).

http://gigaom.com/2008/10/31/ultra-wideband-near-death-as-wiquest-shuts-down/

http://gigaom.com/2008/04/09/wireless-hd-is-the-new-front-in-a-standards-war/

http://gigaom.com/2008/08/15/vcs-hope-to-see-wi-fi-everywhere/

I just want to be able to stream video from my iPhone to my TV set without any wires, don't care what the technology is, and I'm thinking of getting a WiFi enabled TV with Netflix embedded in it around the end of this year.

MIT Medialab's Sixth Sense

2009-04-20T22:54:00.000-07:00

This is exactly the kind of user interface development that I've been looking forward to as part of my Millicomputing meme. I hope it inspires some products soon...

iPhone 3.0 hardware interface implications

2009-03-26T22:41:00.000-07:00

An insightful posting by Daniel Eran on the iPhone 3.0 hardware interface controller, and its significance.

All I would like to add is that I really hope wideband/wireless USB is a feature of the next iPhone hardware. 480Mbits/s without any wires, and with wireless video output, so your pocket can drive your projector sunglasses, TV, meeting room projector or computer screen without even leaving your pocket.

Intel licenses Atom to TSMC

2009-03-02T21:20:00.000-08:00

Intel licenses low power Atom CPU design to TSMC. This is another step along the road of bringing the Intel x64 architecture down into low power embedded devices to compete with ARM.

TI OMAP4 is based on ARM Cortex A9

2009-02-16T10:19:00.000-08:00

Nice summary on GigaOM about the new chip just launched by TI.

This continues along the direction I outlined two years ago, high def video and graphics processing, very long battery life, high speed networking and lots of CPU capacity in a multicore chip, sampling soon, and in products in 2010.

Extended iPhone cookbook XML parser

2009-01-04T23:59:00.000-08:00

Erica Sadun published a tree based XML parser for the iPhone that works quite well. I needed to extend it to parse attributes and I posted my code back to Erica's google code forum. This works quite efficiently, my app parses complex XML for 100 movies in less than a second as it picks out the instant format ones.

Parsing API Results - XML vs. JSON

2008-10-29T08:56:00.000-07:00

There is an option to get JSON format or XML format back from the API. I'm investigating ways to parse JSON to see if its easier than XML (simply add output=json to the URL call).

My open source google code efforts are on hold for now since I have much more demand for new features in my app than people asking for source code. Let me know if you want code and I will assist on a case by case basis. Later on I will post cleaned up code to the google code project.

Netflix API - Parsing the results of an API call - Part 5

2008-10-10T22:52:00.000-07:00

One of the first API calls I needed to call gets the user's name and other information. In my case all I need is the first and last names, and the flag that says whether the user can use instant watching. This flag is set true for main accounts, and false for account profiles that don't have an instant queue.

The return data from the API call is in XML format, so this code shows how to use the XML parser to pick information out of the data returned from an API call. I provide a convenience method that combines the first and last names, and I didn't need the other link information, so it is ignored, but the code could easily be extended to pick it out as needed.

Some of the code below doesn't display properly. I'm going to set it up in google code soon to make it easier to manage.


//
//  NetflixUser.h
//  Instant Flix
//
//  Created by Adrian Cockcroft on 10/8/08.
//  Copyright 2008 millicomputing.com
// Licenced using Creative Commons Attribution Share-Alike
//  http://creativecommons.org/licenses/by-sa/3.0/

#import 


@interface NetflixUser : NSObject {
NSData  *rawData;
NSString *first_name;
NSString *last_name;
bool  can_instant_watch;
NSXMLParser *userParser;
NSString *currentElement;
}

@property(readonly) bool can_instant_watch;

- (NetflixUser *)initWithAPIResponse:(NSData *)response;
- (NSString *)name;

@end


//
//  NetflixUser.m
//  Instant Flix
//
//  Created by Adrian Cockcroft on 10/8/08.
//  Copyright 2008 millicomputing.com
// Licenced using Creative Commons Attribution Share-Alike
//  http://creativecommons.org/licenses/by-sa/3.0/

#import "NetflixUser.h"

/* Sample returned raw data

[userid]
Adrian
Cockcroft
true













*/

@implementation NetflixUser

@synthesize can_instant_watch;

- (NetflixUser *)initWithAPIResponse:(NSData *)response {
rawData = response;
[rawData retain];
first_name = nil;
last_name = nil;
can_instant_watch = NO;

//NSString *responseBody = [[NSString alloc] initWithData:response encoding:NSUTF8StringEncoding];
//NSLog(@"NetflixUser: %@", responseBody);

userParser = [[NSXMLParser alloc] initWithData:rawData];

  // Set self as the delegate of the parser so that it will receive the parser delegate methods callbacks.
  [userParser setDelegate:self];
  [userParser setShouldProcessNamespaces:NO];
  [userParser setShouldReportNamespacePrefixes:NO];
  [userParser setShouldResolveExternalEntities:NO];
  [userParser parse];

return self;
}

- (NSString *)name {
return [first_name stringByAppendingFormat:@" %@", last_name];
}

- (void)parserDidStartDocument:(NSXMLParser *)parser{
//NSLog(@"started parsing");
}

- (void)parser:(NSXMLParser *)parser parseErrorOccurred:(NSError *)parseError {
NSString * errorString = [NSString stringWithFormat:@"Unable to parse XML (Error code %i)", [parseError code]];
NSLog(@"error: %@", errorString);
}

- (void)parser:(NSXMLParser *)parser didStartElement:(NSString *)elementName namespaceURI:(NSString *)namespaceURI qualifiedName:(NSString *)qName attributes:(NSDictionary *)attributeDict{  
  //NSLog(@"found this element: %@", elementName);
currentElement = [elementName copy];
}

- (void)parser:(NSXMLParser *)parser didEndElement:(NSString *)elementName namespaceURI:(NSString *)namespaceURI qualifiedName:(NSString *)qName{   
//NSLog(@"ended element: %@", elementName);
currentElement = nil;
}

- (void)parser:(NSXMLParser *)parser foundCharacters:(NSString *)string{
//NSLog(@"found characters: %@", string);
// save the characters for the current item...
if ([currentElement isEqualToString:@"first_name"]) {
 first_name = [string copy];
} else if ([currentElement isEqualToString:@"last_name"]) {
 last_name = [string copy];
} else if ([currentElement isEqualToString:@"can_instant_watch"]) {
 if ([string isEqualToString:@"true"]) {
  can_instant_watch = YES;
 } else {
  can_instant_watch = NO;
 }
}
}

- (void)parserDidEndDocument:(NSXMLParser *)parser {
//NSLog(@"found %@ who %@ instant watch", self.name, (can_instant_watch? @"can": @"cannot"));
}

@end

Netflix API - Netflix Specific OAuth iPhone Code - Part 4

2008-10-01T00:09:00.000-07:00

I have already discussed how to get OAuth to build for an iPhone in Part 2. To use OAuth to call Netflix there are two small changes needed. The first one is that the way that characters are escaped in URLs needs to be tightened up a bit, otherwise the signature strings will work some of the time and fail when they happen to contain the wrong character sequence. This took a while to figure out...

The file NSString+URLEncoding.m needs to have a few characters (space, plus and asterisk) added as shown below:

Note, I'm having a hard time getting code to look good here, its hard to find a way to render code in a narrow column that doesn't mess up the formatting and works in more than one browser using blogger.com tools and templates.


- (NSString *)encodedURLString {
NSString *result = (NSString *)CFURLCreateStringByAddingPercentEscapes(kCFAllocatorDefault,
                                             (CFStringRef)self,
                                             NULL,
                                             CFSTR("?=& +*"),     // legal URL characters to be escaped
                                             kCFStringEncodingUTF8); // encoding
return result;
}

- (NSString *)encodedURLParameterString {
   NSString *result = (NSString *)CFURLCreateStringByAddingPercentEscapes(kCFAllocatorDefault,
                                             (CFStringRef)self,
                                             NULL,
                                             CFSTR(":/= +*"),
                                             kCFStringEncodingUTF8);
return result;
}

The second thing is that when the authentication is complete, Netflix returns a token that contains an encoded user identifier, as well as a secret and a key. The standard OAuth code only expects the secret and key, so a new NetflixToken class was created to hold the extra information, and to persist it in the iPhone's defaults store along with the secret and key. This means that once the user has signed into OAuth once, this information is saved and they never have to sign in again unless either Netflix or the User revokes the token. One more method was added to remove an entry from the defaults store, for use during logout.

First the NetflixToken.h header file:


//
//  NetflixToken.h
//  Instant Test
//
//  Created by Adrian Cockcroft on 9/11/08.
//

#import 
#import "OAToken.h"

@interface NetflixToken : NSObject {
@protected
 NSString *key;
 NSString *secret;
 NSString *user;
}
@property(copy, readwrite) NSString *key;
@property(copy, readwrite) NSString *secret;
@property(copy, readwrite) NSString *user;

- (id)initWithKey:(NSString *)aKey secret:(NSString *)aSecret user:(NSString *)aUser;
- (id)initWithUserDefaultsUsingServiceProviderName:(NSString *)provider prefix:(NSString *)prefix;
- (id)initWithHTTPResponseBody:(NSString *)body;
- (int)storeInUserDefaultsWithServiceProviderName:(NSString *)provider prefix:(NSString *)prefix;
- (int)removeFromUserDefaultsWithServiceProviderName:(NSString *)provider prefix:(NSString *)prefix;
- (OAToken *)oaToken;

@end

Then the code itself, this is all based on a simple extension of OAToken, which is part of the OAuth code base mentioned in part 2.


//
//  NetflixToken.m
//  Instant Test
//
//  Created by Adrian Cockcroft on 9/11/08.
//

#import "NetflixToken.h"

@implementation NetflixToken

@synthesize key, secret, user;

#pragma mark init

- (id)init {
    [super init];
    self.key = @"";
    self.secret = @"";
 self.user = @"";
    return self;
}

- (id)initWithKey:(NSString *)aKey secret:(NSString *)aSecret user:(NSString *)aUser {
 [super init];
 self.key = aKey;
 self.secret = aSecret;
 self.user = aUser;
 return self;
}

- (OAToken *)oaToken {
 return [[OAToken alloc] initWithKey:self.key secret:self.secret];
}

- (id)initWithHTTPResponseBody:(NSString *)body {
    [super init];
    NSArray *pairs = [body componentsSeparatedByString:@"&"];
    
    for (NSString *pair in pairs) {
        NSArray *elements = [pair componentsSeparatedByString:@"="];
        if ([[elements objectAtIndex:0] isEqualToString:@"oauth_token"]) {
            self.key = [elements objectAtIndex:1];
        } else if ([[elements objectAtIndex:0] isEqualToString:@"oauth_token_secret"]) {
            self.secret = [elements objectAtIndex:1];
        } else if ([[elements objectAtIndex:0] isEqualToString:@"user_id"]) {
   self.user = [elements objectAtIndex:1];
  }
 }
    
    return self;
}


- (id)initWithUserDefaultsUsingServiceProviderName:(NSString *)provider prefix:(NSString *)prefix
{
 [super init];
 NSString *theKey = [[NSUserDefaults standardUserDefaults] stringForKey:[NSString stringWithFormat:@"OAUTH_%@_%@_KEY", prefix, provider]];
 NSString *theSecret = [[NSUserDefaults standardUserDefaults] stringForKey:[NSString stringWithFormat:@"OAUTH_%@_%@_SECRET", prefix, provider]];
 NSString *theUser = [[NSUserDefaults standardUserDefaults] stringForKey:[NSString stringWithFormat:@"NETFLIX_%@_%@_USER", prefix, provider]];
 if (theKey == NULL || theSecret == NULL)
  return(nil);
 self.key = theKey;
 self.secret = theSecret;
 self.user = theUser;
 return(self);
}


- (int)storeInUserDefaultsWithServiceProviderName:(NSString *)provider prefix:(NSString *)prefix
{
 [[NSUserDefaults standardUserDefaults] setObject:self.key forKey:[NSString stringWithFormat:@"OAUTH_%@_%@_KEY", prefix, provider]];
 [[NSUserDefaults standardUserDefaults] setObject:self.secret forKey:[NSString stringWithFormat:@"OAUTH_%@_%@_SECRET", prefix, provider]];
 [[NSUserDefaults standardUserDefaults] setObject:self.user forKey:[NSString stringWithFormat:@"NETFLIX_%@_%@_USER", prefix, provider]];
 [[NSUserDefaults standardUserDefaults] synchronize];
 return(0);
}

- (int)removeFromUserDefaultsWithServiceProviderName:(NSString *)provider prefix:(NSString *)prefix
{
 [[NSUserDefaults standardUserDefaults] removeObjectForKey:[NSString stringWithFormat:@"OAUTH_%@_%@_KEY", prefix, provider]];
 [[NSUserDefaults standardUserDefaults] removeObjectForKey:[NSString stringWithFormat:@"OAUTH_%@_%@_SECRET", prefix, provider]];
 [[NSUserDefaults standardUserDefaults] removeObjectForKey:[NSString stringWithFormat:@"NETFLIX_%@_%@_USER", prefix, provider]];
 [[NSUserDefaults standardUserDefaults] synchronize];
 return(0);
}
@end

Netflix API - Announcement - Part 3

2008-10-01T00:03:00.000-07:00

Here is the official announcement of the API and how to get access to it.

Starting Wednesday, Oct. 1 the Netflix API is open to all:

The Netflix API:

- Allows access to data for 100,000 movie and TV episode titles on DVD as well as Netflix account access on a user’s behalf
Netflix has more than 2 billion ratings in its database
Netflix members rate more than 2 million movies a day
Netflix ships more than 2 million DVDs on a typical day

- Is free

- Allows commercial use
E.g. if a developer creates an iPhone app and wants to sell it for $0.99, that’s ok

Technically, the Netflix API:

- Includes a REST API, a Javascript API, and ATOM feeds

- Uses OAuth standard security to allow the subscriber to control which applications can access the service on his or her behalf

Developers can get access:

- Starting 10/1

- By self sign up at http://developer.netflix.com

Netflix API - Getting OAuth to work on iPhone - part 2: Adding the OAuth Code

2008-09-17T23:23:00.000-07:00

To start with I found some examples by Nick Dalton that helped me build a simple application that included a Web View, a screen that acts like a web browser but with my custom Objective-C code embedded in it. This is important, because the OAuth sign-in process uses a web page, but on the iPhone, if you spawn a copy of Safari to visit a web page, your application quits first.

Next the Source Code Manager in Xcode was configured to load the OAuthconsumer Objective-C framework via subversion. This was easy and obvious, enter the URL http://oauth.googlecode.com/svn/code/obj-c/ and checkout the code.

When trying to import the framework I discovered that Apple does not allow user specified binary frameworks to be added to iPhone applications. To work around this the source code was copied from the Xcode project for the framework, to the Xcode project for my Instant Test application. I renamed the framework Classes folder as OAuth and copied to my project via drag and drop, choosing to copy the underlying files. The Cocoa Categories, Protocols and Other Sources>Crypto folder were also copied. The Tests folder did not compile for iPhone so don't bother to copy it over.

The standard system Security.Framework doesn't need to be added to the Frameworks folder. I initially thought it did, but its probably only needed for the KeyChain code.

The iPhone doesn't support the KeyChain functionality, so if you try to build for iPhone it will fail. It does however build for the iPhone Simulator, which is confusing. Open up the OAuth source code, and delete the last two files OAToken_KeychainExtensions.h and OAToken_KeychainExtensions.m.

Since the code is no longer a framework, the header file references need to be changed from #include to #include "file" for all the includes in OAuthConsumer.h apart from the first Foundation one.

At this point, before you try and call anything, try a build, it should compile with no errors. If it doesn't, look for missing files.

At this point, you should have all you need to connect to an OAuth service

Netflix API Getting OAuth to work on iPhone - Instant Queue Add part 1: why?

2008-09-16T22:23:00.000-07:00

Why develop an iPhone app? Its "the future", a useful skill, and I can carry whatever I develop in my pocket and make it do whatever I want.

I work at Netflix, and I have instant watching on my TV, I built an application "Instant Queue Add" that lets me add a title to my instant queue using my iPhone in a couple of touches. It takes the Top20 and New Choices RSS feeds to find content, and it spawns a copy of Safari to add to instant queue for each pick. The first time it starts, you have to login, then it remembers the Netflix cookie. However I really want to add more features and avoid spawning a copy of Safari with a screen scraped URL.

The Netflix API uses a new standard called OAuth for Open Authentication. There are lots of features, but its complex, and there is no standard off the shelf library for OAuth on the iPhone. However it is a useful building block for more advanced applications.

In this series of posts, I will document the steps I'm making to get OAuth to work on the iPhone using Objective-C and Xcode. My starting point is this code base and tutorial by Jon R. Crosby, which is based on desktop MacOS X, and doesn't directly support the iPhone.

Intel high speed SSD

2008-09-08T17:30:00.000-07:00

Here is Tech Report's review of Intel's high speed SSD, which confirms the trend I've been talking about for a while. SSD's were always faster for random read and write, now they are faster for sequential read (250MB/s), and the "extreme" Intel model is faster for sequential write as well (170MB/s). They use less power and have comparable MTBF to the best enterprise disk drives.

The remaining disadvantages of total size and cost are being eaten away over time...

The Physical Web

2008-08-06T07:18:00.000-07:00

Nat Torkington on O'Reilly Radar declares a familiar theme...

The next step for computing is to move out from the computers. Every device has the potential to become network-connected, delivering information to or from a web service. The mobile phones in our pockets also let us take apps and network service with us wherever we go. Early hackers are building in this space. Big challenges include: take products to the masses and the environmental impacts. We're at early stages yet, but the room for expansion is huge. Projects to watch include Nokia's reinvention as services company, iPhone, and Google's Android platform.

Consumer grade Flash in an SSD package - WAIF

2008-07-24T04:31:00.000-07:00

RegHardware mentions Raidon's Compact Flash in a 2.5" SATA disk form factor which can be loaded up with cheap CF cards (32 Gig for $100 at the time of writing). The Raidon package holds two CF cards which can be mirrored for safety, or striped/concatenated (its not clear which) using "NRAID" which doesn't require both CF cards to be the same size.

I'd like to see a similar concept go even further using microSDHC, it should be possible to get a Wide Array of Inexpensive Flash (WAIF) based drive with consumer based pricing and very high storage capacity and bandwidth. Its going to be appropriate for read-mostly workloads such as personal use in laptops, static content web serving and archival storage.

Next-Generation Mobile Broadband - The 4G Summit at PARC

2008-07-08T20:59:00.000-07:00

There is a talk at Xerox PARC next week sponsored by the Wireless Communications Alliance. They don't have URL's that link to specific events so here is the full description, currently listed on their site. Unfortunately I have a work commitment so I can't attend, they do mention many of the very advanced ideas that I have been talking about in my Millicomputing talks, such as video conferencing over high bandwidth mobile networks.

Tuesday, July 15th 2008, 4:00pm - 6:00pm
WCA CenterStage presents: Next-Generation Mobile Broadband - The 4G Summit
Venue: Palo Alto Research Center, Palo Alto CA

-- Moderator: Iain Gillott, Founder & President, iGR Inc.

-- Jake MacLeod, Principal Vice President & CTO, Bechtel
-- Barry Davis, Exec Director of Products & Services, Clearwire
-- Jim Orr, Principal Network Architect, Fujitsu Network Communications
-- Jon Hambidge, EVP & Chief Marketing Officer, NextWave Wireless
-- Samir Khazaka, Sr Director Technical Marketing, QUALCOMM
-- Gennady Sirota, VP Product Management, Starent
-- Lee Tjio, Director of Advanced Technology & Strategy, Verizon Wireless

Visionaries who speak about fourth-generation mobile technology (aka 4G) often allude to the tantalizing promise of services and features previously found only in science fiction; interactive holographic video, handheld devices with high-resolution (better than HDTV) images, streaming HD video conferencing and real-time interaction while mobile. 4G also promises a convergence between technologies, for example; mobile payments using near-field communications and handset-based smart cards, personal assistant technologies in which your mobile device will interact with networked devices and services based on your location/schedule/current actions/etc. Implementing the 4G vision of the future will require a bandwidth of at least 100Mbps, which has implications for spectrum policy not supported by current licensing and bandplans.

It's generally accepted that 4G will run over an IP infrastructure, will interoperate with 802.xx technologies (Wi-Fi, WiMAX, Bluetooth, ZigBee, etc), and will need to support data-rates from 100Mbps to as high as 1Gbps. It's also expected that 4G will be a collection of technologies and protocols; versus one single standard. There are at last three major camps (and a few upstarts) that aspire to be the major 4G mobile data service provider and have a dominant influence at defining the mobile broadband market for decades to come. Will it be one of the major camps, or will there be a dark horse that emerges?

On July 15th 2008 the Wireless Communication Alliance CenterStage will proudly present "Next-Generation Mobile Broadband : The 4G Summit". Stakeholders from various camps around the 4G battleground will come together under a flag of truce to debate the strengths and weaknesses of the approach on which they're betting. What's real, and what's simply hype? Who will be the first to achieve 4G ratification, and when is a realistic date when this will happen? What are the implications for technology vendors, service providers, and content developers? Are there any non-Western standards also likely to be contenders? How will the industry address spectrum licensing challenges and bandplans which today would seem to favor FDD versus TDD technologies? How will spectral refarming, cognitive radio, and spectrum-sharing technologies affect the market? Given that the evolution of technology demands that existing 3G systems will have to co-exist with future 4G systems; how will that transition take place and are there business opportunities in helping to facilitate that transition?

Insights, information, and understanding will be the take-aways from this exciting WCA event. How can you afford to not be there? Mark your calendar for July 15th 2008 and plan to attend! We expect this event to sell out, so to ensure your seat we recommend that you register now for this event.

Cost: $20 at the door, $15 in advance via PayPal/Credit Card

Wonderland - Immersive Virtual World

2008-07-08T08:10:00.000-07:00

Interesting discussion of Sun's Wonderland project. Of particular interest is that they are using attenuation and stereo audio to place voices in space, so as you move around people come into earshot. Its also an open source project, and this is the kind of audio interface I've been discussing for a while.

LED Displays in Contact Lenses

2008-07-08T00:42:00.000-07:00

There is some research going on to develop LED Contact Lenses, reported in the Guardian, and mentioned in Guy Kawaski's blog.

This looks like a neat alternative to video eyeglasses, however the image will move with your eye, so you will see it superimposed over everything else. With eyeglasses the image moves relative to your head, so you can focus attention on a specific part of the image as you move your eyes. However, this is probably the least of their problems as they try to develop the idea...

Slides - Usenix 08 - Invited Paper on Millicomputing - Boston June 25

2008-06-27T00:02:00.000-07:00

| View | Upload your own

Computerworld and others on Millicomputing

2008-06-26T20:42:00.000-07:00

Nice writeup of my Usenix talk by By Sharon Machlis, Computerworld June 25, 2008

Infoworld copy.

Version at PC World. There were several blogs also copying this story.

From some comments it appears that some people didn't get that I was talking about using "e-sunglasses" head mounted video displays rather than a big laptop screen while on the move.

Intel Atom Reviews

2008-06-26T07:22:00.000-07:00

The Register has done a useful review of the current state of the Intel Atom CPU systems and motherboards. It is still an order of magnitude more power than a millicomputer, but its an order of magnitude smaller than most other Intel architecture systems, so its interesting to see the current state of the art.

BAFuture event on crowdsourced mobile - buglabs

2008-06-23T20:45:00.000-07:00

I just listened in to a talk from Buglabs, via Ustream video. Worth checking out. Buglabs have a nice set of modules that can be used to build homebrew devices.

http://www.ustream.tv/channel/crowd-and-open-sourced-mobile-devices-future-salon

Video streaming from phone's

2008-06-15T08:49:00.000-07:00

Robert Scoble talks about the capabilities of three services on TechCrunch. When I gave my talk on Millicomputing at the BIL conference, Robert was sitting at the front streaming my talk to Qik from his cellphone.

In the near future I think this will become an important area, as continuously streamed video conferencing becomes ubiquitous, and adds two-way support. The only limitations that prevent it are network bandwidth, battery life and the software that manages the service. Network bandwidth is already sufficient, battery life is improving rapidly, and these three companies (http://qik.com, http://kyte.tv, and http://flixwagon.com) are competing to build the software services that will eventually implement the features I have been talking about. These services are working towards computer assisted telepathy.

Arm vs. Intel

2008-06-02T07:13:00.000-07:00

An article in Infoworld confirms the things I've been talking about for the last year or so.

Arm are responding to the threat from Intel, and talking about low power enterprise servers.

Samsung 256GB SSD coming later this year

2008-05-28T07:39:00.000-07:00

The current generation of SSD's are smaller and slower than regular disks in the same form factor, but this is starting to change as products like this Samsung 256GB SSD reach the market. It has a 200 Mbyte/s read speed and 160Mbyte/s sequential write speed, in a standard 2.5" disk form factor.

So the sequence goes like this:

1) SSD's will be faster for random access (already happened)

2) SSD's will be faster for sequential access (coming later this year)

3) SSD's will be higher capacity (maybe next year?)

4) SSD's will be lower cost per GB (when production volumes ramp up)

The advantage in random access and reliability (due to no moving parts) means that relatively fewer SSDs are needed than spinning rust disks to provide the same availability and performance, so the end user cost per configured GB should switch in favor of SSDs earlier.

ATT plans 20Mbit/s to your phone in 2009

2008-05-15T00:09:00.000-07:00

AppleInsider reports ATT's plans for faster wireless networks. 20Mbit/s in 2009 moving up to 100Mbit/s in subsequent years.

Some parts of the world are already running at these speeds, but this validates my mobile millicomputing story. There is going to be an excess of bandwidth to your pocket. The applications that work out how to leverage that capacity are the ones that will take off over the coming years. Streaming video is obvious, and its all about the price and the variety of content. It's the non-obvious applications that will shape the future.

Ubiquitous Computing

2008-05-13T08:59:00.000-07:00

Here is Nat Torkington on O'Reilly Radar talking about Ubiquitous Computing. Its a useful jumping off point into several leading researchers sites.

I added this comment:

The technology required to support ubiquitous computing is reaching a tipping point, in the next year or so all the obstacles will melt away and the devices we carry in our pockets will have an excess of compute power, storage capacity, network bandwidth, and battery capacity. The developer space is moving from "death by 1000 ports" on very limited platforms to two that matter, iPhone and Android that have raised the baseline and opened up to a new breed of applications. I've been tracking and predicting this on my Millicomputing blog (at http://www.millicomputing.com ) and talking about it at conferences like BIL, eComm etc. We have also been building our own open source homebrew mobile phone hardware....

There seems to be a current focus on urban computing, and integrating people with the dense mesh of location aware services and communication opportunities that exist in cities. I'm more interested in the effects of taking "friction" out of communications between people. This is a concept that I picked up while working at eBay. In effect eBay took friction out of selling, PayPal took friction out of payments, and Skype took friction out of communicating. That is what made those businesses take off rapidly.

So far mobile phones have also taken friction out of communicating, we don't need to be tied to a wired location to communicate. Skype has removed the frictions of cost and ease of use, and has provided improved audio and video quality while you are at your desk or toting your laptop. One of the missing links is mobile Skype, it's still a bit slow and inconvenient to have Skype in your pocket, but the mobile versions of Skype are improving and the hardware needed to make them mainstream is on the way.

I'm still seeing most people thinking of their pocket device as a relatively dumb client terminal that hooks up to web services, I think this is a blinkered view. The thing in your pocket will become your server, and the compelling applications will be the ones that take most advantage of what can be done right there right now....

LBNL "ipod supercomputer"

2008-05-09T10:08:00.000-07:00

http://uk.news.yahoo.com/vdunet/20080507/ttc-us-boffins-design-ipod-supercomputer-6315470.html

M. Wehner, L. Oliker, J. Shalf, "Towards Ultra-High Resolution Models of Climate and Weather", Internation Journal of High Performance Computing Applications (IJHPCA), April, 2008.

This system proposes a custom CPU core, which makes sense given the large scale design. Its somewhat similar to the SiCortex machine in many ways.

Speaking at UK CMG TEC Conference - May 19-21st 2008

2008-05-09T07:55:00.000-07:00

UKCMG TEC 2008 is near Northampton in the UK. I'm giving two of the same presentations as I gave at the US CMG, the enterprise version of my Millicomputing talk, and a half day workshop on Unix/Linux Performance.

eComm08 Millicomputing Talk on Video

2008-05-08T12:13:00.000-07:00

Video of my talk at eComm08, along with about 40 other talks...

Wireless USB / UltraWideBand Networking - fast, low power, local wireless

2008-03-20T18:37:00.000-07:00

I started talking about wireless video from mobile devices before I discovered that the technology is well under way, chip-sets are available and its using a cool new technology that has lots of nice characteristics including very low power usage.

Infoworld on Game Changing Technologies

Nice review of UltraWideBand networking.

Summary of news on UWB, including HD video output.

eComm08 Millicomputing Talk

2008-03-17T13:36:00.000-07:00

Here is a nice picture by Duncan Davidson of me, speaking at eComm.

The talk went well and the conference was excellent. Very well setup and run as a single stream of short talks. The Computer History Museum in Mountainview was an inspired choice of venue.

My slides in pdf form at at this link.

The talk is a more technical version of the BIL talk I gave a few weeks ago. It is focussed on the next few generations of mobile devices - "The Future In Your Pocket". Some key ideas from the talks and discussions I had on this subject at BIL and eComm08:

Pocket devices double capacity every year (laptop every two years)
Intel will drive power down further to compete with ARM
Intel x86/x64 in my pocket means I can use desktop versions of apps like Skype
Wireless high-definition video out is a key new feature we could use today
It's a server in my pocket, web services, video streaming etc.
More performance at lower power in the future allows always-on services
Ambient presence placing OpenAL 3D audio sources in the back of your head
Stereo audio sunglasses with video camera, heads-up display
Brainwave sensor (See Neurosky) input controller, records your stress/mood
Lifelogging - permanently archive everything you hear, see and feel
Telepathy - real-time, many to many lifelogging - immersive relationships
Virb-ing - virtual/real blending Second Life or WoW into lifelogs and telepathy
The killer app for teenagers in 2010 perhaps?

BIL Talk: Millicomputing - The future in your pocket

2008-02-29T01:12:00.000-08:00

I've written a less technical but more provocative overview of where Millicomputing could go in the mobile space and posted slides as html and slides as pdf.

These will be presented at some point over the coming weekend at the BIL un-conference.

Millicomputing at the BIL Unconference

2008-02-24T13:00:00.000-08:00

You have hopefully heard of the TED conference (Technology Entertainment and Design), its happening in Monterey CA next week, and an impromptu un-conference called BIL is dis-organizing itself as a follow on event in the public park next door starting Saturday March 1st at 11am. I'm going with a bunch of friends, anyone can just turn up, bring your own camp chair, food etc.

Here is the speaker page for millicomputing at BIL

I've signed up for Twitter, since that seems to be a good way to communicate at this kind of event. I'll also try out Twitxr (pronounced twitcher) which is a simple way to collect photos as it happens.

Emerging Communications Conference - EComm March 12-14

2008-02-12T01:39:00.000-08:00

The EComm Conference is a spiritual successor to last year's O'Reilly ETel conference (I was there). It's the brainchild of Lee Dryburgh, who took over when O'Reilly decided not to repeat ETel, and has created a very interesting conference with a lot of good speakers.

Its taking place at the Computer History Museum in Mountainview, CA March 12-14th. The speakers are rapid fire in short time slots, and I'm presenting on Millicomputing on the morning of March 14th for 15 minutes...

I'm going to focus my talk on a hardware roadmap for mobile CPU's and Flash over the next few years, to give people some idea of the capabilities to expect from portable communication devices, and to discuss the battle that is expected as Intel and ARM come at the market from opposite ends of the spectrum.

On Wednesday 12th, the Homebrew Mobile Phone club will be holding a special meeting in the Museum, held jointly with the EComm event.

Samsung S3C6410 mobile processor

2008-02-12T01:35:00.000-08:00

Samsung's latest device runs at 667MHz and includes video capture acceleration that claims to use much less power to compress or decompress video streams.

The device is sampling in Q2 and shipping in volume later in 2008, and there is a lot of speculation going around that this may be the CPU that Apple uses in its next generation 3G capable iPhone.

Intel Silverthorne Details

2008-02-06T23:11:00.000-08:00

It appears to be an interesting step in the right direction, while running in a 500mW to 2W power range, it is a full 64bit (x86/x64) architecture CPU. Its a borderline Millicomputer, but the first mainstream 64bit CPU to get into this space.

The Ars Technica review has all the details.

Looking into the future, Intel is moving in on ARM from above, with a 64bit architecture that it will be able to power-reduce further while keeping all the desktop oriented software investments intact. ARM is coming up from below, with its own legacy of 32bit software that is built for low power and constrained functionality systems. They don't overlap yet, but they will overlap in the next year or so.

Its a millicomputer if your leg doesn't get hot when you have it in your pocket. By that measure, I think Silverthorne isn't quite there yet. I'm waiting for the next step down...

Intel's Low Power Roadmap from CES - Menlow, Moorestown, PATA Z-P140

2007-12-17T09:51:00.000-08:00

Some new information released from Intel emphasizes low power devices with their Menlow architecture for 2008, and Moorestown for 2009/10, and states that Solid State Disks based on Flash are the future. They have a high speed parallel ATA interface (PATA) to a module that is 12x18mm, i.e. around the size of a miniSD format, bigger than microSD. However it runs at 40MByte/s read and 30Mbyte/s write speed with an ATA command set, rather than microSDHC C4 at 13MByte/s or C6 at 20Mbyte/s. The Z-P140 runs at 1.1mW idle, and 300mW operating, and has a 2.5Million hour MTBF.

Menlow looks as if its heading down into the millicomputer territory (i.e. entire CPU and memory less than 1000mW), and moorestown is another 10x reduction in power consumption.

Flash based SSD from Samsung

2007-11-11T11:20:00.000-08:00

A nice review of a Flash based SSD in Engadget. This 64GB drive from Samsung is a drop-in replacement for a 2.5" hard drive. Its fast, but still too expensive for common use.

The Tera-Millicomputer from Sicortex

2007-11-05T00:55:00.000-08:00

If you take a 1GHz, 1 GFLOP MIPS based core that uses 600mW of power, that fits the definition of a millicomputer (depending upon how much RAM you also add). Now if you put 6 of these cores on a single chip, along with DRAM and PCI-Express controllers and a high speed fabric interconnect for 10W it's quite interesting. Packaging almost 6000 cores on an interconnect fabric in a single rack takes it to the logical conclusion. Thats what Sicortex have done.

Even better, they have the *best* web site architecture tour I have ever seen. Its worth browsing it just for fun. Visit the page here, and click on the buttons to animate the architecture. Don't miss the "Linux" button and step through the Hardware Environment starting at the "Node Level" button.

New advances in non-volatile memory - PMC, PRAM and NRAM

2007-10-31T12:56:00.000-07:00

The end is in sight for spinning rust....

While flash based memory is nibbling at the edges of the disk industry, some new techniques are opening up prospects of even greater capacity and speeds in the next few years.

PMC stands for Programmable Metallization Cell, see this MIT review article.

PRAM stands for Phase-change Random Access Memory, some recent news seems to indicate that good progress is being made on PRAM as well as larger and faster Flash memory.

Update:

This SC07 article from The Register states that NRAM stands for Nanotube RAM, the makers claim that they will beat Flash on every metric in a few years time, and several of the big semiconductor companies are looking into the technology.

More on the Cortex A9

2007-10-03T23:59:00.000-07:00

Another update from Ashlee Vance at The Register gives more details on the Cortex A9. They say its a four issue superscalar (the Cortex A8 is dual issue), 8 times the performance of the iPhone CPU, runs at 250mW and should be in devices in around 2010. The A8 was announced in 2005 and three years is a typical lead time for CPU architectures to get into production, so I expect A8 based devices sometime next year, perhaps even a faster iPhone...

There is also some discussion of Intel moving down into this space over the next few years. Excellent! Competition will drive the market to develop faster and I don't personally care what the instruction set is any more (I used to...)

Multicore ARM Chips - Cortex A9

2007-10-03T16:38:00.000-07:00

Performance is cranking up, now we have four core ARM chips on the horizon...

Infoworld article on the announcement.

Each of these cores seems to be based on the dual issue 1GHz Cortex A8 design that was announced a few years ago, and which isn't quite shipping yet in products.

So to put this in perspective, the CPU in the Gumstix Verdex and the iPhone is around 600MHz single issue, the Cortex A8 is around three times the raw performance and the new announcement is about twelve times the raw performance. These all seem to be around the same levels of power consumption, in the few hundred milliwatt range.

I would not expect to have the multicore ARM Cortex A9 in actual products in my pocket for a few years, but its good to have a roadmap into the future of millicomputing.

Qualcomm Scorpion - ARM Cortex based CPU

2007-09-05T08:12:00.000-07:00

The first devices based on the new ARM Cortex generation are announced from Qualcomm, their Scorpion design runs at 1GHz, 250-500mW, and the Cortex based pipeline is dual issue, so the raw instruction issue rate is twice per MHz that of older ARM designs such as the PXA320.

Millicomputing Applications - ETL

2007-08-28T10:39:00.000-07:00

Millicomputers have a very different balance of compute/memory/network/io resources compared to more conventional architectures. They are lower in absolute terms for compute/memory/network, but much higher for random access io.

The performance per watt and the price/performance are very competitive for compute/memory/network, as long as applications can be run at a smaller "grain size". However for io, the aggregate performance of a large number of direct attached flash devices is amazing.

One possible application is from the data warehousing space. Known as ETL, this is the Extract Transform and Load step that pulls data from online transaction processing systems, such as the collection of database back-ends for a web site, and puts it into a form that can be queried to answer questions about the business. There has been a lot of work put into making the ETL processes into decomposable parallel applications, and there is an open source ETL implementation in Java called KETL. KETL was originally written several years ago, when the typical systems of the day were similar in capacity to the millicomputers we have today, so I'm hopeful that the grain size will fit.

KETL is io intensive and it also supports running on a cluster of networked computers, so overall it looks like a plausible fit for an enterprise millicomputer application.

The Future of Millicomputing

2007-08-27T08:57:00.000-07:00

There is a gap between the performance and memory capacity of Millicomputers and mainstream CPUs, that gap is shrinking but how fast, and what are the next steps?

The base technology from ARM can be seen in their Cortex designs. These were disclosed in late 2005, but have yet to appear in actual products. The overall performance is around 3-4 times the performance of the current generation of ARM based devices.

Since Intel sold off their ARM based CPU business to Marvell, it leaves them clear to move their core 32bit x86 platform architecture down into the millicomputing space.

So in the next few years, I expect to see x86 and ARM based system on a chip architectures with overlapping performance and power consumption characteristics in the millicomputing space.

Lower Power x86 Systems

2007-08-24T09:01:00.000-07:00

There is quite a lot of activity in the low power x86 compatible space. The latest CPU from VIA is touted as a 1W CPU, with 0.1W standby power, but when the complete chipset and RAM are added its substantially higher, more like 10W. This article in LinuxDevices.com surveys the whole space very nicely.

The power trend is downwards, these chip sets are aimed at consumer devices, but not in the battery powered space. For commodity devices we can divide by the order of magnitude in power consumption and environmental conditions.

100-1000W Datacenter Server (air conditioned room)

10-100W Home PC/Laptop Space (fan cooled, on when in use, ambient room temp)

1-10W Home consumer devices (fanless, always on, ambient room temp)

100-1000 mW Battery powered millicomputers (always on, cool enough for your pocket)

Interesting technology and product disruptions occur when we mix these spaces. In some ways, the original compute farms that Google built were leveraging the low end home PC power/price/performance point into the datacenter. There are additional opportunities to leverage home consumer devices and millicomputers into the enterprise space.

Millicomputer Performance Benchmarks

2007-07-22T09:30:00.000-07:00

Performance benchmarks for several Millicomputer CPUs are being measured by several Homebrew Mobile Club members at our wiki.

The benchmark used is an old simple CPU benchmark called nbench, which is basically the original Byte magazine benchmark collection from ten years ago. Its very easy to get running. I looked at the industry standard SPEC benchmarks, but they are more vendor oriented and are not freely available. They don't have any useful results posted for ARM architecture systems.

At this point the results are hard to compare since the compiler options in use are somewhat varied and results for recent enterprise server CPUs have not been posted.

Intel's low power plays

2007-06-30T09:58:00.000-07:00

There is an interesting "Power Plays" discussion of Intel's focus on low power at Ars Technica.

There is a lot of work going on to optimize current enterprise server designs to use less power, this is good, but its not the order of magnitude difference that a move to Millicomputing based designs would provide.

The most interesting new technology described is an interconnect that uses very low power and which scales its clock rate and power consumption according to the bandwidth demand. This brings variable capacity to the network layer, and I'd love to see some very low power Intel CPUs with this technology integrated.

The Flashiest Storage for the Millicluster

2007-06-28T07:59:00.001-07:00

Per-module Flash uses the tiny microSDHC format which is about half an inch square (the picture shown is about three times actual size), see http://www.getflashmemory.info/category/microsdhc/. The older microSD format limits to 2GB (available one off for less than $20 each), and microSDHC expands this limit to 32GB using a FAT32 derived on-card filesystem. At present 4GB cards are available and 8 GByte cards have been announced. Streaming read and write performance for microSDHC is much higher than before at about 20MByte/s. Writes are just as fast as reads, and the file-system automatically avoids wearing out any one location in the flash memory.

There is no seek time! Random access at 1000’s of IOPS is only limited by the device driver efficiency, and will be benchmarked. Raw performance is 112 x 4 GB = 448 Gbytes/RU, 18.8 TB/Rack. 112 x 20 MB/s = 2240 MB/s/RU, 94 GB/s/Rack. The implications for storage performance in general are profound. The reason it is so fast is that the storage capacity is solid state, in a single chip and it is directly connected to the CPU chip. There is nothing getting in the way!

Enterprise Millicomputer Server Comparisons

2007-06-28T07:54:00.000-07:00

To provide a competitive comparison two high end 1U Enterprise servers were priced online at http://www.sun.com - one Opteron and one low power Niagara SPARC. The Sun x4100 Opteron server uses about 400W. Its CPU performance is probably double that of an ARM at the same GHz so 2.8 GHz four cores x 2 = 22.4 GHz. Configured with the maximum of 32 GB RAM results in a $13K list price. The Sun T1000 Niagara uses about 200W of power. Its 1.0 GHz 8 core CPU has 32 threads. Lets call this 32 GHz, which is quite optimistic. With a maximum of 16 GB RAM it has a $15K list price. The Enterprise Millicomputer with OPiuM i.MX31 based modules uses less than 160W, probably much less. 532 MHz x 112 = 60 GHz and 256MB x 112 = 28GB RAM. With modules and flash costing perhaps $130 each, lower cost power supplies etc., a similar price point around $15K seems plausible.

Millicomputer networking has higher network bandwidth and there is a big additional saving as there is no need for an external load balancer appliance.

Millicomputer Storage wins with no contest! 2x146GB disks 240 IOPS vs. ~500000 IOPS, 448 GB Flash.

This is all on paper, actual benchmarks are needed, but the point is that the raw performance looks interesting enough to make it worth running the benchmarks....

Enterprise Millicluster System Specifications

2007-06-27T23:58:00.000-07:00

The specifications below are very tentative, I have tried to be conservative, but this is a paper design at this point, and a real design could end up significantly more or less dense in terms of compute and power usage per rack unit (RU).

The Millicluster board I described is half the width of a typical enterprise motherboard. Its only 0.4" thick, so we can stack four of them high (4x0.4=1.6") in the 1.75" height of a 1U package.

Hence a standard 1U Enterprise Server Package contains Eight Milliclusters. This has a compute density of 112 OPiuM modules per RU, 4704 modules in a 42RU rack. The power consumption peaks under 160 Watts/RU, and idles at less than 24 Watts/RU. Maximum rating would be less than 6.7KW/Rack, which is quite reasonable. The CPU performance totals 60 GHz/RU, 2,520 GHz/Rack. There is 28 GBytes/RU of RAM, 1,172 GBytes/Rack.

The network has 8 Load balancer/bridge-routers per RU with 8 Gbits/RU module bandwidth on 16 redundant Gbit ports. An Ethernet switch could be added to the design to reduce the port count at a cost of a few watts and dollars. For storage a microSDHC flash memory socket at each module would hold a 2 GB microSD for very low cost, 4 GB for capacity, 8 GB in 2008.

There are many optional interfaces that could be used for specific applications. All modules include an ATA disk controller if needed, so each Millicluster could have connectors to support hard disks and DVD-ROM players. For graphics the i.MX31 modules include an OpenGL based 3D graphics accelerator and an LCD display driver with touch screen input. There is a camera input and video compression engine, with stereo audio and video playback. Modules also include multiple USB and serial interfaces.

Enterprise Millicluster Board Layout

2007-06-23T01:38:00.000-07:00

I came up with the board layout shown to indicate how we might arrange the millicomputer modules, interface bridge, USB network and microSD card holders on a conveniently sized board. Its just under half the size of a typical enterprise server motherboard. The total size is about 5.5” Wide x 12” Deep x 0.4” High. The Ethernet network bridge increases the idle power consumption to no more than 3 Watts, and fully active power consumption is unlikely to exceed 20 Watts, so I don't think heat sinks will be needed.

Enterprise Millicluster

2007-06-23T00:49:00.000-07:00

Taking the components I have already menioned in previous posts, we can assemble them into a small cluster that seems to be a useful size and specification for an Enterprise Server building block. Its a cluster of millicomputers, so we may as well coin the name Millicluster as we go along (and register the millicluster.com etc. domains to point here :-)

Using 8-port USB switches, we could lay out 14 i.MX31 based Millicomputer modules behind a PPC440EPx based Ethernet Bridge that runs Linux so it is general purpose, but it will be pre-configured as a Load Balancer. This gives us a 1 Gbit/sec redundant network (it has two 1 Gbit links in, but only two 480Mbit links to the millicomputer modules). There is a total of 7.5 GHz of CPU, 3.5 GBytes of RAM, and 56 Gbytes of Storage using 4 GByte microSDHC flash memory cards on each Millicomputer.

This depends upon having a high speed 8-port switch, and so far I have found some products from Belkin and D-link that have one upstream port and seven downstream. I'm not sure what chipset they use, but they are inexpensive and have been available for a few years, so this seems reasonable.

x86 Millicomputers are on the way

2007-06-08T08:08:00.000-07:00

News about the x86 architecture working its way down into the millicomputer space via the Pico ITX. This board is 1.8x3" with a 1GHz CPU and 256 or 512MB RAM, but it doesn't give power consumption. However they are positioning it for use in mobile devices so its likely to end up under a watt once they get the chip count down a bit further.

PowerTop monitoring tool

2007-05-18T00:39:00.000-07:00

New tool for Intel/Linux systems monitors power use on a per-process basis. This is excellent, as a few months ago I decided I wanted to have a tool like this, and this saves me from writing it!

There is also a new NO_HZ feature in the latest Linux kernel 2.6-21 that stops the clock when the CPU is idle. PowerTop can be used to see how much power that saves.

microSD and microSDHC Storage for Millicomputers

2007-05-17T00:15:00.000-07:00

Its been a while since I last posted here, mostly due to my change in employer, as mentioned on my personal blog... Now that is settled, I have a backlog of things to discuss here. Also, I'll be at the Techshop pavillion at Maker Faire this Saturday and Sunday afternoons.

There is an interesting price graph for 2GByte microSD at Nextag. It shows an introduction at about $120 last September, trending down to about $25 now. Amazon has it for $25 and Amazon Marketplace for $16.50. Meanwhile, larger devices are shipping using the microSDHC format at 4GByte, and announcements have been made for 8GByte. The microSD format gives a performance of around 20MBytes/s, so for 2KB reads that works out at a maximum of 10,000 random reads/second. Writing at 6MB/s works out at a maximum of 3000 random 2KB writes/s. Even allowing for operating system overhead, several thousand IOPS should be possible at under ten cents per IO per second and $10/GByte.

Of course, most millicomputer CPUs interface directly to devices like microSD, so there is no support device or interface needed. Just a microSD carrier or two, per millicomputer module.

Millicomputer Based Load Balancers

2007-04-17T08:51:00.000-07:00

If we build systems that contain hundreds of modules for web based applications, we need a way to manage the workflow distribution for incoming network traffic. Commercial load balancers cost more than millicomputing modules we want to send load to, so I've been looking around for open source projects that implement various kinds of load balancing. I have found a very good detailed summary article on this subject by Willy Tarreau, author of HAproxy, which he describes as:

HAProxy is a free, very fast and reliable solution offering high availability, load balancing, and proxying for TCP and HTTP-based applications. It is particularly suited for web sites crawling under very high loads while needing persistence or Layer7 processing. Supporting tens of thousands of connections is clearly realistic with today's hardware. Its mode of operation makes its integration into existing architectures very easy and riskless, while still offering the possibility not to expose fragile web servers to the Net...

At the http/application level, I found a description of a simple but powerful tool called balance.

Balance is our surprisingly successful load balancing solution being a simple but powerful generic tcp proxy with round robin load balancing and failover mechanisms. Its behaviour can be controlled at runtime using a simple command line syntax.

Another http load balancer that claims high performance and more features is XLB. It states

XLB is a high performance HTTP load balancer. connection management, caching, ssl, scripting. 300 mbit/sec / 4000 reqs/sec takes 30% cpu on a 2GhZ Xeon. connection pooling to backend servers reduces memory and cpu usage on backends.

One problem with load balancers, is that if they fail, a potentially large number of modules would be out of action. The Ultra Monkey load balancer addresses this issue.

Ultra Monkey 3 makes use of The Linux Virtual Server (LVS) to provide fast load balancing. The Linux-HA framework is used to monitor the linux-directors - the hosts running LVS and doing the load balancing. This is combined with ldirectord which monitors real-server - the hosts that accept end-user's connections. These three core components allow Ultra Monkey 3 to provide highly available and/or load balanced network services.

I haven't used any of these options, so I'm very interested to get recommendations, please comment if you have experience or alternatives to share, and I'll update this post.

In the array of modules scenario, I would dedicate a few modules to provide load balancing services. If the modules are all connected via Ethernet, then any module can be used. If we use the USB network then the central USB master that provides an Ethernet gateway is the natural place to install load balancer services.

Gumstix initial bringup

2007-04-17T00:13:00.000-07:00

I bought a Gumstix Verdex GS-270, along with a small motherboard that has serial, USB and power connectors. For initial bringup I also installed Ubuntu 6.10 Linux on a PC to act as my development host. I've figured out how to get logged in to Linux on the Gumstix, and I'm documenting it step by step here.

This was my goal!

# uname -a
Linux gumstix 2.6.18gum #1 Wed Feb 28 18:05:43 PST 2007 armv5tel unknown

The basic sequence included getting at the serial port on the Dell, configuring it correctly, and figuring out which of the two serial ports on the motherboard has the console output.

Download Ubuntu 600MB CD image - I used OSX, then used Disk Utility to burn it to a CD, and installed it on the PC, quite straightforward.
Ubuntu doesn't include comms by default. I ran aptitude to search for programs and found a comms package that includes mincom and cu, so I installed both of them.
The gumstix Wiki eventually revealed these setup instructions, which are to use mincom, turn off hardware and software flow control and set 115200-N-1 mode.
This picture shows a similar motherboard, with the console port connected to the second serial port, which also worked for me.
I connected the serial and USB cables, plugged in the power supply, and a small green LED glowed on the motherboard, nice confirmation that its on.
After watching various boot messages, I logged in as root, with the initial password gumstix.

Sounds simple, but as usual, this took quite a while for me to figure out from scratch.... There are also instructions on how to interface and develop using Windows or OSX, but I wanted to do some comparative benchmarking of PC vs. ARM running similar releases of Linux 2.6.

Configuration messages at boot:

U-Boot 1.1.4 (Mar 1 2007 - 17:10:55) - PXA270@600 MHz - 1321

*** Welcome to Gumstix ***

U-Boot code: A3F00000 -> A3F25850 BSS: -> A3F5AE70
RAM Configuration:
Bank #0: a0000000 128 MB
Flash: 32 MB

.... some more messages then:

Linux version 2.6.18gum (craig@azazel) (gcc version 4.1.1) #1 Wed Feb 28 18:05:7
CPU: XScale-PXA270 [69054117] revision 7 (ARMv5TE), cr=0000397f
Machine: The Gumstix Platform
Memory policy: ECC disabled, Data cache writeback
Run Mode clock: 208.00MHz (*16)
Turbo Mode clock: 624.00MHz (*3.0, active)
Memory clock: 104.00MHz (/2)
System bus clock: 104.00MHz
CPU0: D VIVT undefined 5 cache
CPU0: I cache: 32768 bytes, associativity 32, 32 byte lines, 32 sets
CPU0: D cache: 32768 bytes, associativity 32, 32 byte lines, 32 sets

Interesting information on 32-way cache associativity, which I did not see mentioned in the specs.

The 32MB flash memory is mounted as a filesystem, with 8MB taken up by the default installation.

# df
Filesystem Size Used Available Use% Mounted on
/dev/mtdblock1 31.8M 8.0M 23.8M 25% /

The system supports IP networking over USB, which I have plugged in but I don't have working yet (its supposed to come up automatically, but doesn't). Thats next.

Vertical and Horizontal Module Arrangements

2007-04-12T22:44:00.000-07:00

Modules are available with edge connectors that can be mounted in bulk on a mother board as shown in the image below. The dimensions match the standard motherboard found in 1U Enterprise server designs, about 12x13 inches. The diagram shows 120 modules, but its quite likely to be possible to pack them in more densely than this.

The alternative is to mount modules flat on the boards as shown in the second diagram. This has the same 12x13 inch area, but is a very thin board, and at least four of them could be stacked in a 1U package, which also comes out to 120 modules.

In practice these board sizes and layouts will need to be adjusted to take into account the mechanical problems of flexing, mounting, cable routing etc. In each case the power and cooling management should be relatively simple, since there is a total peak power of around 100 watts for the entire 1U package, and no localized hot spots.

Some of the module designs have built-in temperature sensors and they all have power voltage sensors, so they can detect and report on environmental conditions across the motherboard.

Millicomputer Module Interconnects

2007-04-11T11:02:00.000-07:00

There are two basic approaches.

One is to get modules that have ethernet built-in (or to add ethernet interfaces to a motherboard) and use ethernet switch chips such as the 8-24 port solutions from Vitesse to cluster the modules together. The individual modules would connect at 100Mbit, and the switches and external interfaces would interconnect at 1Gbit. The single chip ethernet switches have lots of features but can be run as unmanaged devices, so there is very little software needed to implement or manage the network. By directly connecting the networks on a motherboard there is no need to drive the full physical ethernet wire standard between the devices, saving a lot of power. These devices cost a few dollars a port, and dissipate about half a watt per port for fully driven gigabit links. if we can avoid using the Ethernet "PHY" (physical driver) a lot more power can be saved.

Another option is to use the built-in high speed USB2.0 interfaces which run at up to 480Mbit/s and connect them to a USB based central router that has ethernet support, then run IP over USB. This is a bit more complex to implement, but could be faster, lower power and cheaper since it uses an interface that is directly built into the millicomputer CPU. There are other kinds of devices like the AMCC PPC440EPx that are more PC-like, and have ethernet, PCI-bus and high speed USB built-in that could be used to implement a board level controller/router/interface. This device is more powerful than the mobile oriented millicomputer CPUs but dissipates about 3W so its in the next bracket up from a power consumption viewpoint.

PXA270 Module for testing

2007-04-11T10:57:00.000-07:00

I just ordered a Gumstix GS270-XL6P module with 600MHz PXA270 and 128MB RAM. I'll run benchmarks on it, then end up building it into one of the mobile phone designs I'm working on. More later...

Millicomputer Module Specifications

2007-04-05T18:13:00.000-07:00

Here is another Google spreadsheet table of millicomputing module specifications.

There are several approaches, but some of these are edge connector based, include on-board ethernet, and could be stacked on a motherboard in a very dense array.

I think that on a standard 1U motherboard, if we could get five rows of 24 connectors that is 120 individual modules, using less than 100W maximum. The motherboard would just need to provide power and ethernet switch chips. If we also want per node storage, there are many very dense NAND flash ships in the multi-Gigabyte range that could be added to the design.

So is that interesting? I think so...

Millicomputer CPU Specifications

2007-04-03T16:40:00.000-07:00

I've started a table of CPU specifications as a google spreadsheet.

I'm mostly interested in the CPU clock rate, CPU caches, RAM bandwidth and size.

All these devices are very flexible, and are mostly configured with relatively small amounts of memory for embedded applications. However, they have a decently fast clock rate, and can interface to at least two SDRAM chips. These chips are 32bits wide and currently contain a total of 128MB each. The CPUs support up to 256MB per chip, so the next generation SDRAM devices can double overall capacity.

Compared to current "enterprise CPUs" they are much slower than Opterons but probably comparable to a single thread on a Niagara.

The next comparison table I want to put together is for board level devices, such as the Gumstix range.

Freescale i.MX31 Overview

2007-03-30T18:13:00.000-07:00

The full specifications are available on the web, without needing an NDA. Freescale should be commended for their open attitude.

See below for the block diagram of the functions of this device, taken from the Freescale web site, a picture really is worth a thousand words in this case...

All this and the price is about $20 in small volume!

Flash Solid State Disk (SSD) for Millicomputers

2007-03-29T23:26:00.000-07:00

The days of keeping bits on spinning rust are coming to an end....

Both Samsung and SANdisk have announced 32GB SSDs. SANdisk's comsumes 0.9W max (competing disks take 1.9W) and fits in a 1.8 or 2.5" drive form factor with ATA interface. The sequential performance of these SSDs is similar to normal disks for reads, a bit slower for pure writes, but as soon as you start doing random reads or writes they are an order of magnitude faster than disks. The smaller the random accesses the bigger the relative speedup. The latest announcement from Samsung is a 1.8" 64GB version, and there is some discussion about the growth of this market in the press release.

This makes perfect sense for millcomputers. Small millicomputers can be directly connected to gigabytes of NAND flash via the SDIO interface, and larger millicomputers can use ATA interfaces to connect to flash-SSDs. The extra random performance of the SSD offsets the lack of disk spindles in a compact design and will make IO intensive workloads extremely competitive for millicomputing.

The MTBF (reliability) of SSDs is also far higher than disks. A mirrored pair of disks may be replaced with a single SSD since it has much higher reliability. This helps offset the current price premium paid for the SSD.

In the past SSDs have been built using technologies that were far more expensive than disks. Flash based SSDs have now reduced the gap, and the trend is that SSDs will eventually become bigger and cheaper than disks, the only question is when, and my answer is sooner than you think!

Update: here is a detailed benchmark review from Tomshardware.com.

What is a Millicomputer? Why talk about Millicomputing?

2007-03-29T13:28:00.000-07:00

While researching devices for my home brew mobile phone, I've realized that the current generation of CPUs for mobile devices are actually seriously powerful, very low cost and use almost no power. The performance per watt and per dollar seems to be an order of magnitude better than the PC-class CPUs that are common in commodity servers nowadays. The absolute performance and memory capacity is lower, but is comparable to common PC hardware from a few years ago, and could be useful for more than running a high end phone or portable games machine. Devices such as the Marvel PXA270 and Freescale i.MX31 run at over 500MHz, some include floating point units, they support at least 128MB of RAM (a single chip), and a myriad of I/O interfaces, with Linux 2.6 support.

While the current mainstream CPUs were driven by the development of the home PC market, this generation is driven by the development of the mobile, battery powered device market, which is a very large. For example the worldwide cellphone market is something like a billion devices a year.

I think that there could be some interesting general purpose computer systems built from low power devices (CPUs that use less than one watt). I looked around but wasn't sure what to search for... I do know about the systems that are sold for embedded use, but they are typically configured using lower speed and lower memory options.

Does anyone know of vendors selling general purpose millicomputer based systems?

I need a name for this class of system, so I'm going to call them Millicomputers, and I'm going to explore this area in public on this blog, and using the principles of open hardware that we have adopted for the homebrew mobile phone club, I expect to help build some.

I originally asked this question on my main blog, and asked a lot of people in person, but didn't find a pre-existing name or any objections to this concept.