Stephen O’Grady, a founder at the technology analyst company RedMonk, said the technology industry often has swung back and forth between more standard computing systems and specialized gear.
A little tip of the hat to Andrew Feldman, CEO of SeaMicro the startup company that announced it’s first product last week. The giant 512 cpu computer is being covered in this NYTimes article to spotlight the ‘exotic’ technologies both hardware and software some companies use to deploy huge web apps. It’s part NoSQL part low power massive parallelism.
From where I stand, the SM10000 looks like the type of product that if you could benefit from having it, you’ve been waiting for something like it. In other words, you will have been asking for something like the SM10000 for quite a while already. SeaMicro is simply granting your wish.
This announcement that has been making the rounds this Monday June 14th has hit Wired.com, Anandtech, Slashdot, everywhere. It is a press release full court press. But it is an interesting product on paper for anyone who is doing analysis of datasets using large numbers of CPUs for regressions or large scale simulations too. And it is at it’s core virtual Machines, with virtual peripherals (memory, disk, networking). I don’t know how you benchmark something like this, but it is impressive in its low power consumption and size. It only takes up 10U of a 42U rack. It fits 512 CPUs in that 10U area as well.
Imagine 324 of these plugged in and racked up
This takes me back to the days of RLX Technologies when blade servers were so new nobody knew what they were good for. The top of the line RLX unit had 324 CPUs in a 42U rack. And each blade had a Transmeta Crusoe processor which was designed to run at a lower clock speed and much more efficiently from a thermal standpoint. When managed by the RLX chassis hardware and software and paired up to an F5 Networks load balancer BIG-IP, the whole thing was an elegant design. However the advantage of using Transmeta’s CPU was lost on a lot of people, including technology journalists who bashed it for being too low performance for most IT shops and data centers. Nobody had considered the total cost of ownership including the cooling and electricity. In those days, clock speed was the only measure of a server’s usefulness.
Enter Google into the data center market, and the whole scale changes. Google didn’t care about clock speed nearly as much as lowering its total overall costs for its huge data centers. Even the technical journalists began to understand the cost savings of lowering the clock speed a few hundred megahertz and placing servers more densely into a fixed sized data center. Movements in the High Performance computing also led to large scale installations of commodity servers being all bound together into one massively parallel super computer. More space was needed for physical machines racked up in the data centers. Everyone could see the only way to build out was to build more data centers, build bigger data centers or pack more servers into the existing footprint of current data centers. Manufacturers like Compaq got into the Blade server market, along with IBM and Hewlett Packard. Everyone engineered their own proprietary interfaces and architectures, but all of them focused on the top of the line server CPUs from Intel. As a result, the heat dissipation was enormous and the densities of these blade centers was pretty low (possibly 14 CPUs in a 4U rack mount).
Look at all those CPUs on one motherboard!
IBM began to experiment with lower clocked PowerPC chips in a massively parallel super computer called the Blue Gene. In my opinion this started to change people’s belief about what direction data center architectures could go. The density of the ‘drawers’ in the Blue Gene server cabinets is pretty high. Lot more CPUs, power supplies, storage and RAM in each unit than in a comparable base level commodity server from Dell or HP (the previous most common building block for the massively parallel super computers). Given these trends it’s very promising to see what Seamicro has done with its first product. I’m not saying this is a super computer in a 10U box, but there are plenty of workloads that would fit within the scope of this server’s capabilities. And what’s cooler is the virtual abstraction of all the hardware from the RAM, to the networking to the storage. It’s like the golden age of IBM machine partitioning and Virtual Machines but on an Intel architecture. Depending on how quickly they can ramp up production and market their goods, Seamicro might be game changer or it might be a takeover target from the likes of HP or IBM.
Some people may remember the poorly marketed and badly implemented Microsoft ReadyBoost technology hyped prior to the launch of Windows Vista. Microsoft’s intention was to speed throughput on machines without sufficient RAM memory to cache large parts of the Windows OS and shared libraries. By using a small Flash memory module on
Intel Turbo Memory to be used as ReadyDrive storage cache
the motherboard (Intel’s Turbo Memory) or by using a USB connected Flash memory stick one could create a Flash memory cache that would offset the effect of having 512MB or less RAM installed. In early testing done by folks like Anandtech and Tom’s Hardware system performance suffered terribly on computers with more than the 512MB of RAM targeted by Microsoft. By trying to use these techniques to offset the lack of RAM on computers with more than 512MB of RAM the computers ran slower using Vista. I had great hopes ReadyBoost at the time the flash cache method of speeding throughput on a desktop PC was heralding a new early of desktop PC performance. In the end it was all a myth created by the Microsoft marketing department.
Some time has passed since then Vista was released. RAM prices have slowly gone down. Even low end machines have more than adequate RAM installed to run Vista or now Windows 7 (no more machines with 512MB of RAM). The necessity of working around those limits of RAM is unnecessary. However total system level I/O has seen some gains through using somewhat expensive Flash based SSD (solid state disks). Really this is what we have all been waiting for all along. It’s flash memory modules like the ones Intel tried using for it’s ReadyDrive capable Turbo Memory technology. However these were wired into a PCIe controller and optimized for fast I/O, faster than a real spinning hard disk. The advantage over the ReadyBoost was the speed of the PCIe interface connected to the Flash memory chips. Enterprise data centers have begun using some Flash SSDs as caches with some very high end product using all Flash SSDs in their storage arrays. The entry level price though can be daunting to say the least. 500GB SSD disks are the top of the line, premium priced products and not likely to be sold in large quantity until the prices come down.
Seagate is now offering a product that has a hybrid Flash cache and spinning disk all tied into one SATA disk controller.
Seagate Momentus XT
The beauty of this design is the OS doesn’t enter into the fray. So it’s OS agnostic. Similarly the OS doesn’t try to be a disk controller. Seagate manages all the details on its side of the SATA controller and OS just sees what it thinks is a hard disk that it sends read/write commands. In theory this sounds like a step up from simple spinning disks and maybe a step below a full flash based SSD. What is the performance of a hybrid drive like this?
As it turns out The Register did publish a follow-up with a quick benchmark (performed by Seagate) of the Seagate Moments XT compared to middle and top of the line spinning hard drives. The Seagate hybrid drive performs almost as well as an the Western Digital SSD included in the benchmark. That flash memory caches the stuff that needs quick access, and is able to refine what it stores over time based on what it is accessed most often by the OS. Your boot times speed up, file read/write times speed up all as a result of the internal controller on the hybrid drive. The availability if you check Amazon’s website is 1-2months which means you and I cannot yet purchase this item. But it’s encourage and I would like to see some more innovation in this product category. No doubt lots of optimization and algorithms can be tried out to balance the Flash memory and spinning hard disks. I say this because of the static ram cache that’s built into the Momentus XT which is 32MBytes in size. Decide when data goes in and out, which cache it uses (RAM or Flash) and when it finally gets written to disk is one of those difficult Computer Science type optimization problems. And there are likely as many answers as there are Computer Scientists to compute the problem. There will be lots of room to innovate if this product segment takes hold.
Great article and lots of hardcore important details like drivers and throughput. It’s early days yet for the PCI based SSDs, so there’s going to be lots of changes and architectures until a great design or a cheap design begins to dominate the market. And while some PCIe cards may not be ready for the Enterprise Data Center, there may be a market in the high end gamer fanboy product segment. Stay Tuned!
The next step up from a regular sata based Solid State Disk is the PCIe based solid state disk. They bypass the SATA bottleneck and go straight through the PCI-Express bus, and are able to achieve better throughput. The access time is similar to a normal SSD, as that limit is imposed by the NAND chips themselves, and not the controller. So how is this different than taking a high end raid controller in a PCIe slot and slapping 8 or 12 good SSDs o … Read More
For many years it has been possible to make your own DVDs with free software tools. Over the course of the past decade, DVD creation evolved from the exclusive domain of the media publishing companies to something basically anyone could do on their home computer.
The move towards Blu-ray encoding is very encouraging. In reading the article I don’t see a mention of CUDA or OpenCL acceleration of the encoding process. As was the case for MPEG-2 a glaring need for acceleration of the process was painfully obvious once people started converting long form videos. I know x264 encoding can be accelerated by splitting threads across CPUs on a multi-core processor. But why not unleash the floodgates and get some extra horsepower from the ATI or nVidia graphics card too. We’re talking large frames and large frame rates and the only way to guarantee adoption of the new format is to make the encoding process fast, fast, fast.
Last year, Samsung told the world it had teamed with Instrinsity on a 1GHz ARM chip known as the Hummingbird, and Samsung manufactures the ARM chips underpinning the Apple iPhone, a smaller version of the iPad. This has led many to assume that the Hummingbird architecture is the basis for the the A4.
I am sure that Apple’s ability to act quickly and independently helped win them not just design expertise, but an actual nearly finished CPU in the form of the Hummingbird project. There does now seem to be a smartphone Megahertz War similar to the bad old days of desktop computing when AMD and Intel fought it out 1 gigahertz at a time. We will see what comes of this when the new iPhones come out this Summer. A4 may not translate into a handheld cpu form factor. But looking at the iFixit teardown of the iPad makes me think the iPad motherboard is almost the size of a cell phone! So who knows, maybe A4 is scalable down to iPhone as well. We’ll find out in June I’m sure when Apple hosts its Worldwide Developers Conference (WWDC) in San Francisco, CA.
It doesn't matter what device you use, it is all about the software
TomTom is releasing a new personal navigation device (PND) called the TomTom Live 1000. As part of this article from MacNN they mention TomTom is attempting to get into the App Store market by creating its own marketplace for TomTom specific software add-ons (like the Apple App Store). The reason is the cold war going on between device manufacturers gaining the upper hand by wholesale adoption of a closed application software universe. Google is doing it with Android and Apple has done it with the iPhone and iPad. Going all the way back to the iPod, there was interest in running games on those handheld devices, but no obvious way to ‘sell’ them, until the App Store came out. Now TomTom is following suit, by redesigning the whole TomTom universe using Webkit as a key component of it’s new OS on TomTom devices (Webkit is also being used in the Android based Garmin A10 phone too). Ambivalent about the added value? Other than trying to gain some market share against PND manufacturers, Harold Goddijn, the CEO of TomTom says it’s all about innovation. They mention in passing the possibility of Augmented Reality apps for TomTom devices. But there’s a small matter of getting a video feed into the PND that can then be layered with the AR software. And honestly even the CEO Tom Goodjin is somewhat ambivalent about seizing the opportunity of Augmented Reality in the TomTom application store universe. As reported on Pocket-lint.com: “Although Goddjin confirmed that the company was looking at the possibility of adding augmented reality in to the mix, the niche technology isn’t a major objective for them.”
It’s not enough to just overlay information on an Apple iPhone or TomTom PND screen showing related points of interest (POI). Like the iPhone Nearest Tube app from Acrossair, knowing the general compass direction to a subway station is useful. But full step-by-step navigating to it seems to be the next logical step, maps and all. What makes me think of this is the recent announcement of the Garmin A10 smartphone with GPS navigation. If Garmin, TomTom or an independent developer could mashup Augmented Reality with their respective navigation engines, whilst throwing in a bit of Google Street View one might, just might have the most useful personal assistant for finding places on foot. Garmin has a whole slew of devices for the hiking, and bicycling market. They even offer walking/pedestrian directions on their automobile navigation devices. So the overlay of Augmented Reality/Point-of-Interest and full-on Garmin Navigation to me would be a truly killer app.
Google now denies they are releasing Google Maps for the iPhone. Take this whole article with a big grain of salt. I’m just glad I didn’t post this immediately after reading the original article on AppleInsider. By the end of the day last Friday April 23rd, Google was denying the rumor already. The moral of the story is look before you leap.
Google this week said it plans to bring Google Maps Navigation, its free turn-by-turn GPS software already available on Android, as an application for Apple’s iPhone.
According to MacUser, Google officials confirmed at a London, England, press conference that its satellite navigation software would be coming to “other” platforms, including the iPhone. No dates for potential availability were given
April 23rd 2010 Amazing, just when you thought you knew what was going on Google can come along and change things entirely. I ask how does one give away GPS navigation? Four years ago this couldn’t have been conceived or dreamed of given the market for GPS navigation. Now, meh, just give it away. I wish Google great success as this is almost compelling enough for me to get an iPhone now.
Previously you needed an Android based Smartphone usually available only on Verizon. But now there’s a multitude off choices, Garmin’s A50 is coming to AT&T and has my favorite navigation interface along with Google Maps if you want it. Which is what I would prefer. I’m hoping Garmin continues to evolve this to integrate any and all live data it cannot incorporate with its stand alone navigation units which to date don’t have live internet connections (whereas TomTom Live! units do). I’m most interested in any live data that might benefit me in a sudden traffic jam or a new Point of Interest not compiled since the last download/update to the navigation software.
Google’s entry to the iPhone navigation arena would force TomTom, Navigon, and now Garmin all to take heed and compete more vigorously especially since Google would be giving its software away. I’m guessing they could promote themselves as being advertisement free alternatives to the Google Maps Navigation?
I wonder: Is there an opportunity for Alan Kay’s Dynabook? An iPad with a Sqeak implementation that enables any user to write his or her own applications, rather than resorting to purchasing an app?
Apple earlier this month instituted a new rule that also effectively blocks meta-platforms: clause 3.3.1, which stipulates that iPhone apps may only be made using Apple-approved programming languages. Many have speculated that the main target of the new rule was Adobe, whose CS5 software, released last week, includes a feature to easily convert Flash-coded software into native iPhone apps.
Some critics expressed concern that beyond attacking Adobe, Apple’s policies would result in collateral damage potentially stifling innovation in the App Store. Scratch appears to be a victim despite its tie to Jobs’ old friend.
What a difference 3 days makes right? Tom’s Hardware did a great retrospective on the ‘originality’ of the iPad and learned a heck of a lot of Computer History along the way. At the end of the article they plug Alan Kay’s Squeak based programming environment called Scratch. It is a free application that is used to create new graphical programs and is used as a means to teach mathematics problem-solving through writing programs in Scratch. The iPad was the next logical step in the distribution of the program, giving kids free access to it whenever and on whatever platform was available. But two days later, the announcement came out the Apple App Store, the only venue by which to purchase or even download software onto the iPhone or the iPad had roundly reject Scratch. The App Store will not allow it to be downloaded and that’s the end of that. The reasoning is Scratch (which is really a programming tool) has an interpreter built-in which allows it to execute the programs written within its programming environment. Java does this, Adobe Flash does this, it’s common with anything that’s like a programming tool. But Apple has forbidden anything that looks, sounds, or smells like a potential way of hijacking or hacking into their devices. So Scratch and Adobe Flash are now both forbidden to run on the Apple iPad. How quickly things change don’t they especially if you read the whole Tom’s Hardware article. Alan Kay and Steve Jobs are presented as really friendly towards one another.
As a phone, Garmin’s entry occupies the lower mid-range with a three-megapixel camera, native T-Mobile 3G and Wi-Fi. Built-in storage hasn’t been mentioned but should be enough to carry offline maps in addition to the usual app and media storage.
After it’s first attempt to create a Garmin branded phone called the G60, Garmin is back once again with the A50. But this time making a much more strategic choice by adopting an open platform: Google’s Android phone OS. I wrote about Garmin’s response to the coming Smartphone onslaught to it’s dominance of the GPS navigation market. This was after I read this article in the NYTimes: Move Over GPS, Here Comes the Smartphone – (July 8, 2009). At that time Navigon which had been in the market for GPS navigation, dropped out and went to software only licensing to device manufacturers. Whispers and rumors indicated TomTom was going to license its software as well. By Fall 2009 TomTom had shipped an iPhone version of its product. It looked like a form of paradigm shift that kills an industry overnight. GPS navigation was evolving to a software only industry. Devices themselves were better handled by the likes of Samsung, Apple, etc. When the Garmin nuviphone finally reached the market, the only review I found was on Consumer Reports. And they were not overly positive in touting what the phone did differently from a a standalone navigation unit. And worse yet, they had spent two years in development of this device only to have it hit the market trumped by the TomTom iPhone App. It was a big mistake and likely to make Garmin more wary of trying another attempt at making a device.
Hope springs eternal it seems at Garmin. They have taken a different tack and are now going the open systems route (to an extent). It seems they don’t have to invent everything themselves. They can still manufacture devices and provide software, but they don’t have to also create an OS that allows things to be modularly integrated (Phone and GPS) and given that they chose Android, things can only get better. I say this in part because over time it has become obvious to me Google is a real fan of GPS navigation and certainly of Maps.
When I bought my first GPS unit from Garmin, I discovered that you can save out routes direct from Google Maps into a format that a Garmin GPS receiver can use. I know in the past Garmin forced it’s users to first purchase a PC application that allowed you to plan and plot routes then save them back to your receiver. Later it was made less expensive and eventually it was included with the purchase of new units. I’ve seen screen shots of this software and it was clunky, black and white, and more like a cartography mapping program than a route planner. On the other hand, Google Maps was as fast and intuitive as driving your car. You click on a start point, and end point and it would draw the route right on top of the satellite photos of your route. You could zoom in and out and see, actually see points of interest on your route. It seems in one stroke Google Maps stole away route planning from Garmin.
In the intervening time Google also decided to get in the Smartphone business to compete with Apple. Many of Google’s web apps are accessed through iPhones, so why not tap into that user base who might be willing to adopt a device from the same people running the datacenter and applications hosted in them? It might not be a huge number of users, but Google has money and time and can continuously improve anything it does until it becomes the most competitive player in a market it has chosen to compete in. Tying this all together one can see the logical progression from Google Maps to Google Smartphone. And even Google came up with some prototypes showing what this might look like:
Google made a video showing how Google Maps, and Streetview could be integrated on an Android 2.0 device. And it looked good. It was everything someone could have wanted, navigation, text to speech directions, the ability to zoom in and out, go to Streetview to get an accurate photo of the street address. There were some bits of unpolished User Interface that they still needed to work on. But prototypes and demos are always rough.
The video they posted led me to believe I would stick to my Garmin device, as it still had some logical organization that it would take years for Google to finally hit upon. My verdict was to wait and see what happened next. With Garmin’s announcement today though, things are even a little more interesting than I thought they would be. I can’t wait to see the demo of the final device when it ships. I definitely want to see how they integrate the navigation interface with the Web based Google Maps. If they’re separated as different Apps, that’s okay I guess but a Mashup of Garmin navigation and Google Maps with Streetview would be a Killer App. Mix in live network connection for updates on traffic, construction, and Points of Interest and there’s no telling how high they will fly. Look at this video from MobileBurn.com :
Now all I need is a robot chauffeur to drive my car for me.
Carpet Bomberz Inc. 