Back in April I wrote an article about various potential disruptors built on the advance of cheap CPU, cheap disk space and faster, cheaper wireless mobile or cellular broadband.
The article was based on the theme of feature and performance overshoot by the mainstream vendors. In their pursuit of the highest margin, price-insensitive and therefore most demanding consumers, suppliers of technology overshoot the mainstream requirements.
This allows low end disruptors a foothold in the market, offering significantly lower price products which just fulfill consumer requirements, or more importantly for the disruptor fulfill a new need within the product area which isn't currently serviced by the major players.
A good example of overshoot is the current PC. Most if not all are more than fast enough for all applications apart from computing gaming. The major vendors Intel and AMD (and Sun) have moved from packing more transistors per CPU per nanometer to having more CPUs per chip, Dual Core and Quad Core for Intel and AMD and Niagara for Sun.
Interestingly enough this change in architecture is starting to disrupt the major leaders in multicore CPUs i.e. the graphics card vendors Nvidia and ATI.
Most graphics cards can have up to 128 mini CPU Cores per card. Many people are starting to use their massive parallel computation abilities in areas outside of rendering pictures and video for games and movies. The vendors have released Software Development Kits (SDKs) to help this new usage and spurn additional demand for their hardware.
This is a case where the technology was complex, the language hard to develop code in and now has changed to approaching a standard module/interface, which is making it easier to develop code. So going from closed and integrated to being open and potentially modular.
So how are the CPU players Intel and AMD disrupting the graphic cards. By pursuing open interfaces on the motherboard to talk to their CPUs and by adding graphics cores to their CPU architecture.
It used to be that onboard graphics equaled poor to substandard graphic abilities, hence the requirement for dedicated graphics cards. As time progresses the CPU vendors can add more cores to the CPUs, some which can be dedicated graphic cores.
The other disruptor in this open interface is FPGA (customized CPU cards) which can optimize in hardware what originally was done purely in software on generalized CPUs.
Have Fun
Paul
DisruptionCity
Developing an understanding of the theory of innovation
Monday, August 17, 2009
Tuesday, April 21, 2009
Oracle integrates towards a new mainframe
With Oracle announcing to acquire Sun (which owns MySQL) their strategy of integrating the enterprise IT stack continues. Now they have two OS products (Oracle Linux) and Sun Solaris. They also now own Java.
I wrote an article in February 2007 that one of the outcomes of Open Source Software (OSS) was the fragmentation of the enterprise IT stack. People could pick and choose the software used. This has lead to increased complexity and Oracle is playing an integrator role now. Basically Oracle can approach businesses and say they can reduce the complexity.
Reading various MySQL blogs and websites, there is a grave concern that Oracle will just let MySQL wither on the vine. I see Oracle using MySQL as a way to allow people to start with MySQL and eventually convert to Oracle Databases as their businesses and database requirements grow. Oracle already owns Innodb, which is the most popular transactional engine for MySQL. Working on making databases customers migration from MySQL to Oracle seamless would be good step for starters.
What does that mean for MySQL?
Given the code is available new databases will appear as forks of that code. Ironically all that will do is increase the amount of complexity in choice for databases.
I wrote an article in February 2007 that one of the outcomes of Open Source Software (OSS) was the fragmentation of the enterprise IT stack. People could pick and choose the software used. This has lead to increased complexity and Oracle is playing an integrator role now. Basically Oracle can approach businesses and say they can reduce the complexity.
Reading various MySQL blogs and websites, there is a grave concern that Oracle will just let MySQL wither on the vine. I see Oracle using MySQL as a way to allow people to start with MySQL and eventually convert to Oracle Databases as their businesses and database requirements grow. Oracle already owns Innodb, which is the most popular transactional engine for MySQL. Working on making databases customers migration from MySQL to Oracle seamless would be good step for starters.
What does that mean for MySQL?
Given the code is available new databases will appear as forks of that code. Ironically all that will do is increase the amount of complexity in choice for databases.
Sunday, April 19, 2009
Good enough means disruption is on the way
I was reading this article from techradar about many different technologies being good enough and in overshooting the requirements for many users (but not the most demanding)
http://www.techradar.com/news/world-of-tech/what-if-our-tech-is-good-enough--589169?src=rss&attr=all
From reading the Innovators Dilemma and the following books, this kind of overshooting of requirements means that there is an area of technology ripe for a disruptor.
The article mentions the following technologies
There is little wonder that the netbooks (smaller laptops) are mentioned. This combines several of the disruptors above. The smaller weight, and less performance means less weight and longer battery life, which dovetails perfectly with the notion of always online applications from anyway.
I have a Dell Mini 9 and it weighs less than some of the books I have. I actually replaced by older laptop for work with this. I can upgrade to 2 Gig and Windows XP works perfectly fine, and the SSD makes IO reads super fast.
The perfect computer device would be a device which has a screen the size of A4, that foldes to A5 (in half), has a touch screen, so you can type on the screen with a full size keyboard and has 3G internet in built, with a battery life around 3-4 hours.
Have Fun
http://www.techradar.com/news/world-of-tech/what-if-our-tech-is-good-enough--589169?src=rss&attr=all
From reading the Innovators Dilemma and the following books, this kind of overshooting of requirements means that there is an area of technology ripe for a disruptor.
The article mentions the following technologies
- Blu-Ray and DVD (Video storage media)
- Windows Vista and Windows 7 (Computer OS)
- Dual and soon multi-core PCs (Computer hardware)
- Faster broadband internet (Internet, connectivity)
- Video storage media (and distribution channel) are disrupted by media PCs and massive low cost storage in the form of Hard drives and soonish Solid State Disk (SSD) technology. Plus the distribution channel is hammered by the ease of downloading material off the net.
- When does having drop shadows and animated stuff on Computer User interface (UI) make much difference. People still struggle with making computers work, most drop into a browser and the browser may be the disrupter here. If you run your computer purely to use tools and software off the internet, why does the underlining OS mean anything.
- This is similar to point 2. The ability to work almost completely online means that the requirement for performance on the local machine is reduced. This always online is also helped by the disruptor to broadband internet (see point 4). The only holdout in this area is gaming, and the continued ability of game developers to suck every last ounce of performance out of the PC and associated videocards and still leave you with a framerate less than the best.
- I had read in the past about Japan and South Korea having massively fast broadband and the uptake of this faster broadband was reasonably small compared to the total internet population. The real disruptor here is wireless internet via 3G technology and 802.11 Wireless. Using the internet via your mobile phone sucked until recently and was also really expensive. Now even in Australia, you can get better than 256 KBits (ADSL1) from basically anywhere. In your car, on the train, with the right telco, in places which only to recently had access to either dialup (56 kbits) or satellite (latency is awful).
There is little wonder that the netbooks (smaller laptops) are mentioned. This combines several of the disruptors above. The smaller weight, and less performance means less weight and longer battery life, which dovetails perfectly with the notion of always online applications from anyway.
I have a Dell Mini 9 and it weighs less than some of the books I have. I actually replaced by older laptop for work with this. I can upgrade to 2 Gig and Windows XP works perfectly fine, and the SSD makes IO reads super fast.
The perfect computer device would be a device which has a screen the size of A4, that foldes to A5 (in half), has a touch screen, so you can type on the screen with a full size keyboard and has 3G internet in built, with a battery life around 3-4 hours.
Have Fun
Sunday, April 13, 2008
Flash Memory or Solid State Disks are disruptive
I asked a rhetorical question back in 2006 "Is Flash Memory disruptive"
The answer would have to be a resounding yes. Already the price per GB is rapidly falling.
People motivated by performance are getting very interested, if not already implementing systems which use Solid State Disks (SSD) as the part of or the primary storage for data.
Specially the solution is to get RAM IO read speeds for read intensive databases. Most databases are 90-98% read versus write. The technology of improving both the write speed and the life of the SSD is improving.
If your dig into computer architecture there is a hierarchy of IO performance. It start with CPU cache and ends with tape drives. See Jim Gray's CyberBricks presentation for more detail.
So Flash Memory as Solid State Disks are disruptive. They are disrupting hard drives and may well force them down the food chain in the IO pecking order. Making them closer to tape as a storage option.
This has not been helped by the inability of hard drives to get faster. Potentially hard drive manufacturers have seen that Solid State Disks are going to disrupt them, and so are no longer spending loads of money on Research and Development (R&D). If this was the case, it will truly be a change from past behaviour as the incumbent technology leaders tend to pursue the most demanding customers until they discover the market has moved on and those demanding customers switch and leave them behind.
It will be interesting to see how this pans out in the future.
Have Fun
Paul
The answer would have to be a resounding yes. Already the price per GB is rapidly falling.
People motivated by performance are getting very interested, if not already implementing systems which use Solid State Disks (SSD) as the part of or the primary storage for data.
Specially the solution is to get RAM IO read speeds for read intensive databases. Most databases are 90-98% read versus write. The technology of improving both the write speed and the life of the SSD is improving.
If your dig into computer architecture there is a hierarchy of IO performance. It start with CPU cache and ends with tape drives. See Jim Gray's CyberBricks presentation for more detail.
So Flash Memory as Solid State Disks are disruptive. They are disrupting hard drives and may well force them down the food chain in the IO pecking order. Making them closer to tape as a storage option.
This has not been helped by the inability of hard drives to get faster. Potentially hard drive manufacturers have seen that Solid State Disks are going to disrupt them, and so are no longer spending loads of money on Research and Development (R&D). If this was the case, it will truly be a change from past behaviour as the incumbent technology leaders tend to pursue the most demanding customers until they discover the market has moved on and those demanding customers switch and leave them behind.
It will be interesting to see how this pans out in the future.
Have Fun
Paul
Monday, July 23, 2007
Innovators Solution
I finally got to the Innovators Solution in the list of books I have to read.
The whole task/circumstance method of discovering what products may succeed if they fulfill a job to be done is an interesting insight. I have read similar in articles talking about programming, software and computers (and databases). Sometimes the disruption is in the ease of use and this disruption can occur almost non-consumers who find the existing products to complex.
Interestingly enough the book mentions Oracle attempt to use thin client computers to get non-consumers of computers using a computer. Computers are still seen as way too complex, especially when most people can't even set the time on their video or DVD player.
Haven't finished yet, however the work dovetails nicely with other stuff I have read over the last year or so.
I have been busy elsewhere blogging about databases and other stuff.
Have Fun
Paul
The whole task/circumstance method of discovering what products may succeed if they fulfill a job to be done is an interesting insight. I have read similar in articles talking about programming, software and computers (and databases). Sometimes the disruption is in the ease of use and this disruption can occur almost non-consumers who find the existing products to complex.
Interestingly enough the book mentions Oracle attempt to use thin client computers to get non-consumers of computers using a computer. Computers are still seen as way too complex, especially when most people can't even set the time on their video or DVD player.
Haven't finished yet, however the work dovetails nicely with other stuff I have read over the last year or so.
I have been busy elsewhere blogging about databases and other stuff.
Have Fun
Paul
Thursday, February 08, 2007
A couple of posts about Database disruption
Zack has some good articles about disruptive software and a specific article about databases.
Interesting reading the comments regarding CAD and Google amongst others. I just haven't made the time to post lately here.
As mentioned in Christensen's book, the money tends to move from integrators to component makers and then back. I get the impression that Oracle is going for the integrator role on the whole enterprise software stack.
Will this work?
How much do managers want to be locked it a single vendor?
Or it is actually to the point again where the whole enterprise software stack has fragmented (brought on in part by Open Source) and business is starting to get worried about reliability and compatibility again?
The whole SOA (Service Orientated Architecture) is trying the stave off this worry. I am inclined to think more businesses will buy into the sales pitch that once Oracle has all components they will be able to integrate them better.
I don't think MySQL should become another Oracle database. Their move to allow new storage engines is going to spawn a data storage revolution ... seriously. There are some very interesting storage models which are more optimized for the increasing needs of web backends and real-time data.
Jim Gray's site (he is missing after going to sea in a boat) is a brain stormer.
So MySQL may just end up moving into a new potentially profitable space which has fragmented niche players away from pure relational databases. So MySQL ends up the tool floating on a sea of new storage engines, like google on top of websites.
Last question: What happens if the database just disappears under the hood. You have a built for your application storage engine with the ease to export/import/perform relational queries.
There is no lost of data movement to and from your database, no closed data architecture, just a storage engine completely optimized for your application.
Have Fun
Paul
Interesting reading the comments regarding CAD and Google amongst others. I just haven't made the time to post lately here.
As mentioned in Christensen's book, the money tends to move from integrators to component makers and then back. I get the impression that Oracle is going for the integrator role on the whole enterprise software stack.
Will this work?
How much do managers want to be locked it a single vendor?
Or it is actually to the point again where the whole enterprise software stack has fragmented (brought on in part by Open Source) and business is starting to get worried about reliability and compatibility again?
The whole SOA (Service Orientated Architecture) is trying the stave off this worry. I am inclined to think more businesses will buy into the sales pitch that once Oracle has all components they will be able to integrate them better.
I don't think MySQL should become another Oracle database. Their move to allow new storage engines is going to spawn a data storage revolution ... seriously. There are some very interesting storage models which are more optimized for the increasing needs of web backends and real-time data.
Jim Gray's site (he is missing after going to sea in a boat) is a brain stormer.
So MySQL may just end up moving into a new potentially profitable space which has fragmented niche players away from pure relational databases. So MySQL ends up the tool floating on a sea of new storage engines, like google on top of websites.
Last question: What happens if the database just disappears under the hood. You have a built for your application storage engine with the ease to export/import/perform relational queries.
There is no lost of data movement to and from your database, no closed data architecture, just a storage engine completely optimized for your application.
Have Fun
Paul
Friday, August 18, 2006
Is Flash memory a disruptive technology?
The main case study in the Innovators Dilemma was the computer hard drive market and how each successive wave of smaller hard drives was disruptive and ultimately destructive for the incumbent manufacturers.
Is it going to happen again with compact flash memory?
Compact flash memory is used by digital cameras, mobile phones and in USB memory sticks (they look like a stick of chewing gum) where hard drives are either too large or in most cases too fragile to be used.
Over time the amount of storage available on this flash memory has increased to the point where you can get 2G USB memory sticks and memory cards. The lifetime of the flash memory is also improving all the time.
I believe there at least two main selling points for flash memory
In the future I believe that flash memory will replace and then surpass hard drive use in modern computers. The cost per megabyte/gigabyte is still high however this will come down over time (as it always has)
Have Fun
Is it going to happen again with compact flash memory?
Compact flash memory is used by digital cameras, mobile phones and in USB memory sticks (they look like a stick of chewing gum) where hard drives are either too large or in most cases too fragile to be used.
Over time the amount of storage available on this flash memory has increased to the point where you can get 2G USB memory sticks and memory cards. The lifetime of the flash memory is also improving all the time.
I believe there at least two main selling points for flash memory
- Robustness, compared to hard drives you can drop these cards and they will still work.
- Ease of replacement for backup or additional storage.
- It has captured a market separate to hard drives i.e. digital cameras, portable memory sticks and now newer mobile phones.
- People are starting to think about replacing laptop hard drives with flash memory. For two reasons: robustness and power usage.
- There is potential for much higher IO (Input/Output) bandwidth with flash.
In the future I believe that flash memory will replace and then surpass hard drive use in modern computers. The cost per megabyte/gigabyte is still high however this will come down over time (as it always has)
Have Fun
Subscribe to:
Posts (Atom)