Comments on: What PaaS Should Learn from the August 2003 Blackouts http://www.saasblogs.com/2008/04/15/what-paas-should-learn-from-the-august-2003-blackouts/ Understanding the Software as a Service Revolution Mon, 08 Sep 2008 01:41:10 +0000 http://wordpress.org/?v=2.5 By: Andrew Biss http://www.saasblogs.com/2008/04/15/what-paas-should-learn-from-the-august-2003-blackouts/#comment-43666 Andrew Biss Wed, 30 Apr 2008 11:39:38 +0000 http://www.saasblogs.com/?p=204#comment-43666 Comparing PaaS with electricity generation is a good basis for talking about SaaS with business decision-makers. From the technical side we can see some of the challenges with a centralised approach. I think that business ISVs must move to PaaS to have any real chance of competing in the future. PaaS is a big help to ISVs, allowing them to stop developing low-level infrastructure. SaaS ISVs can concentrate on building unique features for their vertical niche. As you say, there is a risk when buying into a tool as well as the underlying platform. I think PaaS suppliers know this. They also have a strong vested interest to separate applications from compute and storage services. PaaS suppliers start by building their own infrastructure. However, this is a big continuing cost and will be difficult to manage with future growth. PaaS suppliers have a financial interest to buy-in utility compute and storage services. Keeping the lower levels of the stack separate they can profit from price and service differentials between utilities. By driving a hard bargain with the utilities they can lower their running costs and protect their margins. This will only work, of course, if the applications and the compute platform remain loosely coupled. I think that PaaS suppliers understand this truth. Moving workloads from one platform to another will (in time) become a standard feature. Comparing PaaS with electricity generation is a good basis for talking about SaaS with business decision-makers. From the technical side we can see some of the challenges with a centralised approach.

I think that business ISVs must move to PaaS to have any real chance of competing in the future. PaaS is a big help to ISVs, allowing them to stop developing low-level infrastructure. SaaS ISVs can concentrate on building unique features for their vertical niche.

As you say, there is a risk when buying into a tool as well as the underlying platform. I think PaaS suppliers know this. They also have a strong vested interest to separate applications from compute and storage services.

PaaS suppliers start by building their own infrastructure. However, this is a big continuing cost and will be difficult to manage with future growth. PaaS suppliers have a financial interest to buy-in utility compute and storage services. Keeping the lower levels of the stack separate they can profit from price and service differentials between utilities.

By driving a hard bargain with the utilities they can lower their running costs and protect their margins. This will only work, of course, if the applications and the compute platform remain loosely coupled.

I think that PaaS suppliers understand this truth. Moving workloads from one platform to another will (in time) become a standard feature.

]]> By: Les http://www.saasblogs.com/2008/04/15/what-paas-should-learn-from-the-august-2003-blackouts/#comment-43029 Les Sat, 26 Apr 2008 11:56:39 +0000 http://www.saasblogs.com/?p=204#comment-43029 Good post and great analogy. We've been in this same battle of "best of breed" vs. tightly integrated and controlled monoliths in the software industry for years. Usually in the very earliest stages of a developing market, the monoliths control. Since no real standards exist, the early survivors build their own infrastructures and integrations themselves. But, like the dinosaurs who evolved out of existence, the market ends up dictating the ultimate solution. With open interfaces to start, this process can be accelerated. The question in PaaS is not IF, but only how quickly the market wises up to inevitable decentralization. The democracy of the internet will likely drive this process faster than in the past. Good post and great analogy. We’ve been in this same battle of “best of breed” vs. tightly integrated and controlled monoliths in the software industry for years. Usually in the very earliest stages of a developing market, the monoliths control. Since no real standards exist, the early survivors build their own infrastructures and integrations themselves. But, like the dinosaurs who evolved out of existence, the market ends up dictating the ultimate solution. With open interfaces to start, this process can be accelerated. The question in PaaS is not IF, but only how quickly the market wises up to inevitable decentralization. The democracy of the internet will likely drive this process faster than in the past.

]]> By: Robert Hagens http://www.saasblogs.com/2008/04/15/what-paas-should-learn-from-the-august-2003-blackouts/#comment-42436 Robert Hagens Tue, 22 Apr 2008 18:00:18 +0000 http://www.saasblogs.com/?p=204#comment-42436 I too would be depressed with the potential outcome you describe. However, I think that cost of computing and the open source movement will contribute to our salvation. The one issue I would take with the electric analogy (which is a good one) is that electric power was generally tricky to work with, expensive to install, and so lent itself to centralization. (It was generally un-economic to wire communities with multiple systems). On the other hand, the plethora of open source software and good quality Internet colocation means that it will continue to be easy to host your own services without a tremendous investment. That will be, I hope, a counter balance to the monolithic forces you describe. Never underestimate the market power of open systems! I too would be depressed with the potential outcome you describe. However, I think that cost of computing and the open source movement will contribute to our salvation. The one issue I would take with the electric analogy (which is a good one) is that electric power was generally tricky to work with, expensive to install, and so lent itself to centralization. (It was generally un-economic to wire communities with multiple systems).

On the other hand, the plethora of open source software and good quality Internet colocation means that it will continue to be easy to host your own services without a tremendous investment. That will be, I hope, a counter balance to the monolithic forces you describe. Never underestimate the market power of open systems!

]]> By: PaaS - monolitich or segmented stacks… at diversity.net.nz http://www.saasblogs.com/2008/04/15/what-paas-should-learn-from-the-august-2003-blackouts/#comment-41495 PaaS - monolitich or segmented stacks… at diversity.net.nz Wed, 16 Apr 2008 01:14:36 +0000 http://www.saasblogs.com/?p=204#comment-41495 [...] excellent post over on SaaS blogs locks at two possibilities for PaaS. One is very vertical in nature (AppEngine, [...] […] excellent post over on SaaS blogs locks at two possibilities for PaaS. One is very vertical in nature (AppEngine, […]

]]> By: Michael Nygard http://www.saasblogs.com/2008/04/15/what-paas-should-learn-from-the-august-2003-blackouts/#comment-41426 Michael Nygard Tue, 15 Apr 2008 15:32:59 +0000 http://www.saasblogs.com/?p=204#comment-41426 This is another reason why I believe that many companies will never allow that much of their destiny outside of their control. I suspect that massive centralization isn't the future of cloud computing. Rather, I think we will see more and more clouds. http://www.michaelnygard.com/blog/2008/02/a_cloud_for_everyone_1.html This is another reason why I believe that many companies will never allow that much of their destiny outside of their control.

I suspect that massive centralization isn’t the future of cloud computing. Rather, I think we will see more and more clouds.

http://www.michaelnygard.com/blog/2008/02/a_cloud_for_everyone_1.html

]]>