Wednesday, November 28, 2007

Open Source vs SaaS?

I am amazed to discover lots of discussion about a so called "Open Source vs SaaS" debate.

Surely it is obvious that the barriers to adoption of SaaS and the fears about the uncertainty of provision of service will be reduced through greater patration(as in portability of service). The fastest way of achieving this is through providing an operational means of implementing an open standard - which means Open Source.

It is not a this vs that debate, as Open Source will accelerate the adoption of the XaaS stack and it is complimentary to it.

Next I'll be hearing about the great "Salt vs Vinegar" or "Fish vs Chips" debate ....

I'll have mine "Fish AND Chips" please.


Aurélien Pelletier said...

In a SaaS world you can't scratch your itches like in an open source world. You only have access to a service, not to the source code. The debate is going to move from open source to open data or open API... Services are a bit like DRM they limit what you can do with the data. An open API, or a restfull API should give you direct and full access to the data, they are the MP3 of the data.

Peter J Cooper said...

There is no such word as patration. At least, wikipedia and the Australian Oxford think so.

However, Patras is an interesting Greek port where there was a fight for independence.

swardley said...

Hi Peter,

Patration is a "lost word" dating from around 1650 and has the meaning "perfection or completion of something".

Now I used Patration to describe "the freedom and portability to move from one service provider to another without hinderance or boundaries" at my talks at FOWA / Web 2.0

The reason why I used a "lost" word comes from Robert Lefkowitz, who said when it comes to new computing terms "instead of using words which already have common uses, we should dig up archaic words which might or might not have related meaning and just use those, since no one knows what they mean anyway and it makes you sound erudite to use them."

Hope that helps. I know it's a bit naughty of me - as per my post.

swardley said...

Hi Aurelien,

I agree that with many services in the SaaS world you cannot scratch your itches.

However the purpose of open sourcing a SaaS service is twofold. First to attempt to provide a open standard and secondly to provide an operational means of implementing such an open standard, the latter also helping to promote the former.

Open data is not enough, as there is little point to open data if there is no equivalent service to move it to.

"the freedom and portability to move from one service provider to another without hinderance or boundaries" therefore requires two things - the freedom and portability to move (e.g. open data) and the existence of other service providers.

Achieving this portability, freedom, multiple providers and the generation of an open standard is most likely to be achieved through the use of open source.

Open APIs are a step in the right direction but in the long run I suspect the disruption in this market will come from open source providers.

The thing which creates portability on EC2 is RedHat. You just need multiple different providers of the same RedHat installation. In this case RedHat provides the standard to enable portability for your application and your data.

The value in Facebook is the network - the smart move would not only be to provide open APIs but to open source the entire platform and create an ecosystem of competitors. Do that and open social is history.

It's not about scratching an itch, it's about how to become the open standard.

Anonymous said...

It is more that GPL does not spell out that distributing an open source solution over the Internet is "actually" considered distribution. There are companies like MySQL and SugarCRM who have been completely screwed because they choose the GPL license. I could host SugarCRM for free and make many upgrades while distributing as a SaaS solution and I don't need to feed those updates back to SugarCRM. Google has done this to MySQL for years. The SugarCRM folks have their heads screwed up and it will come back to kill their business model in the future. TigerCRM already offers the exact code for half the price. There are new licenses that cover the basic premise that distributing software over the Internet is actually distributing software. That is basic but OSI did not seem able to get off their thrones to address this problem. SocialText and Projity have adopted these new licenses. We use their Wiki and Project Management solutions. The whole problem with SaaS/Open Source is self inflicted by the OSI

swardley said...

Actually, I have to say the GPL v3 license is near perfect for the SaaS market. Let me explain why.

Let's take a hypothetical open source CRM system with multiple companies providing the system as SaaS.

Now as a consumer of the service you have choice between providers as well as installing your own version. This is as long as none of the providers start to change the primitives, hence there is a market for 3rd party assurance services. Overall as a consumer you have more choice which should help overcome the fears of adoption. Furthermore with enough providers / consumers such a system is more likely to become the defacto standard especially if the area we are dealing with is CODB (cost of doing business).

As the solution is open sourced, each of the providers retains strategic control of their business and are not dependent upon the whims of another company. As they are providing it as SaaS under GPLv3 they can modify the system and improve its implementation without returning such changes - this creates competition between the providers on the implementation details whilst still retaining pressure to commit their changes to source due to the constant cost of upgrade. Note if you change the primitives then consumers would be warned by any 3rd party assurance services over compatibility issues with your service.

In practice you should be able to achieve competition with portability through open source and it is also the most disruptive way that any news starter can attack this market.

So TigerCRM vs SugarCRM - excellent that's competition for you, whether it is price or service. If those two fight it out (or more get involved) and you have portability between them - then it's salesforce that will suffer in the long run.

A good article in this area is


Simon W

swardley said...

As for SugarCRM having their heads screwed on right.

Well it seems that they have decided to go for a small piece of a big pie rather than a big piece of a small one.

With enough competitors in an ecosystem providing sugarCRM as SaaS - they will own the industry.

Very smart move.