As the hosting business is a low margin business support cost is a very critical area that they want to have as low as possible. In many low-end offerings you actually don’t have any support other than calling high priced hotlines. If you do get support with your offering the hoster still wants the amount of calls be a low as possible. But what happens if a hoster  chooses to aggregate applications or modules of applications for their customers? Who is going to take the support cost and who actually will do the support ? There’s no doubt that this will increase price and complexity of servicing the hosters customers well.

If you go to www.microsoft.com/saas you will end up with the definition: “Software deployed as a Service and accessed using Internet technologies”. To me this is very broad probably too broad. Unfortunately I don’t have anything clearer and more precise than this.

From the talks with the parties involved and from the experience that I have with hosters around Europe I want to add some bullet points that appeared to me as very important and critical for SaaS in the scenario where Hoster and ISV become partners.

Many hosters make their money servicing consumers in a low margin business. Many of them have a business brand like Cohaesio vs. Surftown or Schlund&Partner vs. 1&1.

I’m not sure how realistic a scenario would be where a hoster would give up customer relations by only selling delivery platform services. In fact 1&1 has grown from a direct marketing company to the biggest hosting company in the world. The point I’m making here is that maybe this is a different type of service provider which might not be the classic hoster. I wonder what kind of partner is such a SaaS provider?

In a hosted environment an application will not have the full freedom of choosing it’s runtime behavior and the way it is being provisioned. There are many policies defined in a hosted environment that need to be enforced from the service provider. Therefore I see the merge of the two manifests as a design-time step. It can be used to actually generate

·         a set of abstract provisioning tasks

·         the provisioning script(s)

·         the provisioning workflow

This output needs to be inspected and tested in the hosted environment and will then be approved and enabled.
Parallel to generating the provisioning tasks there are a more steps which need to be accomplished:

·         create a product in the product catalog

·         associate pricing and billing information

·         associate service plans

·         define monitoring and tie it together with service plans and billing

Michel Baladi posted a lot of information on the outcome of our efforts regarding the SaaS Delivery platform in his blog. You can find the information at blogs.msdn.com/baladi

The two most significant innovations which came out of the Architecture sessions are those two manifests. The application manifest is something built by the software vendor to actually define the needs of an application in a hosted environment. It contains information for Provisioning as well as information on how to monitor that application and what metering information is being used.
The platform manifest is a definition of hosting resources such as web servers, database servers etc. and their attributes like load-balanced, clustered etc.

These two plus a set of parameters can be merged together to generate a set of installation tasks.

We discussed the shared database approach or multi-tenant data layer and came to the conclusion that running backups is not so much of a problem but restoring data on a tenant by tenant basis would be a complex pattern that is not supported currently, neither by software vendors nor by the Windows-or .NET platform.  I recognize that there are scenarios where sharing the same database can lead into scalability advantages. This would be the case where database size is so small that the overhead of running the system vs. the actual tenant data limits scalability. This is a scenario that is very likely to exist in consumer oriented hosting solutions. Nevertheless as there is no solution to the backup and restore scenario and other parts of the platform seem also not ready -  I would not recommend this scenario to a Hoster or to an ISV.

I'm actually using them myself and find them one of the most useful resources that I'm aware of. Check them out at http://www.microsoft.com/uk/msdn/events/nuggets.aspx.

this week I had the great pleasure to speak about Ajax and WPF/E in Dubai in front of students from various universities around UAE. Honestly I was very nervous presenting in a country with a culture that I did not know. Actually it turned out that the students were less shy than myself and I had great conversations with many of them as well as I had chats with some of threir professors. Yesterday I had the chance to speak to prefessional developers which was another great experience. As mentioned in my presenation (attached) here are some addintional useful links to technology I was referring.

For the attendees: Ajax.net can be found on ajax.asp.net You can find the Ajax Control Toolkit on www.codeplex.com/AjaxControlToolkit.If was also refering to the blog of Nikhil Kothari at http://www.nikhilk.net/ and WPF/E can be found at http://msdn2.microsoft.com/en-us/asp.net/bb187358.aspx 

Please watch the information and announcements being published during MIX07 - I think there will be exciting news.

Today I started my first trip to Estonia where I'll be speaking to students about Web2.0 and related SaaS engagements I'm currently doing. Additionally I'll give an overview on ASP.net AJAX We will also visit a hoster this week.

Presentations from this conference are available here.

With a first ADS (Architectural Design Session) at the Microsoft Innovation Center in Denmark we stepped into a very interesting space. In October we spent three days together with a hoster and a software vendor discussing opportunities and technologies arround Software as a Service.
I think we need to figure out what are the interesting categories arround that theme. For me there are three:

1. Business impact
2. Architecture of a SaaS solution
3. Service Delivery platform

There are lot's of interesting discussions going on in regard to SaaS and its business impact and how it will change the IT world. I hope that my observation adds to this discussion event though it is an assumption and is not verified with hard figures. Here it is:

Hosters and ISV (especially Line Of Business ISVs) potentially approach SaaS from different perspectives.

The classic hoster:

1.       seems to be keen on moving up the stack to become an aggregator and gain more quality type of service offerings through SaaS.

The LOB ISV might have the following goals:

1.       Renting services from hosters (e.g. Operations but also things like Provisioning, Federation etc.)  to actually outsource the application but keep the business relationship with the customer

2.       Sell through the hoster their solution and create a new sales channel and the hoster has the customer/business relationship

IMHO the majority of classic hosters do not refer to themselves as service companies. So in many cases I would suspect that ISV option 1 does not fit to the business model of a hoster. This is rather a services play for service companies. 

I put together this slide to visualize this a little bit. The two curves are guesses – assuming that the economy of scale is higher in Small to medium sized services whereas the closer you get to LOB the scale goes down and therefore the margin decreases as well.