In a recent telephone interview with Ray Ozzie, Microsoft’s Chief Software Architect, a number of these topics were touched upon within the framework of what Microsoft may or may not be doing. There was one interesting point made in the interview “the margins will still be higher to people who build solutions”. A commenter explains this point .

… there are basically 3 tiers of service in the cloud:

1. one is the underlying server farm infrastructure of hardware and bandwidth
2. two is a middleware cloud utility software platform
3. three are the applications that reside on the cloud platform

I think MS will get vertically integrated in all 3, as they usually do. They will have higher margins at levels 2 and especially level 3.

The more interesting point – at least to the end users – was made in the comments …

… as [we] move toward services infrastructure development, how long until we see a new utility computing cost model and standard on par with KwH (electricity) for consumption and usage? mark my words, software is headed toward the electricity industry utilities standards model …

I am not sure where the balance is between our current “flat rate” consumer models and our “usage based” consumer models. For the first 100+ years of communications, people were happy to pay for telecommunications usage. However, now more and more people opt for flat rate packages. At the other end of the spectrum, people consume electricity, water, and heat with a usage model. One model promotes “use it as much and as often as you can” while the other promotes conservation. Some places have even switched models to change user behavior – towns which charge a “fee per bag of trash” is an example.

Companies who charge a flat rate are betting that the average user will not exploit the service beyond a profitable usage rate and when all users are aggregated the result is more profit.

Flat rates work best with application services (described as tier 3 above). While usage rates are accepted for commodities where the profit margins are low and there is no tolerance for gambling on how users will use the service. The internet is struggling with this problem now. It is really a tier 1 service and it is struggling with a tier 3 pricing model. The fact that all high speed internet service providers (Cable, DSL, and cellular data) monitor usage and have legal safeguards to shut off heavy users points to the fact that the “flat rate” model is a risky one and is likely to change soon.

The challenge is to find a price-point. In most of the utility cases, there is no competition so it is not realistic to say “whatever price the market will bare”. This inevitably means regulatory boards, government involvement, and legislation. It is already starting with internet access as customers complain about service “degradations” and the debate over net neutrality. The first step will be for service providers to determine the cost of their infrastructure (something they have already done). Next is for consumers to determine their usage. Most large corporations already do this but small companies and consumers likely have no idea how much or how little they use their internet connection. (I fall into this camp). 

It would be a very interesting study to find out the profitability price point. For obvious reasons, no internet provider will publish their data but its possible to get a good idea if the data of large companies- who have no vested interest in selling their network bandwidth but need to track usage in order to keep their business operational.

Could this information be gathered publicly today ? How would we go about doing it ? Who would be willing to share their cost vs usage ? What “usage model” price point would be acceptable to consumers ? businesses ? government ?

So far this has all focused on the bottom and the top of the internet food chain. There is still the middle. In some cases, it is exposed directly as in cloud computing offerings by Amazon’s AWS. In other cases it is buried within the application offering as in Gmail where users get storage as part of mail (with some restrictions). But the lines are not really as clear as all of this. For example, Amazon charges bandwidth costs for storage. However, if your applications runs in their cloud and you use their storage, then bandwidth between the application and the storage is free.

Cloud computing has a much narrower target audience than either raw internet connectivity or applications. As such, it is likely to get compressed into one or the other layers – at least as far as billing and service models. All of this predicts the “data center” will become an operating expense  and will be squeezed as much as possible, just like  brick and mortar businesses and office space. the more you can make that “someone else’s cost” the easier your business. And a small number of highly efficient companies will continually optimize the offering to maintain a profit margin.

Speculation: All of this could be moot. Internet connectivity has very little value. The value is in the applications, information, and services. Imagine that the internet had a usage fee but browsing Amazon.com or eBay.com were free because those companies covered the cost of internet access as an incentive to use their service. Alternately, the internet connection might be free but you pay for your email service (either directly or because you accept advertisements). This latter model is being considered for cellphone data service. There are many different business models that *might* work. In the end, I predict internet access will go the way of electricity and water but there are other ways.

Let’s start with the big topic right up front – “Multi-Tenancy”. (I’ll call it MT from here on out just to save some typing.) MT means having more than one customer use the hosted solution. Obviously, MT is how you scale your customer base without scaling your cost.

There is a continuum for MT, each being more efficient than the previous.

• each customer has their own separate hardware, software, customizations, and data
• customers share hardware but have their own software, customizations, and data
• customers share hardware and software but have their own instance, customizations, and data
• customers share hardware, software, and instance but have their own, customizations, and data
• customers share hardware, software, instance, customizations, and data is isolated only by security

The progression goes from most costly to most efficient. What drives a specific solution along this line is a combination of the software used in the solution, customer customizations (unique requirements), and security requirements (user and data isolation).

For solutions targeted at individuals – accessing features, access to other users, and access to data – is on a per-user basis. An example of this is Gmail. However, there is another type of solution targeting individuals, which by featureset, does not separate users. An example of this is AOL Instant Messaging.

For solutions targeted at groups of users, then you need to consider how the directory or ‘address book’ allows users to perform actions with others of their group but not be aware of or have access to other groups or individuals in those groups. This is the key attribute of MT solutions.

The directory tends to drive the implementation of most hosted services. Some software can handle multiple directories or multiple groups within a directory; *but* a lot of software does not. In this latter case, you need to start shifting from the left on our continuum.

The ‘feature-set’ is another consideration. Some solutions provide each group may have a different set of features or a different look-n-feel. If the software does not support this, then again, you shift further from the left on our continuum.

Security is a big consideration for some potential customer groups. Some customers require secure data, data isolation, and data encryption ‘at rest’ (such as FIPS 140-2 . If the software does not support this, then again, you shift further from the left on our continuum.

So, “directory”, “customization”, and “security” are matched up the software to be used, resulting in a point on the continuum.

Next, I’ll look at and example using WebSphere Portal.

Stay tuned !

The challenge with looking at persistent chat from the “nose bleed” seats of the Astro Dome is that persistent chat can be implemented in so many different ways and used for so many different things.

• A disaster recovery team may want persistent chats to track issues asynchronously but in near real-time.
• A helpdesk may want a persistent chat to help customers and reduce the number of support personal.
• A Project team may want persistent chat for handle tasks and “to-do” assignments.

What starts to differentiate one implementation from another is (1) the feature set; and (2) the infrastructure or integration with existing systems.

I’ll use three examples -

1. the ubiquitous web forums
2. TeamSessions by Instant Technologies
3. Campfire by 37Signals

(caveat – I only picked these because I have either used them or they were near the top of a Google search). These are by no means exhaustive of the genre nor do they cover all use cases.

Web forums are everywhere. A nice feature about forums is they thread messages very well. It’s easy to have multiple topics covered simultaneously and overlapping. But, forums are really not “chats”. A typical user case would not assume two or more users would be active at the sametime, waiting for replies from one another. Forums are most often web based so they tend to support web metaphors well – text, images,links, etc. Most forums discourage uploads so they are not purposed for content sharing. Web forums have many implementations – some are open source, most will deploy onto a standard web server / application server infatuation. This means you can get hosted forums with most Internet hosting provider services as well as get the software and install it yourself – you decide if you want a hosted service or a private installation (say, on your business network and behind your firewalls).

TeamSessions is a web based solution that leverages a java applet to provide an “IM-like” user experience. It leverages Lotus Sametime for it’s real-time awareness and network infrastructure. (Users are not required to use Sametime as their text messaging solution but TeamSessions integrates well into existing Sametime environments.)

Adding rooms is quick and access to rooms can be restricted or left open to registered users. The rooms support text and images as well as storage for documents. TeamSessions is differentiated by some of its more business centric features such as security features, the ability to manage the growth or team rooms (both issues important to company IT departments), personal alerts (where a user defined monitors’ and gets alerted when user defined keywords are detected in the chat of a room), etc.

While TeamSessions can be hosted, most installations are within a company’s network. Again, it comes back to its integration with Sametime.

Campfire is a web based hosted service. It has a web-friendly user experience and makes it easy for users to have numerous “rooms”. Setting up rooms is quick and access to rooms can be controlled. The rooms support text and images as well as uploading of documents. One strong feature of Campfire is it’s integration across rooms. A user can “slice and dice” activity across all of the rooms they have joined. This makes it more likely that lots of rooms (with narrow topic / focus) can be managed by an end user. Without this capability, tracking and interactive with a large number of chatrooms would be difficult at best. As I mentioned, Campfire is a hosted service. Thus, if you have restrictions that prevent you from using an Internet service or you prefer to deploy a software solution on your own network, this is not an option.

One thing I did not see in the two “real” chatroom solutions above are some of the Web 2.0 features. I expect to see solutions move in this direction with support for RSS/ATOM feed readers at the very least but also the ability to extend the persistent chat solutions with “widgets” from other vendors / sources.

I have not personally used Campfire beyond a quick demo but the video tour really looks compelling. I have used TeamSessions and, in the business case I use it it, it is very effective. Probably the most compelling feature is that it integrates with the instant messaging solution used at work (Lotus Sametime). I had a chance to try out the Sametime 7.5 client with a plug-in that gave me direct access to my TeamSessions chat rooms. It made use the chatrooms as seamless as my regular instant messaging.

One topic I avoided for the time being is “best practices for on-line meetings”.I mention this because chatrooms are often used as virtual meeting rooms.There is a litany of discussion going on about how to best collaborate across virtual teams – especially will all of the telecommuting. I’ll try to tackle that topic some other time.

Disclaimer: I am not a representative of Instant-Technologies or 37Signals. I gain no compensation from them and do not necessarily represent their views. I have attempted to correctly depict the technology described above but do not guarantee accuracy for the statements made.

The service is hosted Sametime instant messaging, web conferencing, chat conferencing, persistent chat rooms, and document sharing all tied together with comprehensive directory services and a WebSphere Portal front end.

The new functions we are installing this evening include a series of Dashboards to provide concurrent usage, and various Sametime Community and Meeting room services. You may recall I talked about some multiple medium coding. This is the output from that work.

The development environment is Eclipse … well, more accurately it is WebSphere Portlet Factory with the Dashboards extensions. All of the performance graphs and tables are rendered with a single portlet.

The graph shows the user concurrency for web conferences and Instant Technologies TeamSessions chat room users.

IBM Lotus Quickr will provide ready-to-use team places where team members — inside or outside the firewall — can share information and collaborate on projects online.
Lotus Quickr comes with Quickr connectors that integrate into your familiar desktop applications so you can “work the way you want to work.”

What’s more, Quickr connectors work with current versions of these applications — avoiding costly desktop upgrades and freeing your valuable IT resources for more important work.

Lotus Quickr is designed to empower grassroots, bottoms-up adoption by letting users invite and encourage others to participate and to bring about change in the way the organization manages information and works together.

Shared content libraries
Quickly set up libraries to organize and share content. Users can access content in Lotus Quickr right from their familiar Microsoft Windows desktop. They simply drag documents from their C-drive folders to personal or shared Lotus Quickr places. Or, they can save documents to Lotus Quickr folders directly from the application they are using.

Source: IBM team collaboration software – Lotus Quickr

Now, I am a long long time user of Lotus software (can you say Notes 1.2 ?). However, I was never a big user of Quickplace. It was too slow – which may be as much about my server and my ability to tune software as anything else so I’m not casting blame.

For those reading this, you also know I am a new blogger. It took me all of about 5 minutes to get WordPress installed with my hosted service. It took me about 2 days worth of playing around to get a look-n-feel that I am happy with and I’ll admit I still have a few invisible issues to iron out. But the whole idea of blogging is becoming more clear to me.

Blogging is not just a place to share photos of the kids and tell about your recent vacation (and trust me, you don’t want that type of blog from me – BORRING). Blogging can be a great source of information. Some blogs are “hints and tips; other blogs are training information; my blog tends to be open ended issues I want to “get out there” for brain storming and reaction.

The software for this form of content sharing is broad and deep – content publishing, file servers, and document management all share aspects of what I am describing. When you build and end-to-end solution, you not only have the service end of things, but also the user experience. That UI serves as both the input and output of the solution.

What has made blogging so interesting is the easy of both in the input and output processes. At the lowest common denominator, all you need is a web browser at the user end of things. As for the service, it has become as simple as paying an Internet hosting service $8/month and you’ve got a full web site with blogging , on-line databases, storage, access control and more.

Now, IBM Lotus is introducing Quickr. I have no hands-on experience with it and my imagination is running wild with ideas of what I will find when I finally “open the box”. I am *so* looking forward to getting Quickr installed ! I want to see the connectors in action and see how I can integrate Quickr with established sources of content.

I’ll be interested in the “software as service” capabilities of Quickr. Blogs have value both as an open and public content expository (my word) as well as a corporate tool (staying well protected behind corporate firewalls. I anticipate lots of companies will explore the latter so I will take some time to look as the hosting service side of things.

… this could be the beginning of a beautiful friendship …

Moore’s Law assumes compute power will double every 18 months (at least as it relates to the processor chip).

Typical computer applications like word processing, graphics editors, presentation packages, etc. are gobbling up more and more computer resources including CPU and memory. It’s ridiculous to think that a 1.5GB memory machine with a 1.83Ghz Duo-Core chipset crawls with some of today’s applications.

I don’t expect Web 2.0 to deliver an equivalent to Photoshop CS. At the same time, I don’t do a lot of high end image editing. I need some simple web graphics and there are websites like Adam Kalsey’s Button Maker that make this possible.

Will we finally see useful Internet service based applications ? Will Moore’s Law shift importance to mega servers “out there” somewhere ?

Will we trust Internet applications ?

Gmail, Flickr, and others are banking on the answer being “yes”. I think these services are amazing but I *don’t* use them. I am having trust issues – will my personal privacy remain intact ? What will Google and Yahoo (the owner of Flickr) do with all of our data ? Will we end up with something like the extreme targeted advertising we saw in Minority Report ?