Private Clouds? Hosted vs Private, what comes next?

So, I saw the article about Cisco entering the server, or private cloud market.  The big enterprise vendors Microsoft with Azure, Sun with Q-layer, have started to want to get in on the cloud hosting companies and several announcements recently have been interesting.  However, in my opinion the market leader is still Amazon Web Services, EC2, S3, etc.  I recently moved my Demisauce server to EC2 and really enjoyed setting it up.

As more and more capability moves to hosted services, open-source gains with linux etc but what else?   Xen was gaining especially with Amazon EC2, however Ubuntu seems to be moving away from xen towards KVM, as does Redhat  (Wow, kvm just needs to get faster though, I run it on my quad core AMD machine at home and it is much, much slower than my VMWare fusion images on my mac).  Redhat is building everything it needs for a CloudOS, with virtual machine management (Ovirt), hypervisor (KVM), security & directory (FreeIPA), and an API for virtual machine configuration (libvirt).

Once again, I can’t help but feel that the value proposition of Open-source is being missed by some of the big-co’s.  Open-source is so valuable because capability and inventions by one can be utilized by other’s.  This really isn’t as possible (ok, you can license technology from each, but that doesn’t happen much) among the proprietary vendors.  So, in the end run I think that the open source stack is going to exceed what the vendor’s provide.

Comment System for Potlatch blog on Demisauce.

I just posted an article about the new Comment system for the blog i am working on.  Full article here…

Here is a small picture, the post has full set of pictures.

Introducing Potlatch Blog: A google app engine personal blog

I was looking for a project to try out on Google App Engine, and also wanted to have something to showcase the comment system in Demisauce that i was working on. So, here is another Personal Blog Engine. You can get the code at my Github repository.

I dual posted this on my new Aaron’s Potlatch Blog on Google App Engine, but you will need to comment here because i don’t have the comments set up there yet.

Yahoo Y!OS strategy signals change in Software Development

I attended Web2.0 Expo last week in San Francisoco, and the thing that really struck me was the Yahoo Y!OS strategy. A bit of a blend of the Opensocial strategy combined with Google App engine, combined with widget development in general. I walked away convinced more than ever that we are seeing a startling shift in software development. Four or five years ago, we talked about developing on top of other “Platforms” and meant JBoss, BEA app server, MS .NET, Sharepoint, Plumtree server, Oracle Portal etc. Now, when we talk about developing on platforms we talk about Facebook platform, OpenSocial, Google App Engine, and Y!OS.

Changes:

• Creation of Community: Previously to get your software/application out you either created the community, or sold it as a “Plug-in”. Now you can leverage other’s communities.
• Scalability: Google App Engine, and now Y!OS help you scale up and out.
• Pages vs features: No longer are you creating a full application, but maybe a portion of a page, a feature instead of an app.

Articles about Y!OS

http://blog.programmableweb.com/2008/04/25/the-yahoo-operating-system/

http://www.readwriteweb.com/archives/yahoo_apps_platform.php

http://www.news.com/8301-13953_3-9928557-80.html

I really think this challenges software developers to have a much broader understanding of the ecosystem out there, the platform upon which to build which includes Y!OS type strategies.  Starting to see the value you add to your customer, and realize what you need to write vs what you can leverage.  Examples of this are  comment systems such as Disqus, as well as feedback systems such as Get Satisfaction.  Now, these applications seem first to be targeting the public consumer space, but I guess will migrate to be able to be more private.

Previously software developers needed to keep on top of the newest code level components such as ORM tools, caching tools, unit testing tools, etc.  Now however, they need to add to the tools the services and platforms such as Google App Engine, Widget services, Y!OS and servers such as the Disqus and Get Satisfaction.

Decision making in Companies vs Web2.0 communities

User powered communities, and user generated content (Web2.0) have started to replace tasks that were provided by employees of companies previously. Most of these initial tasks are simple decisions. Other online communities, Open Source software and Wikipedia we see all decisions, and processes being open, so that the entire operation of a large organization is without a traditional management structure or decision making process. The cost structures of these organizations are much less than the traditional company, so how far can companies go to reflecting these new virtual organizations?

The first wave of these employee tasks being replaced have been fairly simple in nature, and very close to the consumer. Digg.com community, or RSS Readers selecting articles replaces editors, vetting, content checker’s. Youtube members contribute short films, acting, directing, producing the content themselves, replacing or supplementing tv/movies etc. Users providing reviews and ratings on amazon.com replace employees at Best-Buy in sales staff, and research staff, and in purchasing staff deciding what products to stock.

A commonality of the employee functions being replaced: Simple decisions (Consumer “likes” or “dislikes”) Ones where employees were previously trying to act as a proxy for consumers. Examples:
- Best Buy sales staff: providing advice to consumers about the “best products”
- Best Buy Purchasing manager: Deciding out of many products, which to stock, amazon.com doesn’t need to decide, they can stock them all, allow consumers to choose
- newspaper editors: select most interesting articles
- newspaper fact checker’s: Check content, Digg.com readers can mark content as potentially inaccurate

However, more recently a couple of consumer input services have started which are more strategic and interesting: DellIdeaStorm and MyStarbucksIdea: Replace executive decision making on “what to focus on”, which products and projects to work on. These are more complex decisions, has cost trade-off inputs, more strategic decisions.

The “upper and middle management” of a company is there to allocate out scare resources to projects, focus on executive decided “goals” and “strategies”. To shuttle information up to executives, and shuttle down decisions. If consumers can start to dictate what to focus on, how much more of traditional middle management of corporate structures is going to go towards more web 2.0 user contributed input? With more information systems meaning we don’t need information shuttlers, with direct online information, voting/decision making online will we need as much middle management in the future?

What to focus on (Goals and strategies), how to allocate resources to these goals, feedback processes seeing if work meets users needs, can these go away? An open source community or Wikipedia has no/less formal decision making processes, middle management, yet is able to have a common direction. The engineers, tester’s, content editors adding value to an open source project are all “Value added”, that is every single person on the project is involved with directly adding value to the user. Compare this with a company with layers, and layers of middle management which does not directly add value to the product but instead is considered necessary to deliver the companies products/services.

So, how far will companies move towards new decision making processes? Web2.0 style decision making for more strategic processes? And, what tools will this reflect?

Book Review: The Big Switch

I have finished reading “The Big Switch:  Rewiring the World, from Edison to Google“, and although I have been an advocate of cloud computing (S3, EC2) before, this even more strongly reinforced my beliefs about the switch to utility computing.    It talks about the original usage of electrical systems in the world, replacing the water-wheels at manufacturing facilities, but where the company still runs their own power generation.   I particularly found amusing a quote about how the manufacturing companies when faced with electricity that could be bought from a utility, brought up the fears of “what if it goes down”.   This is the exact same fear I hear people in enterprise hosting bring up about Amazon S3, or Google Apps.   The cost and complexity of hosting today in an Enterprise IT environment, with the overwhelming complexity of compliance, disaster recovery planning, data backup, change management, release management, funding prioritization creates a stunningly difficult solution to hosting software.    With the complete ease of signing up and using hosted applications such as Google Apps, it makes me feel sure that yes, the movement will move towards utility computing.  

Demisauce Updates – Help Overlay

I just posted some changes to Demisauce for the help overlay javascript widget.  It is a popup (draggable) built on jquery that would be embedded in your application, but gets its content from the Demisauce server.   So, content for a web application can  be updated through a web based administration and updates are available within your application.

I will try to get some full size screenshots up soon on Demisauce.org web site.

Software as a service, for browsers or servers or both?

Recently I have been working on Demisauce, an open source Software as a Service project to provide services to other applications. Some colleagues of mine were trying to solve a problem at work that involved integration of voting and commenting capability into an existing software application (.NET) using a service approach.  Meanwhile, i was working on a tool to allow a “help overlay” to overlay parts of a web page with a dialog type box of content managed by Demisauce but shown within the browser of a page served by my application.

While talking about where Demisauce is headed, the need to incorporate “services” that are both server to server communication, as well as browser to server became apparent.

Diagram showing services that are both browser to server, and server to server.

With the browser more and more being the controller of MVC architecture, it is aggregating content (mashup) into  a single view and application.  However, that content comes from multiple servers.   However, again, the need to have services such as commenting applications, comments, help, etc that are usable within your web application as a “private” application (internal corporate software applications) instead of the web2.0 consumer/public model prevalent in mashups today.

The services we have targeted initially for this are the help application, to be able to pop open an overlay/dialog box of “help” contextual to your page or application, but have the content for that managed on Demisauce.  To allow for feedback on pages, and the management of that feedback (has anyone processed it), for generic content, email templates, etc.

Email, the root of all evil?

I just read an interesting blog about email and getting rid of it if it is controlling your work life, as well as the article they refer to http://www.theage.com.au/news/essential-gadgets/when-the-inbox-is-on-the-outer/2007/09/19/1190486197735.html is also interesting.  It definitely made me envious, to have an organization that forward thinking in how to communicate.  We push a lot of things out on wiki now, but much of the organization still uses email.   In addition to the focus, I think wiki’s or other online collaboration help with other aspects of communication:

•  Wiki’s are “Permanent”, so people think about what they write more before sending it.
• Wiki’s are “Open” so people don’t write things they probably shouldn’t have to begin with.
• Wiki’s promote people to stick with the facts, and stay away from lengthy political discussion.

However, neither Wiki’s or email address some other bastions of corporate waste, Meetings!

• Decision making:  Andrew McAfee has an interesting article that recognizes decision making as separate from information itself.
• Brainstorming:  Meetings can be (although most aren’t) very effective at brainstorming.

What makes a good goal?

At work, I and my team have been assigned to work on a new project, however, nobody knows what this project is supposed to accomplish. Everyone is talking about providing the “best widget/service in the industry”. Resources, and millions of dollars have been allocated, but nobody can describe why, or what the desired end goal is. Just to have a shinier fancier widget or service?In my experience, this is the wrong way to communicate, collaborate, and motivate. Companies should agree upon goals that are stated in terms of customer experience of what customers really want, and should be described in desired end result that is free of HOW to accomplish this. Couple of examples:

1. GM vs Toyota. GM (or other American auto companies) had been merrily dominating the American auto industry, but Toyota with an obsessive focus on Customer value was addressing quality issues. I can just picture internal goals/metrics at GM at the time about producing “x units” or “y cars/day”. While Toyota focused on producing happy customers. Because GM focused on an internal measurable (x units, or y cars) as opposed to a Customer metric (were they happy?), Toyota eventually won out. That focus on internal deliver-ables which are ASSUMED to be equivalent to what the customer wants are almost never right. All employees must have an understanding of what the goal is, what the customer value, see the toyota usa web site about “Customer First“.

2. Ipod vs “MP3 players”: IPod was not the first MP3 player on the market, but it did come to dominate. Other MP3 players may have had more “features”, but Apple realized that by focusing on what customers really wanted (“Good music listening experience”), they had to expand their definition beyond the “MP3 player’ and encompass some additional portions of the music experience (coolness, and the transfer, management of songs).

In both of the examples, Toyota, and GM made what I call a Proxy Association Error. That is, the Music player was a “proxy” for the music experience, and the automobile was a proxy for the transportation, people didn’t want car’s or players, they wanted music experience and transportation. Once a better proxy was found, the customers abandoned the old way for a better one. Toyota, and Apples competitors didn’t see this, they thought the “Car” and “MP3 player” were the actual thing that people wanted.

So, here are my paradigm for how to distribute work (not tasks) to a company:

1. Get the entire employee population involved in seeking to understand what customer value is (real customer value, not proxy error, the “music experience” not the “mp3 player”).

2. The goal of any work should be stated in how it should help the customer. The Goals should be distributed should be stated in terms that are free of how, and leave the wisdom of the crowds and changing winds of the marketplace open to allow a different solution that still makes the customers happy.

Examples:

“Identify the top causes of dis-satisfaction by customers, and put in place plans to mitigate them” not “Implement improvement of the invoice service”. The first leaves the ability for the employees of the company to contribute their own ideas, the 2nd doesn’t.

I am working on a new Strategic Goal planning tool which helps users create goals, collaborate on them with others, and track accomplishments, with visibility up and down the organization. It is intended to bring the best aspects of the wisdom of the crowds to helping guide a company as exist within the Open Source community.