Tuesday, December 28, 2010

SOA will die!

I think Services Oriented Architecture (SOA) will die with the whole notion of Cloud computing. If you were to follow to the history of Google APIs (Application Programming Interface), you will notice that Google had quite a few XML based Simple Object Access Protocol (SOAP) web services in the 2003 - 2004 time frame. If you were to surf the Google Code base, you will notice that there aren't alot of web services enabled APIs. What does this mean?

I think with phenomenon like Cloud computing, there will be a decrease in system to system level integration. How many of us can say that Enterprise Java Beans (EJBs) will make a come back? Maybe one or two handfuls around. The reality is that IT is evolving so fast that complexity of the accessible API is more simplified. Who would ever have thought that JavaScript is the favored scripting language to access APIs? It does make sense though. JavaScript is here to stay as long as the web browser (workstation or mobile). In this ever changing evolving IT world, services will be designed and developed using JavaScript. I am waiting for a JavaScript specifically for system integration. I like the JavaScript libraries that were built for Facebook and Google. Goodbye SOA and welcome to JSOA aka JavaScript Oriented Architecture in the cloud. I believe cloud will also cause the fall of the relational database management system (RDBMS) and rise to the Reduced HashMap. I will expound on this evolution in the data tier in the near future.

The image is taken from David Chou's blog entry, "Cloud Computing and the Microsoft Platform" at http://ht.ly/3ACCE

Sunday, December 26, 2010

Twitter Lists

For the last two days, I have been organizing Tweets I follow into Twitter Lists. I follow US Government Tweets here to understand what do my fellow IT Geek/bureaucrats tweet about. It's been an interesting exercise. Most of it is plain noise which doesn't really make sense in the grand scheme of things. Twitter can be used as an enabling tool for agency branding strategies. Here are some of the tweets I saw:
Anyway as you can see, Twitter is giving me the power to group the tweets I follow into "lists". This is a great idea. Rather than Tweeter creating the lists for me, it lets me create the lists and lets me publish my lists to the Twitter world.

Why is this so great? Twitter is allowing its users to create user centric communities of interest. This is truly social. Usually organizations, standards' bodies or architectures define the communities of interest which to some degree fail because it doesn't appeal to everyone. I have seen this problem in the "search" world. Search engine technologies like Autonomy let organizations or data architects design the taxonomy for a specific topic and then the search engine moves the documents to the appropriate taxonomy node during indexing. If the taxonomy is not done right then bunch of documents don't get put in the taxonomy and they are grouped in the "other" node. The reason why folksonomies work better than taxonomies is because of the "social" aspect. Folks who know the content are organizing it and not some information architect who is following an archaic process. BTW why do relational data architects call themselves data architects. Most of them don't have a clue on how to build an ontology.

Anyway I see alot of potential with the social function. If anyone has an opinion regarding this matter then please chime in. You can also follow me on Twitter at http://www.twitter.com/eknock

Friday, December 24, 2010

Make "TRON" interoperable

Yesterday I went to the movies with my son Seth. We had a nice Mediterranean dinner. After which, we got Edy's Mint Dibs ice cream and medium Barq's root beer and then we settled in our chairs and watched the Disney movie Tron:Legacy. It brought memories of me when I was 13 year old boy watching the original Tron. I thought it was the coolest movie every. I now have to say this movie is actually not that bad. My son liked the movie too.

Anyway the point of this blog entry is not to expound on the entertainment value of the movie but rather philosophical flaws of the movie. The human characters interact with one system and its programs. The "Tron" world is a universe of one system. This is so 1990s. Is this movie promoting stove-piped systems? I don't think so but it does show you the inherit flaws of a stove-piped systems. Here are the flaws:
  • One world concept - A stove-piped system is a world on its own. The language, the business logic, and hardware and software are designed for that system.
  • Limited scalability - Any optimization and scaling is done within the constraints of the system. My world is the center of the universe. Oh wait, my world (aka my stove-piped system) is the universe.
  • Failure to Adapt - In the movie, there is no sun. Hence we can infer that the "Tron" system doesn't have a program which acts like a sun. Now lets say there is a "Sun" program introduced into the "Tron" system. Will other programs recognize there is infact a "Sun" in the "Tron" system? Probably not because that logic has not be programmed in.
  • Reusable components - Imagine if the "Tron" system had reusable programs which can be used in several other systems. If so how can these programs be ported into other systems. The program "Tron" is imported in the "Tron" system however this program had re-programmed to interact in the "Tron" system.
Anyway if there is a Disney producer or a script writer reading this entry then please consider the following ideas for the next Tron movie:
  • Introduce programs called Interfaces which allow the Tron world to interface with other worlds like the "blue world" (IBM) and the "red world" (Oracle).
  • A cyber war across the network (kinda like an alien war)
  • Virus software in the Tron world
  • Tron on Facebook or MySpace - just kidding
Anyway the movie was great but the movie also shows the flaws in having "stove piped" systems.
Here is the trailer

Wednesday, December 22, 2010

Google TV

I am glad that I trusted my gut instincts and got rid of Dish Network and wanted to watch all of my television via the Internet. I did it because:
  • I didn't want to let my kids watch tv all of the time
  • it saved me alot of money (around $120/month)
  • I wanted to watch TV shows when I wanted to watch them
  • I didn't want to revolve my schedule around a specific time of the week
  • I wanted to pay for content that I was more comfortable. I would have paid more money to watch all NBA games and not the Catholic Channel.
It has worked out. I enjoy ESPN3, Sunday Night Football on NBCSports.com, Netflix, Hulu, YouTube and others. I now have a wider selection on Comcast and I am thinking of getting NEO Cricket channel.

Television content is now geared towards my lifestyle and I am not altering my lifestyle to watch TV. Google TV, Comcast and others are enabling others to do the same.

Tuesday, December 21, 2010

New Design Wordpressism

Lately I have been working with web designers and media experts at the FAA to do the following:
  • Provide architectural guidance for a proposed Web Content Management System (WCMS) which will be deployed on a government cloud.
  • Developing a Communications Architecture Framework
  • Collaboratively developing a FAA social media strategy with Ms. Karen Snyder
After doing web design for FDA/CVM from 2000 to 2004, I have realized that web design has morphed from mechanical javascript, splicing gifs and jpegs into rectangular shapes and controlling design flow using tables into CSS, AJAX, CMS. Speaking about CMS, Wordpress themed sites are spreading like wild fire. It seems to be the rage. An open source technology has strongly influenced the web designer's behavior. In fact the Wordpress layout seems to be de-facto layout for all enterprise, mom and pop web sites. The other thing I am seeing is that Facebook design seems to a baseline layout for social media sites. Too bad Sharepoint is not making a dent from the design site. Oh well does Microsoft do anything well?...oh wait they make money well.

Tuesday, December 14, 2010

It's been awhile...now Semantic Web is DOOMED!

It's been awhile since I last wrote my blog. I start off by saying that "Semantic Web" is dead. I don't think it will ever pick up and it's buzz has been replaced by "Social Media" and News Aggregators. Purpose of Semantic Web is to write code so that machines can learn and make inferences. For example I want to fly to Tahiti and I get my information on a web page which is written in OWL* (which includes Lite, DL and Full) and the machine via the internet is going to give me the best flight rates. This approach seemed great in 2004 however with the phenomenons of social networking and graph mining have squashed that fantasy. Don't get me wrong, Ontologies and OWL files are still needed in research oriented labs however they are not for the common users. Semantic Web promised us rich logic in the XML document which is parsed by the machine. Social Networks state that humans can relate to other humans and they are smart enough to make the right choices. I, personally, would rather trust someone in my social network to tell me about the hidden restaurant in Tahiti, where they serve grilled island crab and a glass of coconut milk, rather than an Oracle database churning on a RedHat Linux server somewhere on the cloud. With smart devices like iPads and iPhones, folks would rather view a YouTube video than to have a semantic web enabled application which is hogging up the device resources.

After looking at the trends, I have to say bye to "Semantic Web" and hello to "Social Web". Speaking about "Social Web", what do you guys think of Rock Melt?

Image taken by happy.apple and stored on Flickr

Saturday, May 15, 2010

Wisdom of Dan Meyer

Couple of days ago my colleague sent me an inspiring video from the TED Talks. The video was called, Dan Meyer: Math class needs a makeover, and it was about the flaws in the way math was taught in the US. Mr. Meyer was dead on with the problem.

After which I spent a couple of days thinking about it and realized that this was also the case with Information Technology (IT) in many companies. Most executive management and middle management want to address most if not all IT problems with tools. Here are some commons answers I have heard for major problems:
  • Need an Enterprise Architecture Capability within our organization
  • Let's buy an Enterprise Architecture Tool
  • Need to understand the current state of sensitive information in our organization
  • We need a tool to do analysis
  • Need an Application Portfolio Management (APM) Capability within our organization
  • Let's buy an APM tool or upgrade our existing APM tool
  • Need to pick a technology and product to address significant IT problem
  • Let's buy the recommended tool from Gartner or Forrester Research
  • Need to introduce metrics for IT in our organization
  • Let's buy or build a Dashboard tool.
As you can see, the formulation of the problem and ability to formulate a strategy on how to solve the problem are replaced by simply buying or building a tool.

It's time to stop watching "Two and a Half" and meet with your customers and understand their problems and anticipate future problems from the customers' broken processes.

Wednesday, March 17, 2010

EA can be so many things!

I have been working as an Enterprise Architect (EA) for the last two years however I am currently running the EA program within my organization. I have been involved in public sector EA for the last five years. I started off in the XML space, eventually moving into the Services Oriented Architecture and now primarily looking at all aspects of EA. EA can be challenging, fun and fulfilling but in the end it is like any other architecture disciplines, an art form. I believe to be a good Enterprise Architect or a Chief Enterprise Architect, you need:
  • To have a vision. Enterprise Architects have be visionary. If you don't have vision then it is time to call yourself a Enterprise Engineering. Architects are not problem solvers but they are more than that. They strategically anticipate problems, risks and provide mechanisms to mitigate them.
  • Be a hard worker. People think that Enterprise Architects live in a world on conceptual models and your canvas is the abstract world. WRONG! Enterprise Architects need to know strategy, technology, business processes, data and other domains and should be able to connect the dots. If they don't know a certain domain then they need to have a trusting relationship that domain SME. For example my knowledge in network engineering is quite light but I know folks who can design a network in their sleep.
  • To communicate. Communicate with your customers and see how EA will help them. Communicate with other architects and share thought processes. Understanding and disseminating ideas, processes and methodologies is a key.
  • Think on your feet. A good EA should be able to convince his point by using any information he has in his arsenal like data models, road maps, regulations, compliance, governance, applications, finance, etc., etc.
  • Understand your opportunities and boundaries. A good EA can easily identify opportunities that will enable him or her to promote their vision. He or she should be identify land mines and avoid them.
  • To have allies in all parts of the organization. Since EA doesn't have a good view in an organization, you need to build allies in different parts of the organization. This way you have a pulse on the organization and evolve the implementation of your EA vision.
The beauty of EA is that it can clearly identify gaps, issues and opportunities but it takes a savvy architect to easily identify it.