Plug and Play Databases

Plug and Play Databases: A Mythical or Near-future Technology Proposal

This article is again one of my humble attempts to take everyone’s attention to the fact that the Big Bang in IT industry is at our front door. There is no escape in facing the BigData issue whilst your in-house systems are outputting more and more bits at any time. In the most innocent way, no matter where you are or what you are doing, you are the subject of an event tracked by some computing machine such as a GSM transmitter. The scenario is getting complicated when you want to look back the traces of information left behind by your systems. The best case appears if the systems are able to talk to each other in the same lingua. Imagine the worst case in which there is a mess with full of inconsistent, repeated, unidentifiable data being signaled in the air.

Whatever you would do in that case is called yet another reaction. We all have our reasons to defense the information chaos surrounding our corporates. One may complain about the diversity and multiplicity of applications whereas other may criticize the poor communication among the divisions. The most common reaction mechanism applied is building enterprise-level integration systems or buying one, then fingers crossed. It does not mean that they do not work as they supposed. I would like to spot on the main point or root cause of the problem, nature of the data.

Networking vs Adapting

At the time Tim Bernes-Lee introduced us the World Wide Web, textual and media-oriented documents were melted in a big bowl and spread over the world as it is still happening. The trouble is of course not the contents of the bowl but the standards and machine-readability. Take this article for instance, can you jump in another page when you click on any word unless I explicitly anchor it to a URL? Possibly not if your browser is somehow running like IBM’s Watson (Actually I am not sure if Watson can do that though it is the ultimate Jeopardy master ever). For now, let’s put that problem in our pocket: What you see is what you get! You should ask your favorite search engine if you need more.

Stepping into the second wave, the Web 2.0, brought us a new discovery, the social web. Clicking on “Like” or “+1” buttons, connecting the friends and getting them into the circles, catching new opportunities and keeping in touch with BizWorld and finally an intention to know about what your fellow is doing or thinking. There may be countless services here I could not recall but I am pretty sure that they should make available magnificent services for their fans and adding value to their lives. So what’s the problem here? In the age of information, people who adapt those changes are sticking to one or more social platforms where they see the jobs get done. It is a good thing to see people sharing things with each other and finding alternative communication ways. It is a fantastic collaborative effort. On the flip side, the famous buzz word BigData comes on the way leveraged by billions of events and transactions occurring in real-time social web. Anyone who is interested in the information in his/her allowed sandbox(or relevant realm) can extract that piece of information through services provided. So, when you look at deeper, each social web application is actually an information nebula with stroked boundaries. But the point of interest is not again the data, it is the service you ask for. Consequently, there are various types of service protocols or so called as APIs which let the applications talk to each other. We should accept that there will be no one-only cloud in the future. It looks like a fact. But there ought be better ways for striking from one to another. That’s the second obstacle or problem I put in the pocket: a networked world solely based on service exchange.

Mounting the Cloud

If it is time to call back the title, the Plug’n Play technology seems to be a convenient metaphor for our solution investigation. Here is the Wikipedia definition of what Plug and play is:

“In computing, plug and play is a term used to describe the characteristic of a computer bus, or device specification, which facilitates the discovery of a hardware component in a system, without the need for physical device configuration, or user intervention in resolving resource conflicts.”

It is admirable to see how your environmental devices such as USB drives can seamlessly be part of your original environment using the Plug and Play interfaces. The idea is upon auto-configuration, auto-discovery and zero intervention.

In the perspective of data clouds, each cloud may represent a single entry point of service where any other cloud can get into. You may be humming by asking “now another service?”. No, it is already there, the HTTP service could be what we only need. URI is already defined as a standard to locate anything where it dwells. The only thing necessary is thinking in data-oriented. Getting machines to really read the data with its meaning could be achievable if and only if clouds store, define and preserve the data with what they stands for regarding to the shared knowledge-base as known as an ontology.

Above-mentioned clouds can vary from private or public ones. But the solution lies on naturally linking the data not using synthetic services or adaptor mechanisms. Legacy systems might be seen as missed the train. Actually, we all miss the train if we still insist on discarding the interoperatability as a must-have characteristic of the systems we built.

A Proactive Approach to Data Integration Problems

Where to start from? This is not an easy question and there is no silver bullet. However, luckily we have some standards and technologies let us drive our architectures in a model driven way. Here, I do not cross the well-known Model Driven Design or Architecture paradigm. It is all about any piece of data not just business objects. One of the good starting points could be investing on ontologies. Take your data warehouse and examine what you have got and delve into the details. Extract main entities (or classes) and highlight the relationships between your entities. There will pop up many raw values which represent your properties. After all, you need to search the literature or ontology banks. There are many researches around the world that common vocabularies are emerging and being shared. Social and financial ontologies are almost complete and many others are in progress. Pick up the shared ontologies to be applied for your system. After that,there are two options (as far as I know) you may follow up. (1) Build up a native data-centric, linked data set if you start from scratch, (2) keep your existing data sources as they are but decorate them with a semantic layer on top to link between the underlying data and your standard ontological definitions. Resulting system will lead an interoperable and extensible infrastructure to the neighborhood. Data will be serviceable with a payload of what it means.

The Audience

My primary target was the CIOs or any related decision-makers who has having trouble with their corporate data integration systems. Usually the audience paragraph is put somewhere between introduction and the first section. But it might be discouraging for the ones who would like to see the benefits and opportunities in semantic web or -as my metamorphic naming- the Plug’n Play Data Sources on the article, but not his/her title is listed in this paragraph.

Thank you,

SEL�UK BOZDA�

KOMTA� Bilgi Y�netimi ve Dan��manl�k - Yaz�l�m Ekip Lideri

2002 y�l�nda Hacettepe �niversitesi Bilgisayar Bilimleri ve M�hendisli�i B�l�m�nden mezun oldu.
2002 - 2007 aras� Ankara'da yaz�l�m m�hendisli�i ve uzman yaz�l�m m�hendisi olarak �al��t�ktan sonra 2007 - 2010 y�llar� aras�nda yurtd��nda yaz�l�m ekip liderli�i yapt�.

2010 y�l�nda T�rkiye'ye d�nd� ve KOMTA� Bilgi Y�netimi'nde ArGe departman�nda ekip lideri olarak g�reve ba�lad�. Halen bu g�revini devam ettiren Sel�uk, kurumsal veri analizi yaz�l�mlar� ve y�ksek performansl� bili�im konular�yla ilgilenmektedir.

SEL�UK BOZDA�

Content on this page requires a newer version of Adobe Flash Player.