'Disparate applications pose challelenge'

Creating global data warehouse is an essential step to preserve data integrity, said Eddie Chandok, President, Infogain India. He spoke at length about data integrity in an interaction with B.V.Shiva Shankar, Associate Editor, CIOL, Cybermedia. Excerpts:

CIOL: Everybody is talking about Data security and integrity….

Eddie Chandhok: I differentiate between data security and data integrity. I have a slightly different understanding of data integrity. Data integrity can mean security; it can also mean cleanliness and appropriateness of the data. We in our parlance say the data integrity means the good data.

The business decisions that are made based on this data will be appropriate. At times we run into situations where there is lots of data, but its integrity is questionable. In the case that it comes from different sources, there could be issues with the alignment and completeness.

There is a big movement in the market to get more and more information about customers. Integrity, completeness is a big factor there and also validity of data. If I have poor integrity of data or poor quality of data and do not have appropriate processes for constantly upgrading completeness as well as appropriateness and correctness of the data, over time it will deteriorate.

CIOL: For Which vertical it is most critical?

EC: It is actually difficult to rank. Because every vertical will say that data integrity is important for us. It is obviously important in banking, health-care and health sciences. Think about medical records for example. If there is some inappropriate information about some test results, drug interaction, about past medical history, it can have a negative impact on your future. It applies to banking also. In banking data security is very important and the data leakage results in negative consequences.

I can say where ever data is important, data integrity is important. It depends on the mode and situation of your business. If it is a medical business, it is a life-threatening situation. It is not the case with a toy business or an automotive business. If you really want me to rank it, I have to say Life sciences are the most important space. BFSI will come next and consumer market is probably in the third place.

CIOL: To what extent mirroring of data is helpful in preserving its integrity?

EC: Mirroring of data does not help you either complete the dada or increase the correctness of the data, all it help you in disaster recovery perspective or access perspective. Some companies have one global data warehouse that puts you at risk. If there is a problem with that data warehouse for what ever reason you will lose the data, that's one aspect. Secondly, if have the data in only one location access becomes difficult, at times. Suppose, I am sitting in Asia-Pacific, it would be easier if my data warehouse were in Singapore as opposed to sitting in Texas. So mirroring gives you an automatic back up and it also gives you an easy access. But it will not ensure integrity.

CIOL: What is the role of standardization here?

EC:I think there is always a tussle between standardization and creativity. My feeling is that standardization is appropriate in certain areas. For example it is great that we could standardize the windows operating system. I think it has led to the advent and transferability of skills across companies. People move around between companies, people move around between countries is a lot easier. So standardization is good from that perspective.

CIOL: Tell us about the key applications you have deployed to ensure data integrity

EC: We have our internal data, the data that is used to run Infogain Company. Then we have customer's data that we have access to, and used to provide services to them. Specifically as far as infogain is concerned, we have multiple locations globally but we are trying to work towards completely global systems. We have only system of records that assures data integrity. But there are still problems in it, you know, people have the same kind of names, then we assure that the particular name alone is used. Internally for our company data is not an issue for us because the size of data is relatively small.

The second part is the access to data and that's where the challenge is a lot more from many perspectives. We have to ensure the secrecy of the data that we handle. Clients like HP and Cisco are very much concerned about their intellectual property and data. We have to ensure that there are internal processes to protect their data. We ensure data security. We do not allow our employees to access to other websites so that the data is not sent out through mails. These are a few procedures to ensure data protection.

Apart from this, we are working with many our customers to constantly upgrade the cleanliness, the appropriateness and the correctness of the data.

CIOL: what are major challenges in this regard?

EC: The major challenge for data integrity in most organizations, ours included, that the same type of data is stored in many different places. It automatically creates a situation where there is no one version of the truth. Your information is stored in four places, the name, and phone number, each in different systems. This is a great challenge for data integration. If there is just one place then it is easy, but there are many places where the data is stored. If you have two companies interacting with each other and you have more than one system, which has information about the same interaction, it leads to redundancy. This has to be managed.

CIOL: Current trends…

EC: Some extent in trying to create global datawarehouses. , which basically means that master information should be stored only once in any organization, like customer lists, vendor list, supplier list etc… There should be only one type of system that is performing a certain task. May be concept of SOA. In that concept there is typically only one system that provides certain service to everybody else as oppose to having more than one system providing the same services. Single services managed by one system, there is global data warehouses. You don't want to move away from those problems. But what has happened in the past is that companies have not centralized as much. They have run more de-centralized mode and as a consequence of that the different entities across organizations have built different systems around databases and you have. I think, CIOs today are getting a lot more control; they have been able to centralize this to some extent. Third trend is significant amount of outsourcing.