Smart City Consulting - my blogs

Smart City blog posts links-

1) What Indian government should consider for planning Smart City

2) Planning For A Smart City

20) Tackling Top Telco Fraud's with Big Data and Complex Event Processing - Part- II

I am going to share a sample architecture that will help implement solutions to prevent fraud. The architecture combines Big Data technologies with Complex Event Processing (CEP) to provide a Smarter, Proactive system along with a Dashboard that will 'show relevant fraud events or potential fraud events' which will help prevent fraud.
The high level diagram of the Big Data-Complex Event Processing solution is self explanatory and the key points to be noted are :
1) Technologies  & products implemented in sample solution
2) Real Time Architecture to handle large amount of data
2) The way data in ingested, filtered & crunched
3) The way events are 'acted on' on-the-fly
4) The operations dashboard that will provide alerts and notifications so manual checking will be reduced
5) How analytics is going to be plugged into this architecture

Please feel free to mail me if you have questions.

19) Tackling Top Telco Fraud's with Big Data and Complex Event Processing - Part- I

Those of you who have not worked for a telecom company would be surprised to learn about the 'different type of fraud' faced by the telecom industry. Frauds are biggest cause of revenue loss to telecom companies and in general the top 5 types of fraud that a telecom faces are as follows (priority might differ for different  telcom companies)-

1) Sim Card Cloning Fraud : 

Fraud happens because there is no efficient way to check cloning, CSP have to perform manual checks to detect cloning, not efficient

2) Subscription Fraud Using Fake Identity: 
Subscription fraud involves the acquisition of telecommunications services using stolen or false credentials and/or identity  with no intention of paying. With subscription fraud, service providers lose revenue.
3) Roaming Data Fraud : 
Delay in getting Customer Roaming Call Data from roaming partner to home network paves the way for fraudsters to make roaming calls resulting in financial loss for the CSP, which is categorized a roaming fraud.

4) Internal fraud activity :   
Internal manipulation of the system, Account adjustments via Voucher Administration Terminal & Tampering with billing/rating systems are frauds that can be done by people who have access to systems
 
5) Frauds related to prepaid telecom services: 
A retail agent or call center agent may attach a value-added service (VAS) to an unsuspecting subscriber. For example, a ringtone can be added without the customer’s knowledge or permission, resulting in a commission for the agent. Currently there is no effective means to analyze the data and detect such kind of fraud. 

18) Telcos, challenge of Big Data & even Bigger Challenge of Telecom Fraudsters

Big Data brings an interesting set of technologies like Hadoop, Sentiment Analytics, Predictive Analytics. Big Data also brought a new generation of solutions. The only challenge as I have mentioned before is that technology & business need to sit together to analyze and identify the 'right' use cases before investing in Big Data.
      I have implemented Complex integration solution for telcom companies and realized that telco is a world by itself. An outsider will never realize the complex working of a telco and the huge volume of data that they generate! Telcom companies have always been in forefront of adopting latest technology innovations and rightly so. The ever growing customer base & network of Telco world generates humongous amount of data. Telcos requires their software to crunch data faster, filter out the noise & process faster.  Big Data makes way for Distributed Data Processing hubs for various Telco use cases. I am listing a few use cases that I have discussed with some of the top think tanks and innovators of Telcos.

Telecommunication Companies, Fraud Prevention & Big Data

One of the biggest challenge faced by telecom companies is fraud. Telecom companies loose more money to fraud then one can imagine. The fraudster use innovative technology to commit fraud so to detect and prevent fraud we need to use Predictive Analysis coupled with Big Data Processing on the large data generated by each phone, each switch, each tower, every second.

Take a quick look at some facts -

1.Telecom fraud is estimated  at $40 B globally and it is the single biggest cause of revenue loss for operators, costing them between 3% and 5% of their annual revenue. With rising competition & extremely low average revenue per user (ARPU), detecting fraud and plugging revenue leaks have become extremely important to reduce costs.


2.One study reports that the internal fraud (40.3%), roaming fraud (11.4%), pre-paid (10.8%), subscription (11.6%) and premium (13.1%) are the most important in terms of losses by values. 

3.Fraud connected to prepaid accounts is much easier to commit and harder to combat, since there is very little information on the subscriber, unlike postpaid accounts, where a credit check is usually done. Entry-level fraudulent activities such as subscription and impersonation are very serious since the cost is coming straight from the bottom line in the form of commissions and incentives.

4.The fraud management becomes more and more important as the new methods of access become available such as Cable networks,  Wireless networks, DSL, Satellite, Metropolitan Optical Networks running Ethernet, Broadband Wireless Systems (radio, microwave, or infrared).

5.Although there is an abundance of data generated by mobile devices and systems a large amount of data is not processed in real time.

6.Telco would like to detect critical events and patterns across all its data sources in real time, perform advanced in-memory analysis in real time and take preventive or corrective action in real time to providing better service to its customer and reducing the financial looses