Monday, July 29, 2013

Will CMS Really Stop Terrorists?

(This piece first appeared on  

The Indian government plans to spend $132 million on setting up its brand new Central Monitoring System this year. Several articles have raised valid questions about privacy violations, including this one by Danish Raza. Elsewhere, Pranesh Prakash has raised important points about how CMS may actually violate several laws and at least one Supreme Court verdict.

I ask a much more basic question: will CMS work? Can it really help security agencies eavesdrop on criminals and terrorists, despite several known technical hurdles?

In 2008, a prominent Brazilian banker and investor named Daniel Dantas was arrested and charged with money laundering and tax evasion along with a former mayor of Sao Paulo. For five months, the Brazilian National Institute of Criminology tried to read the contents of his hard drive but failed to crack it. Dantas had encrypted his data using a free program called Truecrypt. The INC sent the hard drive to the FBI in the US, which spent a whole year trying to crack it; it too failed. Dantas’s use of encryption likely helped him escape the money laundering and tax evasion charges. He was ultimately convicted of attempting to bribe a police officer.

This story illustrates a fundamental loophole at the heart of CMS. A criminal, using free and easy-to-use software, can protect his data from even the most advanced surveillance tools available in law enforcement. NSA whistle blower Edward Snowden himself used encrypted email to communicate with journalists at the Guardian. In an online chat where he took questions from the public, Snowden noted that encryption was “one of the few things that you can rely on” to protect you from the eavesdropping behemoth created of the NSA.

It should hardly be surprising then, that terror groups have been encrypting their emails and data for at least the last five years. In fact Al Qaeda developed its own encryption software called ‘Mujahideen Secrets’, to encrypt emails, chat sessions and files. Version two of Mujahideen Secrets even included a tool to delete files securely so that they could not be recovered using special software if the computer was captured. Al Qaeda’s links to several terror groups operating in India has been widely reported in the past. It is not inconceivable that they have shared their encryption software with their comrades-in-arms.

Over the years it has become easier to encrypt one’s communication. YouTube tutorials train even novice users to set up email encryption within minutes. Phone calls, text messages and online chats can also be encrypted with free, easy-to-install apps.

The biggest problem with encryption is that it is virtually impossible to break the code in a time frame that’s useful for law-enforcement purposes. Without getting too technical, modern encryption relies calculating the prime factors of very, very large integers. In 2009, a group of some of the world’s best-known mathematicians and cryptographers reported that it took them four years to factor a 768-bit integer. They estimated it would take 1,000 times longer to factorise a 1024-bit integer. GPG, which is the most widely-used email encryption software, allows users up to 4096-bit encryption. Unless you have the password to the encrypted files, it would take you a very long time to crack the encryption.

Here’s an example to help you understand why encryption makes CMS redundant. Let’s say the system intercepts an encrypted email sent by a LeT handler in Karachi to a sleeper cell in Mumbai. The email contains instructions to detonate a bomb in a specific market at a specific time four days from now. Even if India’s intelligence agencies managed to link up every computer they had available to process the encryption, they would still not be able to crack it in time to learn the details and stop the attack.

What about ‘Metadata’?
It should be noted that encryption only protects the body of the email. The metadata, including the sender’s and receiver’s email addresses remain unencrypted, else the service provider would be unable to send the email to its destination. Law enforcement agencies often partner with email providers to track down the exact computer on which tell-tale emails were read.

However, this method of tracing criminals has a limitation. Programs such as TOR and Hotspot Shield disguise the IP address of a user’s PC. For example, when I use TOR, Facebook will often ask me to confirm my identity as it sees me as logging in from an unfamiliar location. TOR has thousands of servers around the world through which it bounces your data before sending it to its destination.
There is another limitation to using metadata. Due to obvious legal hurdles, CMS will only be deployed to capture communication within India. If terrorists were planning an attack from elsewhere in India’s neighbourhood (as happened with 26/11), we would have to rely on that country’s intelligence services for an alert. Good luck with that!

To make untraceable phone calls, terrorists have been known to use “burner” phones. These are pre-paid phones that are easily available in the US and other countries that do not require an ID for such mobile connections. They can be topped up using cash, which makes their prolonged using even more untraceable.

Even if CMS allowed spooks to listen to these calls, it would not be able to tell who was talking to whom. From details that emerged following the Abbottabad operation that killed Osama bin Laden, we also know that terrorists have been trained to turn off their phones and remove the battery to prevent being tracked even while not on a call.

So what is CMS good for?
If terrorist communications can easily be hidden from CMS, you have to wonder why the government is going through all the effort and expense to set up such a system. What good can come off the mass hoovering of data of ordinary citizens’?

Imagine if CMS intercepted a ‘BBM chat’ between two businessmen, who were discussing a contract that could affect the business interests of a government MP.

Imagine the government getting access to emails exchanged between a journalist and a source in the IAS who wants to expose a major corruption scandal involving a cabinet minister.

Imagine if the government had access to phone calls between two opposition politicians discussing election strategies.

What if CMS tracks a PhD candidate who is researching Naxal terror and has downloaded Naxal pamphlets? What if this researcher has been able to establish contact with Naxals for an interview. Can the government use such data to charge him with participating in a Naxal conspiracy, even if his only intention was to research their motivations? In a country where chief ministers label their critics as “Naxals” for merely raising questions, are we certain we want such unmitigated power in the government’s hands?

These are all questions well worth asking, especially since the ostensible reason for setting up the CMS—monitoring terrorists and criminals—is a fool’s errand at best.