Google, Mozilla, Facebook & Twitter Toughening Its Security to Thwart Government Snoops
NOVEMBER 22, 2013, 4:22 PM
Twitter Toughening Its Security to Thwart Government
Snoops
By NICOLE PERLROTH and VINDU GOEL
Jacob Hoffman-Andrews, right, a security engineer at
Twitter, had been pushing the company to adopt forward secrecy for some time,
but did not get much support for the project until the recent revelations about
the National Security Agency's surveillance practices.
Noah Berger for The New York Times
A year ago, hardly anyone, save for cryptographers, had
heard of Perfect Forward Secrecy. Now, some customers are demanding it, and
technology companies are adding it, one by one, in large part to make
government eavesdropping more difficult.
On Friday, Twitter will announce that it has added
Perfect Forward Secrecy, after similar announcements by Google, Mozilla and
Facebook. The technology adds an extra layer of security to Web encryption to
thwart eavesdropping, or at least make the National Security Agency’s job much,
much harder. (Update: Twitter has announced the security change on its blog.)
Until Edward J. Snowden began leaking classified documents
last summer, billions of people relied on a more common type of security called
Transport Layer Security or Secure Sockets Layer (S.S.L.) technology to protect
the transmission of sensitive data like passwords, financial details,
intellectual property and personal information. That technology is familiar to
many Web users through the “https” and padlock symbol at the beginning of Web
addresses that are encrypted.
But leaked N.S.A. documents make clear that the agency is
recording high volumes of encrypted Internet traffic and retaining it for later
cryptanalysis. And it’s hardly the only one: Iran, North Korea, and China all
store vast amounts of Internet traffic. More recently, Saudi Arabia has been
actively trying to intercept mobile data for Twitter and other communication
tools.
The reason governments go to great lengths to store
scrambled data is that if they later get the private S.S.L. keys to decrypt
that data — via court order, hacking into a company’s servers where they are
stored or through cryptanalysis — they can go back and decrypt past
communications for millions of users.
Perfect Forward Secrecy ensures that even if an
organization recording web traffic gets access to a company’s private keys, it
cannot go back and unscramble past communications all at once. Perfect Forward
Secrecy encrypts each web session with an ephemeral key that is discarded once
the session is over. A determined adversary could still decrypt past
communications, but with Perfect Forward Secrecy the keys for each individual
session would have to be cracked to read the sessions’ contents.
Perfect Forward Secrecy was invented more than 20 years
ago, and Paul Kocher, a leading cryptographer, put support for Perfect Forward
Secrecy into the S.S.L .protocol. But companies have been reluctant to use it
because it slows website and browser performance, uses resources and because —
until Snowden — most consumers did not even know it existed. Unlike S.S.L.
technology, there is no indication to a user that Perfect Forward Secrecy is
enabled.
This tougher security is quickly becoming a must-have for
Internet companies.
Earlier this week, Marissa Mayer, the chief executive of
Yahoo, announced that Yahoo would introduce new security features in 2014. But,
on Twitter, some consumers were quick to point out that Perfect Forward Secrecy
was conspicuously absent from her blog post.
“With security, there are always the things you know you
ought to do,” Mr. Kocher said in an interview. “But it’s not until you have a
clear adversary that it’s much easier to justify the resources to go fix the
problem.”
At Twitter, Jacob Hoffman-Andrews, a security engineer,
had been pushing the company to adopt forward secrecy for some time, but did
not get much support for the project until the Snowden leaks.
That showed “there really were organizations out there in
the world that were scooping up encrypted data just so they could try to attack
it at a large scale,” said Jeff Hodges, another Twitter software engineer. “We
were like, oh, we need to actually spend some more time and really do this
right.”
Actually installing and turning on the technology took
only a few months, once Twitter decided to do it, both men said in an
interview. That was in part because Google, an early pioneer in the technology,
had worked out many of the kinks in Perfect Forward Secrecy and shared its
knowledge with the security community.
Perfect Forward Secrecy does add a slight delay to a
user’s initial connection to Twitter — about 150 milliseconds in the United
States and up to a second in countries like Brazil that are farther away from Twitter’s
servers. But the company said the extra protection was worth the delay.
Twitter said it turned on Perfect Forward Secrecy on Oct.
21, although it refrained from publicizing the change immediately to make sure
there were no problems.
Twitter said it hoped that its example would prompt other
companies to adopt the technology.
“A lot of services that don’t think they need it actually
do,” Mr. Hodges said.
Comments
Post a Comment