WASHINGTON – The United States, concerned that Iran is behind a string of cyber-attacks against U.S. banking sites, has considered delivering a formal warning through diplomatic channels, but has not pursued the idea out of fears that doing so could escalate hostilities, according to American officials.
At the same time, the officials said, the disruptive activity against the websites has not yet reached a level of harm that justifies a retaliatory strike.
The internal discussion reflects the complex nature of deciding when and how the United States should respond to hostile cyber-actions from other countries. It also reflects the pressure the administration is under from banking industry officials, who want to know what amount of pain or damage will justify a government response.
Administration officials say it is difficult and unwise to be too precise about potential responses because they do not want to set red lines that, if crossed, might obligate them to act.
“You’re always going to see the government be more cautious and incremental in response to most incidents than the private sector probably would like,” said Michael Daniel, the White House cybersecurity coordinator. “But that’s because the risk of misattribution and escalation is real, and we always have to consider the broad foreign policy implications of our actions.”
The White House does not want an incident that is “mostly annoying” to harm relations with other nations, he said, “or worse, result in conflict.”
This much is clear: The last eight months of disruptions to bank websites, caused by efforts to crash servers with torrents of computer traffic, have not been severe enough to trigger a military response, cyber or otherwise.
“Not even close,” said one military official. “But at some point, does it become a question of the public losing confidence in the banking system? That’s one of the questions, among many others, in the discussion of when a threshold is crossed.”
U.S. intelligence officials have said they believe the disruptions, called distributed denial of service (DDOS) attacks, are the work of the Iranian government.
More worrisome to the government and banking officials are destructive attacks that corrupt financial data or cause panic that can harm the economy. In one of the most damaging such incidents, a cyber-attack last summer wiped data from computers at Saudi Arabia’s state-owned oil company Saudi Aramco, rendering them inoperable.
Daniel said he believes companies need to do more to defend their own networks as part of a “spectrum of responsibility” that includes the public and private sectors.
At a March meeting of banking executives hosted by the Treasury Department, U.S. officials made clear to the chief executive officers that they could not simply rely on the government for cybersecurity.
Daniel said he thinks the business leaders are taking the problem to heart. “Now companies are going to have to take the next step,” he said. “Do you have a backup plan for your systems if they’re hacked? Did you actually test it?”
One key group whose role is still being debated are Internet service providers, whose pipes, cables and servers make up the backbone of the Internet.
Some banking officials say they would like the providers — including Verizon, AT&T and Century Link — to do more to block malicious traffic headed toward their networks.
Gen. Keith Alexander, the director of the National Security Agency, suggested that the Internet companies would be best positioned to block an Aramco-type attack with help from the NSA. The agency, he said, would provide the signature for the malware that could be used to screen network traffic to stop the attack. But doing so is fraught with legal, political and operational challenges. If a company is asked by the government to screen the traffic to stop an attack, it could be seen as acting as a government agent, exposing the firm to legal action.
Alexander, industry officials say, would like the Internet providers to be able to screen traffic entering U.S. networks, but privacy laws prevent them from doing so unless, for instance, they have customer consent or a court order.