It didn't start (and doesn't end) with OPM

The lax security that made OPM an easy target permeates government, says ICIT study.

Shutterstock image (by wk1003mike): open security lock on a computer circuit board.

(wk1003mike / Shutterstock)

What: An Institute for Critical Infrastructure Technology report titled "Handing Over the Keys to the Castle: OPM Demonstrates that Antiquated Security Practices Harm National Security."

Why: In its report, ICIT breaks down the fallout from the biggest federal cyber hack in U.S. history.

Lessons learned: multifactor authentication, activity logging (which is tough to non-existent on OPM's legacy systems) and user behavior analytics are crucial.

Exfiltrated data isn't the only concern – if OPM can't trace the path and actions taken by hackers in its systems, it won't be able to detect malicious changes (such as giving a security clearance to someone who shouldn’t have one). Better logging would have made breach cleanup easier, the report argues.

The report also cautions against learning the wrong lessons from OPM.

For instance, while OPM officials have correctly stated that encrypting files wouldn't have helped deter hackers in this particular breach, agencies shouldn't ignore encryption, because it can be useful in other types of intrusions.

The report notes that, while lawmakers are fingering China, OPM's security was so atrocious that basically any hacker who knew how weak it was would have rushed to attack OPM. The list of possible culprits is vast because "rudimentary" attacks, including spear-phishing and decoy websites, facilitated the breach.

A list of future potential targets is similarly vast.

The report notes nearly every other federal agency, from the IRS to the Department of Transportation, suffers from some or all of the security deficiencies that marked OPM.

The report calls on agencies to seize the moment to overhaul their security postures – but notes they'll need funding infusions to do it right.

Verbatim: "By far, the greatest failure at OPM was their lack of comprehensive governing policy. Governance controls such as disabling any old or unused accounts, limiting account functions and access to least privileged, limiting the login time of accounts, constraining and defining acceptable access times, limiting the ability and function of remote access, and mandating regular patching and modernizing systems. The number of privileged accounts should be minimized, when possible, the access of each account should be restricted to a fraction of the total systems, and the implementation of change review by other privileged accounts, should be considered. … Multifactor authentication…would have stopped attacks like this from ever realizing success."

NEXT STORY: Q&A: Rep. Ted Lieu