Here's what agencies need to know about the next phase.
Federal agencies are off to an extremely promising start deploying an anti-spoofing email tool but they still have work to do.
The Homeland Security Department issued a binding operational directive October 16 requiring agencies to secure their email through DMARC and STARTTLS, and web pages through HTTPS by January 15.
When the mandate first emerged, only 18 percent of the 1,315 federal domains had a Domain-based Message Authentication, Reporting and Conformance—or DMARC—record. In just three months, that number has more than doubled to more than 50 percent.
The federal government now has a higher rate of DMARC deployment than almost any commercial sector we’ve looked at, including the Fortune 500 (34 percent), major U.S. banks (32 percent), and even startup “unicorns” (31 percent).
Based on this trend, I expect the vast majority of the government’s domains will have DMARC records within the next few months, even as some agencies miss the January 15 deadline.
It Doesn’t Have to Be Difficult
Implementing DMARC, especially in monitoring mode, does not have to be costly, risky or difficult. Once awareness of that fact becomes widespread, I expect many more agencies will become compliant with the directive’s initial requirement to publish a monitor-only DMARC record.
DMARC can be implemented on any domain with a simple five-minute copy/paste, but given the change control mechanisms that govern Domain Name System updates in most organizations, that could realistically take several days to complete. The DMARC record has no impact on other DNS services (such as the availability of the domain’s web servers) and, as long as the policy for the DMARC record is set to “none,” for monitor-only mode, it will have no effect on whether email messages get delivered.
The most basic DMARC record also allows domain owners to specify an email address to receive DMARC aggregate reports, which provides an invaluable tool for collecting data on how the domain is being used by email senders. These reports can either be sent to the agency itself for parsing and analysis, or they can be sent to a third-party DMARC vendor.
When agencies turn on DMARC reporting, they will begin to see exactly which IP addresses have been sending email using the agencies’ domains. Using a third party can provide the agency with easier to read analysis, including the resolution of IP addresses into human-readable email services that are sending on behalf of the agency. Phishers who are trying to impersonate the agencies with fraudulent emails will also show up in these aggregate reports.
The Next Step: Enforcement
Making sense of the aggregate reports and figuring out exactly which services to whitelist and which ones to block is more challenging, of course. Even with perfect visibility, there will be internal sleuthing needed to determine which email services are authorized to send on the agency’s behalf, which are clearly fraudulent, and which may be “shadow email services” (services that are not officially sanctioned nor vetted by the agency, but instead were turned on by a well-meaning but unauthorized employee). Federal agencies have until October 16, 2018 to meet that challenge: That’s when the DHS directive requires agencies to set their DMARC policies to “reject,” which means mail servers worldwide will reject all messages failing authentication.
A policy of reject or quarantine (which directs failing messages to spam folders) is what we call email authentication enforcement because it lets domain owners control who can send email on their behalf and who gets blocked.
Done properly, DMARC at enforcement allows an agency to whitelist mail servers and cloud services that it wants to authorize while blocking everything else: phishers, spammers, and even unapproved cloud services that the well-meaning staffer might have set up without going through the proper vetting and compliance checks.
Until October 16, agencies can keep their DMARC policies at “none,” simply monitoring the DMARC aggregate reports and gathering data as they plan the move to an enforcement policy.
We know from our experience in the commercial sector that DMARC enforcement can be challenging. The numbers show that on average 70-80 percent of all DMARC implementations fail to achieve enforcement within nine months. There are three main paths that agencies can take over the next nine months: do-it-yourself, hiring a consultant, or automating DMARC—all with varying degrees of success rates and levels of effort.
For now, agencies simply need to publish a basic DMARC record and begin monitoring the data on who’s sending on their behalf. That’s a valuable first step, gives each agency actionable data and awareness, and it’s eminently achievable.
Alexander García-Tobar is the CEO and co-founder of ValiMail.