25 July 2017

MUCH ADO ABOUT NOTHING? CYBER COMMAND AND THE NSA


Last week, word began to spread that the Trump administration was considering granting new powers to U.S. Cyber Command. Lolita Baldor of the Associated Press had the scoop, discussing two related but separate steps under consideration: first, to elevate U.S. Cyber Command to the status of a unified command and second, to break the current “dual-hat” arrangement with the National Security Agency (NSA), whereby the commander of U.S. Cyber Command is the same individual as the director of the NSA.

It is worth noting, however, four things: First, these two steps (elevation and separation) have been under consideration for years. Second, there were good reasons at the time why the Obama administration didn’t act on them. Third, elevation and separation should, in theory, operationally empower U.S. Cyber Command, but in practice Cyber Command may ironically find itself with less capability to offer. And finally, Cyber Command has already quietly amassed non-operational power and authority within the Department of Defense, making it one of the most independent commands, second only to the U.S. Special Operations Command. As such, while this weekend’s news is a good sign of the continued maturation of Cyber Command (and the acknowledgment of that maturation by the White House), there’s less here than meets the eye.

Let’s review Cyber Command’s origins and its assigned missions before tackling the news. (Please accept my apologies in advance for some acronym salad.) For the short-story long, see chapter 8 of Playing to the Edge by Michael Hayden and the early parts of Jay Healey’s Fierce Domain. Long-story short, the NSA had been the nation’s leading signals intelligence agency for decades. But after 9/11, as new opportunities emerged to create effects against adversaries during declared hostilities, Pentagon leadership became uncomfortable with the notion that the intelligence missions of collection and analysis would be conducted by the same organization that would disrupt or degrade, even destroy, targets through cyber-attacks during an armed conflict. In 2002, U.S. Strategic Command was given responsibility for cyberspace, and two little-known subordinate organizations emerged to manage it: Joint Task Force-Global Network Operations (JTF-GNO) would handle guarding the Defense Department’s networks while Joint Functional Component Command-Network Warfare (JFCC-NW) would be responsible for missions we’d think of as offense. Because there was so much overlap between the NSA and the emerging JFCC-NW, the Department of Defense created the “dual-hat” by making the NSA director (then Hayden) the commander of JFCC-NW. As the threats to the Department of Defense in cyberspace increased throughout the 2000s, Secretary of Defense Robert Gates consolidated JTF-GNO and JFCC-NW under a new U.S. Cyber Command in 2010, but it was still subordinate to U.S. Strategic Command and still dual-hatted with the NSA director. That’s more or less where we find ourselves today.

Since then, U.S. Cyber Command has been charged with three missions: defend the Defense Department’s networks and systems, provide offensive support to other commands in the event of a contingency, and defend the nation from a cyber-attack of significant consequence (less than two percent of incidents would qualify as “significant”).

Advocates of more autonomy and authority for U.S. Cyber Command have often bemoaned its subordinate status to U.S. Strategic Command. The theory is that having to work through Strategic Command slows down operational approval, coordination, or whatever else needs to happen. Based on my experience in the Cyber Policy office of the Office of the Secretary of Defense, I am of the view that a stove-piped Joint Staff had more to do with delays and miscommunication than anything else; nor could I ever find a function Cyber Command might be asked to execute that could only be performed by a full, unified command (like Strategic Command) but not by a sub-unified command (like Cyber Command). We looked at this several times during the last administration: If the secretary of defense wanted the sub-unified command to execute, they could and would. It wasn’t a problem, so elevating the command wasn’t necessary. So, while I don’t think there are any big wins to be had by the recent news about the Trump administration wanting to elevate Cyber Command, I don’t think it hurts to do it either. And it might not ultimately be up to the White House: The 2017 NDAA requires the administration to elevate Cyber Command.

Breaking the dual-hatted relationship with the NSA is more complicated. There are very good reasons why JFCC-NW was born with the NSA as its commander, as there is a lot of overlap between the organizations. This overlap is intuitive to those who’ve worked in the business, but hard to explain in brief here. I’ll just quote Hayden on this point: “[I]n the cyber domain the technical and operational aspects of defense, espionage, and cyberattack are frankly indistinguishable — they are all the same thing.” It’s obviously more complicated than this, but at a high level, I think this was the rationale.

There were studies undertaken about the implications of breaking the dual-hat before the Snowden affair, but his disclosures forced policymakers to confront the issue head-on. At that time, it was thought that breaking the dual-hat could improve perceptions about privacy and civil liberties at the NSA, but in December 2013 the Obama administration decided to maintain the arrangement. Senior leaders felt it was too soon to separate Cyber Command. Its readiness and resources were growing but insufficient, and it was still too reliant on NSA talent and services for its missions.

Working with the two organizations, I found that the relationship between the two was akin to a mix between hostage-taking and Stockholm syndrome — except each organization kept mixing up which was the hostage and which was the hostage-taker. One day, U.S. Cyber Command would demand NSA support due to the latter’s responsibility as a combat support agency. The next day, the command would cave and say that NSA had other, more important priorities. And NSA too would resist a request from Cyber Command, then embrace it, and then fight it. The overlap and dependence was that tight.

For that reason, among others, I understand the argument about needing to separate Cyber Command from NSA so that the former can pursue its missions (especially to defend the nation and to support other commands) with greater independence from signals intelligence. But there’s a risk here that would be dangerous to miss: When Cyber Command needs NSA support, the fact that it’s the same person in charge of both organization can break what might otherwise be a log-jam. Splitting the dual-hat could result in the NSA isolating itself and refocusing on its own core missions (the collection of signals intelligence and providing information assurance) while minimizing its support to Cyber Command.

Just because there are risks does not mean the Trump administration should leave the current arrangement in place. The question is not whether, but when and how, to break the dual-hat. One priority for the White House and Secretary Mattis will be to have a clear understanding with the new NSA director (who may well be a civilian for the first time) about how he or she sees the relationship with Cyber Command, and then how the administration monitors the relationship to ensure the NSA doesn’t abandon Cyber Command outright.

The selection of who will next lead Cyber Command will also be a priority. Someone like the current commander of Army Cyber Command, Lt. Gen. Paul Nakasone, is an ideal candidate: He has years of experience in the cyber effects business, time in the Pentagon and the field, and he understands the roles of civilians, fellow military officers, and senior political types. Another name that’s been floated is Lt. Gen. William Mayville, currently the Director of the Joint Staff. His time as the Joint Staff’s chief information officer and with Joint Special Operations Command would make him a strong leader for Cyber Command as well.

The good news for the future of the U.S. military’s cyber operations is that, regardless of whether or not Cyber Command is elevated as a unified command or separated from the NSA, Congress has quietly been empowering Cyber Command with greater authorities and independence through legislation. My colleague Charley Snyder and I assessed all the additional powers conferred in the 2017 NDAA over at Lawfare, but I’d like to single out the authority related to requirements: Being able to set its own requirements for the conduct of cyber operations, as well as validating the requirements of other defense components, matters more than this bland bureaucratic language might suggest. With the independent acquisition authority Congress gave it in a previous NDAA, Cyber Command can now accelerate acquisition and procurement to keep up with new requirements without the usual deliberations chaired by the Joint Staff. Special Operations Command is the only other military outfit with that kind of freedom, and it makes a big difference.

But the big question will be this: Regardless of these crucial authorities and any new command arrangements, what will Cyber Command’s role be in protecting the country from threats like Russian information operations? Maybe it’s time we get away from using “cyber” as the description of what needs to be done, and instead think about what an Information Warfare Command would look like. How should the United States wage such a fight, and how should it protect itself? I am pleased the Trump administration is considering organizational changes to support a higher profile for cyber operations, but we really need answers to these bigger policy questions.

Michael Sulmeyer is the Director of the Cyber Security Project at the Harvard Kennedy School’s Belfer Center for Science and International Affairs. He also served in the Office of the Secretary of Defense, Cyber Policy, from 2012-2015.

No comments: