How can we secure the smart grid?

I recently gave a talk on how power electronics will play a vital role in the creation of smart grids, promising to improve reliability while promoting a more distributed generation architecture coupled with a sensing/controls/communication backbone, such as that proposed in [1] referred to as Gen2PG, the second-generation power grid. One of the questions from the audience at the end of the talk was in reference to the security of a smart grid that relies on Internet-like technologies.

The public’s general perception is that such technologies are error-prone, unreliable and a security risk. Perhaps this perception is one that has been derived over the past two to three decades, as personal computers have become more ubiquitous and people’s experiences with identity theft, viruses and spam have made them very aware of how vulnerable they are. When you couple these perceptions with the additional bad news of irresponsible behavior from those in critical positions involving either security or oversight, there is clearly a reticence among the public to having any information access behind their meter.

After some reflection, I decided to address the security question here in this forum. Security threats on the smart grid could include: those who would seek to take control of the grid; disrupt the grid’s operations; defraud the electric power companies through either electricity theft or embezzlement of funds; or steal customer data, identities or funds. Monitoring of a home’s power signature would be an indication of absence, which when taken together with other simple surveillance would leave the residence easily targeted. Along these lines, NIST recently released a second draft of its Smart Grid Cyber Security Strategy and Requirements that identifies more than 100 interfaces that pertain to smart grid applications and classifies them according to the level of damage that could result from a security breach [2].

Perhaps the place to start is to acknowledge that no system is failsafe if someone on the inside compromises the system. However, even then, checks and balances can be put in place to quickly detect unauthorized use and access. The threats outlined above can be generally separated into those that involve the personal information of the customers and those that impact the grid’s normal operation.

As long as systems remain physically distinct and separate, security can be guaranteed. For example, if my home computer is not accessible by the smart grid communications protocols, then it cannot be hacked. However, as an example, if I choose to add the hardware and software to communicate from my home computer through the smart meter, say over power lines, to the power company to pay my electric bill, broadcast usage data, etc., then I have opened myself up for the possibility of attack in a fully integrated system. My illustration serves to provide my first recommendation on smart grid security: Let the electric power companies figure this out first for their operations before jumping into the home.

Billions are being spent by government and industry to tackle this problem. In fact, in late 2009 a university consortium led by the University of Illinois was funded by U.S. Department of Energy and the U.S. Department of Homeland Security to focus on just this issue [3]. The center, known as the Trustworthy Cyber Infrastructure for the Power Grid also includes Dartmouth, Washington State University and the University of California at Davis. Undoubtedly, this effort along with electric power utilities and third-party grid security firms will address the operational needs of grid security first. Based on these solutions, we as homeowners can then begin to assess the degree to which we are comfortable opening up our side of the meter.

So the answer to the title question of this article is to secure the hardware and software for grid operations first, then focus on those additional items needed for the home. Security of such vital national infrastructure as our electric power grid is of serious concern to everyone. Before the smart grid rolls out too far, strong solutions must be deployed at the hardware and software levels to address anti-tampering. This will include “grid-qualifying” technologies such as the communication chipsets, microprocessors, software, control algorithms, power electronics, communications protocols and data encryption algorithms. This qualification will have to address vulnerabilities associated with both built systems, components of systems as they are manufactured, and how these components and systems detect and respond once attacked, including the ability to determine whether an attack has compromised a system and made it more vulnerable to a future attack.

The good news is that we have learned a great deal from the computing attacks of the last 30 years. The bad news is that we still have much to learn to expand sound security to such a geographically diverse power grid to prevent widespread disruption.

[1] H. A. Mantooth, R. Dougal, “Center for GRid-connected Advanced Power Electronic Systems – GRAPES,” IEEE Power Electronics Newsletter, vol. 24, no. 1, pp. 29-31, 1^st Quarter 2010, http://www.ieee-pels.org/publications/newsletter.

[2] http://www.nist.gov/smartgrid

[3]   J. Applequist, “U.S. Departments of Energy and Homeland Security Establish Major Resilient Smart Grid Program at the University of Illinois,” http://cs.illinois.edu/news/2009/Oct26-2, Oct. 26, 2009.

Mr. H. Alan Mantooth is an IEEE Fellow and has published numerous articles on models, modeling strategies, and electronics design and holds patents on software architecture and algorithms for modeling tools.