http://www.computerworld.com/securitytopics/security/story/0,,91899,00.html Opinion by Marc Gartenberg APRIL 08, 2004 COMPUTERWORLD For the fourth year in a row, the federal government released its "Report Card on Computer Security at Federal Departments and Agencies" (download PDF) [1]. The average grade for fiscal 2003 was a D (65). The overall average grade in 2002 was an F (55); in 2001, it was also an F (53). Since 2000 was the first year that any measurements were taken, that year's score was "Incomplete" with a letter grade of D-. Looking at the situation through the lens of an eternal optimist (and realist), maybe, just maybe, agency heads, the Office of Management and Budget and Congress will start looking for ways to get these agencies where they should be. An empire in the age of technology can and should be able to get passing grades in information security. As an alternative to looking at the trends and drawing the conclusion that things aren't really that bad since, after all, the overall score is improving, let's examine instead the underlying factors that led to these scores. Then we can see why our dear Uncle Sam needs some help, and we can offer some suggestions. Through this analysis, it will become clear that the issues are related to establishing, maintaining and measuring enterprise security management strategy as part of the systems development life cycle so that no government agency or company ever has to settle for a D. Why the bad grades? To answer that, we need to examine the factors upon which the scorecard is based. These include certification and accreditation processes and recognize subtle distinctions in the categories of IT systems, namely general support systems and major applications (a.k.a. mission-critical applications), and realize that there are still many legacy systems long overdue for retirement. Ready? First, here's the process, which sounds simple. The National Institute of Standards and Technology (NIST), a component of the U.S. Department of Commerce, publishes and updates its policy guidance for information security. Federal agency security chiefs are supposed to see that these guidelines are followed within their agencies. The problem is that the NIST guidance isn't very concise regarding implementation. It also isn't an operational procedure manual. Rather, to a great extent, it's a higher-level management policy document. This creates a gap between knowing what to do and how to do it. Yet the scorecard rates an agency on how well it implements the guidelines. Then there's the reporting scheme, which is handled by each agency's own Office of the Inspector General. These offices are designed as semi-autonomous bodies operating within and under the jurisdiction of each agency head. Another factor is that this year agencies had to meet the requirements of the Federal Information Security Management Act (FISMA). This law expands on the information-security evaluation and reporting requirements enacted in 2001 under the Government Information Security Reform Act (GISRA). Under FISMA, agencies must demonstrate their progress in areas including risk management, contingency and continuity procedures to ensure that their mission-critical and general-support systems are protected. This includes annual IT security reviews, reporting and remediation planning on systems at all stages of the systems development life cycle. So while agencies in previous years were showing improvements on the standards according to GISRA, the fact that the regulations changed midstream caused many agencies to have a problem meeting the new mandates. When all these dependent variables are synthesized, you can see that getting an A isn't all that easy. Is it fission or fusion? Interestingly, though, FISMA has been a long time coming, and the federal security chiefs had fair time and warning before the sun set on GISRA. It's also noteworthy that the grades in previous years weren't much better -- how could they get any worse? Last year's grade average was an F. This year it's a D. So, maybe things are getting better. It's hard to say, since each agency, regardless of size, is given an equal weight in determining the overall average, so a couple of A's such as for the National Science Foundation and the Nuclear Regulatory Commission (whew!) helped improve the scores a bit overall. The ROI of security Federal and industry chief security officers will agree that it's hard to build a tangible case for increased security appropriations. This is primarily because it's hard to quantitatively justify increased spending on IT security because there's little tangible immediate return on investment. That's what makes IT security policy development all the more challenging. Namely, proving that the need exists and that a properly formed strategy can mitigate risks, protect critical information assets and ensure confidentiality, integrity and availability. One way to increase appropriations, though, is to fail a security audit and place the blame on inadequate funding, which is essentially what's happening. The fiscal 2005 federal budget increases IT spending by about 10% over fiscal 2004, to close to $60 billion. A company, especially a public one, has to maintain solid earnings while building equity. Meanwhile, it takes leadership and vision to recognize the value of a solid IT enterprise security policy. Much has been written on demonstrating ROI for IT security, so I won't get too granular here. Suffice it to say that it doesn't take too much effort to perform a solid risk assessment and produce a risk-level matrix that clearly demonstrates the risk thresholds of any enterprise. The tough part is gaining support and momentum for developing a solid set of plans (contingency, continuity of operations, training and education) and to ensure that these plans get the critical executive-level support within the organization. Lessons learned - the "P word" That would be policy, and that's just what NIST provides. But that's not enough. The federal government - and this applies to industry as well - needs guidance but also needs procedures to follow. The average IT professional needs a set of standards to subscribe to and a set of guidelines on how to meet those standards. That's the missing piece, which I hope will be recognized and developed sometime soon. It was another year of dismal federal IT security grades, and the complexities and threats in the world aren't diminishing. Government agencies and related organizations have their work cut out for them, but the pieces are there. By optimizing talent, focusing on embedding security into the systems development cycle and including further refinement of continuity planning, along with the continual retirement of legacy systems, the overall grades should show improvement next year. [1] http://public.ansi.org/ansionline/Documents/Standards%20Activities/Homeland%20Security%20Standards%20Panel/ComputerSecurity.pdf _________________________________________ ISN mailing list Sponsored by: OSVDB.org
This archive was generated by hypermail 2b30 : Mon Apr 12 2004 - 04:52:40 PDT