Page 20 - FCW, September/October 2019

Cybersecurity and Modernization
Demystifying data
analytics and protection
Machine learning and cloud technology can help agencies

Rob Roy
Public Sector CTO, Micro Focus Government Solutions
phishing campaign and take over a user’s credentials. That could allow them to
log in as a legitimate user and download the contents of the database. All that data is unencrypted, which means all
the personally identifiable information is exposed. Therefore, it’s important for the government to move to the next stage of protecting data in use.
Moving legacy systems
to the cloud
When it comes to modernizing the government’s approach to cybersecurity, legacy systems present some special
MANY CYBERSECURITY BREACHES happen when an adversary takes over a legitimate
identity. Therefore, a good network defense strategy relies on machine learning to monitor user behavior and make sense of data from billions of cybersecurity events, which is impossible for a human to do.
For example, anomalous user behavior is marked by a series of activities outside their normal behavior, such as a U.S. employee logging in at 3 a.m. from an IP address in China, accessing a database he’s not supposed to have access to and downloading files to a local computer.
A network administrator seeing those activities can conclude that a breach is in progress and shut that person off.
With unsupervised machine learning, agencies can start putting the onus on computers to recognize unusual behavior. Unlike the rules-based approach, unsupervised machine learning lets the technology develop an understanding of how the network’s users typically behave and alert administrators when something abnormal occurs, increasing the likelihood that a rogue event is detected and a response is orchestrated at machine speed.
Letting computers do what
they do best
Machine learning is rooted in analytics. There’s no lack of data to analyze in the government, and in fact, most agencies have more data than they know what to do with. Understanding what to look for in all that raw information requires the help of data scientists, who can tell agencies how to solve a problem or answer a question by
identifying the relevant data to analyze. Then the technology can step in and do
the repetitious and very rapid work that computers do well. The resulting analysis provides the insights that leaders need to make good decisions.
Throughout the process, agencies need to safeguard the data and comply with rules governing privacy. The government has a mandate to encrypt data when it’s at rest, but data is vulnerable if it’s not also encrypted while it’s being used.
For example, if the information in a large agency database is unencrypted for use, hackers could mount a successful
davooda/Shutterstock/FCW Staff
S-20 SPONSORED CONTENT