Page 63 - Cyber Warnings

• Distributed systems and big data can now offer ad-hoc analysis, allowing queries
constructed using multiple data field dimensions, against billion row data sets, to deliver
answers in questions

8. Deploy adaptive baselining for more accurate anomaly & DDoS detection
• Legacy monitoring system limitations rely on manual policy changes that usually fall
behind, or overly broad that suffer from the effect of averages

• Distributed systems and big data make it possible to individually track millions of network
elements and adaptively select the most “interesting” set of elements to baseline and
measure for anomalies, which can deliver 30% better detection of anomalous network
conditions, including DDoS attacks

9. API-friendly big data can make metadata useful for automated remediation
• The IT industry is very intrigued with the possibilities of using automation and
orchestration for network remediation

• Big data’s ability to detect (and thus describe) very precise network conditions in a multi-
dimensional fashion provides the semantic bridge to value from network orchestration,
and having strong API’s makes auto-remediation possible

10. Use ad-hoc exploration of metadata to proactively understand emerging conditions
• Exploratory analyses enabled by ad-hoc querying allows for experts to look out over the
horizon for emerging conditions that will impact how you plan and monitor your network
and infrastructure

• This proactive approach allows for agile changes to detection policies so you don’t miss
important phenomena

63 Cyber Warnings E-Magazine December 2016 Edition
Copyright © Cyber Defense Magazine, All rights reserved worldwide