Security is not black and white, but has many different shades of grey. Many attackers, and their associated IP addresses, only target specific industry segments or remain active for only a short period of time. Those IP addresses are also used by legitimate users. Simply blocking an IP address could affect the legitimate users, which causes a negative business impact. 

The changing threat landscape also forces enterprises to constantly improve their ability to act on suspicious client behaviors in a way that further reduces the risk of successful DDoS or application layer attacks and at the same time minimizes the impact to legitimate users. Client Reputation increases the accuracy of security decisions that separate malicious traffic from legitimate traffic. Kona Site Defender primarily focuses on threat vectors, while Client Reputation provides a complementary view on clients — the potential attack sources. 

Many IP reputation solutions that are on the market today create only a single score per client or IP address, which is the same for all customers. Client Reputation, however, uses a state-of-the-art, proprietary risk analysis engine that computes a risk score for every source IP address, customized for every customer. This custom risk-based scoring model is significantly more accurate than generic scoring, and it has shown that actions taken based on the risk score are less likely to negatively impact legitimate clients and users. The quality of risk scoring is driven by the knowledge that can be extracted from Big Data. Akamai leads the Content Delivery Network market as a central hub in the Internet ecosystem, serving 15-30% of all web traffic at any given moment, interacting with 1.3 billion client devices every day. The data is analyzed by Cloud Security Intelligence, Akamai’s Big Data security platform. The breadth and scope of this platform enables Akamai to deliver a client reputation service well beyond anything available in the market today.

 

- Cross-customer correlation: Correlation of client requests across different customers and identification of malicious intent.

- Multiple risk score categories: Ability to associate potentially malicious activity with the following types of attackers: