The previous article gave great coverage of API Discovery functionality on the Volterra platform. This one gives an overview of what is available in terms of API control features.
Volterra platform provides a rich set of techniques to securely deliver applications and APIs. It goes beyond industry-standard features like WAF and DoS and introduces cutting-edge AI/ML technologies that significantly increase accuracy and decrease the number of false positives. The following paragraphs give an overview of available technologies and provide links to get started.
Per API Request Analysis
Per requests analysis name is self-explaining. The technique analyzes every single API call and extracts various metrics out of it such as request/response size, latency, etc. Based on observed data it builds a model that identifies a valid request structure for a particular API endpoint. Once the model is applied to the traffic it automatically catches any request that deviates from learned models and notifies a user. This approach is good to discover any per request-based threats.
TS anomaly detection is somewhat similar to the previous mechanism but instead of looking into each API request, it monitors the traffic stream as a whole. Therefore, metrics of interest relate to the entire stream such as requests rate, errors, latency, and throughput. Then statistical algorithms detect the following types of anomalies:
Unexpected traffic patterns
Such detection instantly notifies an administrator about any event that potentially means a service disruption. Whether it is a traffic drop caused by a misconfiguration or a traffic spike due to a DoS attack.
Behavior analysis uses AI and ML to profile valid user behavior. It detects clear suspicious behavior based on WAF events, login failures, L7 policy violations, and other anomalous activity based on user behavioral aspects. Once enabled an administrator can observe a whole list of suspicious users in the HTTP load balancer app firewall dashboard.
As a platform with a mature security stack, Volterra's WAF matches and exceeds all industry-standard features. In addition to that, it significantly simplifies the WAF configuration. It only asks for application language, CMS, and web server type to set up the best level of protection for an application.
Another name of the service policy is the L7 policy. When applied to a load balancer it allows to granularly define what traffic is allowed to a service and what is not in terms of L7 metrics such as hostname, method, path, or any other property available from a request. Service policy consists of rules that match a request by one or more metrics and then apply one of the following actions allow, deny or check next.
Rate limiting protects an application or an API from excessive usage via limiting the rate of requests from a single user or client executed in a period of time. There is a variety of ways to identify a user:
By IP address
By HTTP Header
By query string parameter
The feature can limit the rate or total amount of requests which an entity is allowed to execute.
Similar to rate limiting DoS protection helps to keep service up and running under an excessive amount of traffic. However, opposite to rate limiting Volterra's DoS protection operates on L3/4 and deals with much higher volumes of traffic.
Self-explaining feature as well. Allows to set up lists of trusted and untrusted IP prefixes. Clients from trusted prefixes bypass all security checks. Clients from untrusted prefixes are simply blocked before any security check happens. Both lists support the expiration time setting. Meaning that rule continues to exist but isn't applied to the traffic after expiration time comes. That is handy to set up temporary rules e.g. for testing or security purposes.
All security features above are available on top of Volterra's HTTP load balancer. A combination of those provides reliable protection for an API and automatically supports it in up and running conditions.