mitigating multiple
play

Mitigating Multiple professionals can mitigate even previously - PowerPoint PPT Presentation

By recognizing the four main categories of attack, security Security professionals need to understand how to plug the Mitigating Multiple professionals can mitigate even previously unknown vectors: security gap from Layers 3 to 7, and protect


  1. By recognizing the four main categories of attack, security Security professionals need to understand how to plug the Mitigating Multiple professionals can mitigate even previously unknown vectors: security gap from Layers 3 to 7, and protect against multi-layer 1. Volumetric: Flooding attacks, with a full proxy security architecture. It's time to rethink 2. Computational Asymmetric: Consuming CPU cycles and refine the enterprise security architecture, so organizations 3. Stateful Asymmetric: Abusing memory can remain agile and resilient against future threats. DDoS A tu ack Vectors 4. Vulnerability-based: Exploiting software vulnerabilities The following mindmap shows the detection methods (left) for 5. Blended DDoS: Combination of multiple attack vectors DDoS attack categories (middle) and the mitigations (right). DETECTION MITIGATION Use Web Application Firewall heuristic Signature Based Rate Limiting (L3-L7) latency based detection Strengths: Strengths: - Fast, easy for hardware - Ease of hardware implementation Method: Resource Attacked: implementation - Fast deterministic Packet or Flow Flood Network Bandwidth Set proper thresholds for load - Deterministic/ predictable - False positive rate Considerations: Considerations: Volumetric - Reactive - Dependent on 5-tuple/header info Use Web Application Firewall - Some may not be able to distinguish to distinguish “Good” vs. “Bad” heuristic Transaction volumetric “Good” vs. “Bad” Per Second (TPS) based detection • UDP Packet Floods • ARP/ICMP Floods • DNS Reflection Attack • HTTP flood Client Challenge (L7-L8) Heuristic Flow Analysis Strengths: Strengths: - Use client response to lower - Good at “Good” vs.“Bad” false-pos/neg. rate - Pro-actively finds anomalies Method: Method: - Weed out botnets to protect server Vulnerability/Exploit Create packets/requests Stateful Asymmetric Create malformed/ resources Considerations: requiring security and crafted requests & - Computational challenge can limit - May require “baseline-ing” server infrastructure • SYN Floods packets targeting • LAND Attack per-attacker rate under attack to software security holes • Fragmentation Attacks • Bad TCP Options/Size maintain state • Slow-Loris/Post, Slow Post/GET • Invalid DNS Opcode Considerations: • FTP Ephemeral Opens, Security Appliance Resource Monitoring • Apache killer, PostOfDoom Resource Attacked: Resource Attacked: - May not work with all listener • Slow file download • Apache Struts Software stack of security Memory of security and types (Forwarding, BigTCP) Strengths: and server infrastructure server infrastructure - Based on attack’s target (not specific to attack mechanism) - Low false positive/negative rate Use Web Application Reputation List (L3-L7) Use profile definitions and Firewall flow definition for - Feedback-driven security Computational Asymmetric resource monitoring application logic DOS appliance self-defense mechanism Strengths: • SSL Renegotiation - Detect in Layer 7 and block in Considerations: Layer 3 • Heavy URL’s - Protects only resources that are - Real-time updates • XML DND, XML external monitored entity logic (e.g.: Ask where Set proper protocol - Not server-aware; doesn’t directly protection Considerations: are the closest ATMs?) protect server - Does work against many volumetric network attacks (spoofed source addresses) Method: Resource Attacked: Server Resource Monitoring Create requests that Compute Resources Set proper have large of security and server timeouts Strengths: computational cost infrastructure - Based on attack’s target (not on security and (Middleware/DB) of Use custom scripts for zero day attack and specific to attack mechanism) server infrastructure Server Full Proxy Architecture (L3-L8) other vulnerability exploits protection - Low false positive/negative rate under attack - Server-centric Strengths: - Feedback-driven - Manipulate packages - Programmability Considerations: - Flexibility Application - Protects only resources that are Web Delivery Network Anti-Fraud monitored Application Controller (ADC) Firewall Protection Firewall Session DDoS Protection Reference Architecture Protection Get the DDoS Protection Exclusive 8 Resources! User/End point Next-Generqation Firewall Corporate Users OSI BUILDING http://delivr.com/2wgtk • The first tier at the • The second tier is for 7 Application perimeter is layer 3 application-aware, Tier 1 Tier 2 and 4 network CPU-intensive defense Network attacks: Financial firewall services SSL attacks: mechanisms ICMP flood, SSL renegotiation, Services 6 Multiple ISP UDP flood, SSL flood Strategy SYN flood • Simple load • SSL termination balancing Session Legitimate 5 to a second tier Users • Web application firewall E-Commerce ISPa/b • IP reputation • Mitigate asymmetric and Network Application database and DNS SSL-based DDoS attacks DNS attacks: 4 HTTP attacks: DDoS DNS amplification, Slowloris, query flood, Attacker slow POST, DDoS Protection • Mitigates dictionary attack, recursive POST/GET Subscriber DNS poisoning Network volumetric and Cloud Scrubbing 3 Service DNS DDoS attacks IPS Threat Feed Intelligence 2 Anonymous Anonymous Strategic Point of Control 1 Scanner Botnet Attackers Proxies Requests Sources : F5 Security Forums h"p://resources.idgenterprise.com/original/AST-0127081_Mi@ga@ng_Mul@ple_DDoSA"ack_Vectors_Infographic.PDF

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend