Performance: Parallel subnet scanning across /22, /23, /24 networks

Description

Implement parallelized subnet scanning for DC discovery when targeting large network ranges (e.g., 10.0.0.0/22).

Current Behavior

• Sequential scanning of all hosts in /24 subnet
• Single-threaded port fingerprinting (88, 389, 636, etc.)
• Slow discovery on large networks (1000+ hosts)

Proposed Implementation

1. Adaptive subnet sizing

   • /24 (256 hosts) → scan all
   • /23 (512 hosts) → split into /24s, parallelize
   • /22 (1024 hosts) → split into /24s, parallelize
   • Larger ranges → use config-specified subnet size

2. Concurrent scanning

   • ThreadPoolExecutor with 32 workers
   • Parallel port scans across subnets
   • Batch results by subnet
   • Report live hosts as discovered (streaming)

3. Optimization

   • ICMP ping before port scan (reduce attempts)
   • Early termination if DC found
   • DNS reverse lookup parallelization
   • Cache ARP table for local networks

Requirements

• Minimum 10x speedup on /22 scans
• Memory usage < 200MB even with 256+ concurrent scans
• Support both local networks (ARP) and remote networks (ICMP/TCP)
• Respect per-host timeout configuration

Files to Update

• adpentest/core.py - subnet_sweep() function
• Add: parallel_subnet_scan() function

Success Criteria

• /24 scan: < 10 seconds
• /23 scan: < 15 seconds
• /22 scan: < 30 seconds

Start by reading subnet_sweep() in adpentest/core.py to understand the current sequential scan and its timeout behavior. Implement the requested parallel_subnet_scan() flow around the stated subnet sizes, worker limit, streaming results, and local versus remote network requirements. Done means the /24, /23, and /22 timing targets are met while staying under 200MB and respecting per-host timeouts.