The default installation of NetScreen-Security Manager before Feature Pack 1 does not enable encryption for communication with devices running ScreenOS 5.0, which allows remote attackers to obtain sensitive information via sniffing.
Netscreen running ScreenOS 4.0.0r6 and earlier allows remote attackers to cause a denial of service via a malformed SSH packet to the Secure Command Shell (SCS) management interface, as demonstrated via certain CRC32 exploits, a different vulnerability than CVE-2001-0144.
Firewalls from multiple vendors empty state tables more slowly than they are filled, which allows remote attackers to flood state tables with packet flooding attacks such as (1) TCP SYN flood, (2) UDP flood, or (3) Crikey CRC Flood, which causes the firewall to refuse any new connections.
Buffer overflow in NetScreen-Remote 8.0 allows remote attackers to cause a denial of service and possibly execute arbitrary code via crafted Internet Key Exchange (IKE) response packets, possibly including (1) a large Security Parameter Index (SPI) field, (2) large number of payloads, or (3) a long payload.
The web interface (WebUI) of NetScreen ScreenOS before 2.6.1r8, and certain 2.8.x and 3.0.x versions before 3.0.3r1, allows remote attackers to cause a denial of service (crash) via a long user name.
NetScreen ScreenOS before 2.6.1 does not support a maximum number of concurrent sessions for a system, which allows an attacker on the trusted network to cause a denial of service (resource exhaustion) via a port scan to an external network, which consumes all available connections.
NetScreen ScreenOS prior to 2.5r6 on the NetScreen-10 and Netscreen-100 can allow a local attacker to bypass the DMZ 'denial' policy via specific traffic patterns.