Protected profile
What a protected profile is, how it protects your resource, and why there may be several profiles.
What it is
A protected profile is a set of traffic filtering rules at the network layer (L3/L4) applied to your subnets to protect against DDoS attacks.
In simple terms, it is mitigation logic: which packets are legitimate, which are malicious, and what to do with them.
How it works
Technically, a profile is a configuration file. Its copies are stored on the nodes of the mitigation system. When your traffic passes through the system, it reads the file to determine how to filter the packets.
There may be one or several profiles. For example, if your DNS and corporate VPN require different filtering rules, we create a separate profile for each one.
What it consists of
Each profile contains two elements:
-
Subnets or sets of IP addresses. A list of addresses that the mitigation system protects using this profile’s rules.
-
A rule chain. A set of traffic filtering rules performed by the mitigation system. Each rule includes a condition—traffic parameters such as protocol, port, packet size, and others—and an action: what to do with a packet that matches the condition, such as allow it to your servers, drop it, or apply TCP or DNS authentication to the connection.
When it is used
A profile is used during a DDoS attack.
Under normal conditions, traffic reaches your servers directly without passing through the mitigation system. As soon as we detect an attack, the route changes: traffic enters the system, is filtered using the profile’s rules, and only then reaches your resource.
|
When there is no DDoS attack, no mitigation mechanisms work, including Black lists and blackhole. |
How it is created
We create and configure the protected profile. You only need to complete the questionnaire sent by our engineers.
The minimum information needed to connect protection is a list of your subnets and individual IP addresses. If you do not provide any other information in the questionnaire, we create a profile that includes:
-
The specified subnets and IP addresses.
-
A standard set of rules.
The standard set of rules is a filtering policy developed through years of experience protecting customers from DDoS attacks. It includes rules against common attack vectors that are regularly encountered in practice. This set can protect a resource even without additional configuration. We update it regularly as attackers develop new techniques.
If you complete the questionnaire in more detail, we configure protection specifically for your service. Additional information helps us take account of your architecture and traffic characteristics. The more details you provide, the more accurate the protection and the lower the risk of false blocks.
What the standard set of rules includes
Our standard filtering rules include:
-
Dropping inbound fragmented traffic.
-
Dropping all protocols except TCP, UDP, ESP, AH, ICMP, and GRE.
-
Blocking packets whose source and destination ports are both in the 0–1023 range. Exceptions: 123 → 123 (NTP) and 500 → 500 (IPsec).
-
Dropping inbound ICMP traffic. Exceptions: ICMP Echo Request and ICMP Echo Reply.
-
Dropping inbound ICMP traffic with a packet length greater than 200 bytes.
-
Dropping inbound traffic identified as NTP Amplification.
-
Dropping inbound traffic from potentially vulnerable UDP ports, including 17, 19, and 80.
-
Dropping inbound traffic sent to potentially vulnerable UDP ports, including 0, 22, and 80.
-
Dropping inbound UDP traffic addressed to the network or broadcast address (for example,
192.0.2.0or192.0.2.255in the192.0.2.0/24subnet). -
Dropping inbound DNS traffic originating from port 53 with a packet length greater than 1,000 bytes.
-
Dropping inbound DNS traffic sent to port 53 with a packet length greater than 1,000 bytes.
-
Dropping inbound traffic sent from system ports to ports 4444, 8080, 8443, and 8088.
-
Dropping TCP packets with SYN, SYN-ACK, or RST flags and a packet length greater than 100 bytes.
-
TCP SYN Authentication, which counters TCP SYN floods by authenticating the client.
-
DNS Authentication, which counters DNS floods by authenticating the client.
-
Blocking a source IP address for 300 seconds when it exceeds 2,000 UDP packets per second.
-
Blocking a source IP address for 300 seconds when it exceeds 50 TCP SYN packets per second.
We deliberately provide a general description and do not disclose every rule in detail, so attackers cannot use this information to prepare attacks against your services.
How rules work
Filtering rules are the basis of a protected profile. They define the conditions under which traffic is considered legitimate or suspicious and specify what to do with it.
Example of a real rule: if an inbound IP packet is fragmented, drop it.
Why this rule exists: fragmented packets are often used in DDoS attacks to confuse traffic analysis and bypass protection. This traffic is rare in normal service operation, so it can be blocked safely.
A rule consists of a “condition + action” pair. The condition describes traffic parameters: protocol; source IP address and ASN; ports; packet size; packet fragmentation; TCP flags; payload content; whether the source belongs to White lists or Black lists; and others. The action defines what the mitigation system does with traffic that matches the condition. One of three actions is possible:
-
Allow the packet to pass.
-
Drop the packet.
-
Apply TCP or DNS authentication to the connection.
Rules can be strict or threshold-based. Strict rules trigger immediately when the condition is met. For example, if a packet is fragmented, it is dropped immediately. Threshold-based rules respond only when limits are exceeded. For example, if one IP address sends suspiciously many packets of the same type, its traffic is blocked temporarily.
The rules in a profile are combined into a chain — an ordered list. The mitigation system sends each packet through this list: it checks whether the packet matches the first rule, then the second, then the third, and so on until it finds a match. Once it does, it applies the action specified in the rule.
While processing traffic, the mitigation system can assign a label to a packet. A label identifies a particular type of traffic. For example, IP_FRAG means fragmented traffic. In protected profile traffic statistics, you can view charts that show how much traffic with different labels reached the mitigation system.
How to change a profile
We make improvements proactively. When a new widespread attack technique appears, we develop rules to counter it and add them to all customer profiles. This is how protection continuously evolves and adapts to evolving attacker techniques.
We also monitor how protection works for each customer. If we see that some malicious traffic was not filtered during an attack, we adjust the profile and strengthen filtering.
If you have requests for the filtering policy, contact support.
We will update the profile at your request by strengthening or relaxing individual rules to account for the characteristics of your service.