Rate limiting

From Halon, SMTP software for hosting providers
Jump to: navigation, search

Rate limiting is a crucial defense method against service misuse. The Halon SMTP software implements this in a global rate controlling service called rated. It handles rate limiting for all processes (HSL contexts), based on an "items per time interval" algorithm. Our implementation features a sliding rate limit (in contrast to leaking bucket; which has a fixed output rate regardless of exactly when the rate was exceeded).

Implementation

Rate limits may be "hit" using the HSL rate() function, and sometimes using the graphical blocks in the flows. It returns true as long as the rate() call interval is less than its rate limit, but false when exceeded. If the count parameter is zero, the current rate is returned instead. When a "hit" is counted, its expiration time is the fourth parameter (interval).

// get the rate limit number, without increasing
if (rate("failed-login:saslusername"$saslusername060) >= 3)
    
Reject("Too many failed logins");

// increase, and reject if exceeded
if ($is_spam)
    if (
rate("spammers"$senderip10086400) == false)
        
Reject("You may only send 100 suspicious messages per day"); 

Rate limits may be read and cleared using the SOAP API.

Clustering

Rate limits are synchronized in the cluster if a HMAC-SHA1 key is specified on the system settings page. The protocol is based on UDP and uses port 13131. In terms of security it may be wise to only activate this feature on local networks, as the HMAC-SHA1 only serves as a packet-authenticity firewall. It does not provide encryption nor protect against replay attacks. It should probably be protected by other means such as a VPN tunnel or psychical security (DMZ).

Performance

The rate limits performance is roughly O(log N + log M) and lookups are currently implemented as a C++ std::multimap.

Rate limit page in Web UI

The rate limit page in the Web UI may display different fields depending on how it is configured. Entries may be alphabetically paged and filtered by name and by a minimal count (threshold).

Result Description
Namespace The namespace or "bucket".
Entry The name of the rate limit entry.
See table below.
Rate count divided by the interval.
Interval The time between first and last count.
Most recent When a count was last added to this entry.

Cluster

If viewed in a cluster (/cluster/ratelimit/). The rate limit may or may not be synchronized.

Synchronized

Result Description
Count The maximum count a single unit has; this should preferable be the same on all units.
Min  If shown, it is the lowest count a single node has.

Non-synchronized

Result Description
Same as for synchronized
Sum The total sum of all counts on all units.

Single unit

If viewed on a single unit (/ratelimit/).

Result Description
Count The count this unit has.
Local count If shown, the count created on this unit (otherwise the same as Count).

Send rates from Python

import socket, struct, hashlib, hmac

key = 'hash-key'

m = struct.pack('!bbIII64s512s', 1, 4, 0, 0, 3600, 'namespace', 'entry');
m += hmac.new(key.decode('hex'), m, hashlib.sha1).digest()

s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
s.settimeout(1)
s.sendto(m, ('192.168.0.12', 13131))