Category: systems

Traffic limits with HAProxy stick-table

A traffic rate limiting feature is required to keep an HTTP website backend safe from abusive or malfunctioning clients.  This requires the ability to track user sessions of a particular type and/or from a given IP address. HAProxy is an HTTP proxy which (when configured as reverse proxy to protect a website), receives client requests in its frontend and sends those requests to servers in its backend.   The config file has corresponding frontend and backend sections. Haproxy also has an in-memory table to store state related to incoming HTTP connections, indexed by a key such as client IP address.  This table is called a stick-table – it is enabled using the ‘stick-table’ directive in the haproxy config file.

The stick-table directive allows  specifying the key, the size of the table, the duration an entry (key) is kept in seconds and various counts such as currently active connections, connection rate, http request rate, http error rate etc.

Stick tables are very useful for rate-limiting traffic and tagging traffic that meets certain criteria such as a high connection or error rate with a header which can be used by the backend to log the traffic.

The origin of this rate-limiting feature request along with an example is at . Serverfault is a high traffic website so it is a good indication if the feature works for them.

frontend http
    bind *:2550

stick-table type ip size 200k expire 10m store gpc0

# check the source before tracking counters, that will allow it to
# expire the entry even if there is still activity.
acl whitelist src
acl source_is_abuser src_get_gpc0(http) gt 0
use_backend ease-up-y0 if source_is_abuser
tcp-request connection track-sc1 src if ! source_is_abuser

acl is_test1 hdr_sub(host) -i
acl is_test2 hdr_sub(host) -i

use_backend test1  if is_test1
use_backend test2  if is_test2

backend test1 
stick-table type ip size 200k expire 30s store conn_rate(100s),bytes_out_rate(60s) 
acl whitelist src

# values below are specific to the backend
tcp-request content  track-sc2 src
acl conn_rate_abuse  sc2_conn_rate gt 3
acl data_rate_abuse  sc2_bytes_out_rate  gt 20000000

# abuse is marked in the frontend so that it's shared between all sites
acl mark_as_abuser   sc1_inc_gpc0 gt 0
tcp-request content  reject if conn_rate_abuse !whitelist mark_as_abuser
tcp-request content  reject if data_rate_abuse mark_as_abuser

server local_apache localhost:80

Note that the frontend and backend sections have their own stick-table sections.

A general strategy would be to allow enough buffer for legitimate traffic to pass in, drop abnormally high traffic and flag intermediate risk traffic to the backend so it can either drop it or log the request for appropriate action, including potentially adding the IP to an abusers list for correlation, reverse lookup and other analysis. These objectives are achievable with stick-tables.

An overview of the HAProxy config file with the sections global, defaults, frontend, backend is here.

Stick tables use elastic binary trees-

Related, for analysis of packet captures in DDoS context, a useful tool is python dpkt – .



Javascript Timing and Meltdown

In response to meltdown/spectre side-channel vulnerabilities, which are based on fine grained observation of the CPU to infer cache state of an adjacent process or VM, a mitigration response by browsers was the reduction of the time resolution of various time apis, especially in javascript.

The authors responded with alternative sources of finding fine grained timing, available to browsers. An interpolation method allows obtaining of a fine resolution of 15 μs, from a timer that is rounded down to multiples of 100 ms.

The javascript  high resolution time api is still widely available and described at with a reference to previous work on cache attacks in Practical cache attacks in JS

A meltdown PoC is at, to test the timing attack in its own process. The instruction RDTSC returns the Time Stamp Counter (TSC), a 64-bit register that counts the number of cycles since reset, and so has a resolution of 0.5ns on a 2GHz CPU.

int main() {
 unsigned long i;
 i = __rdtsc();
 printf("%lld\n", i);

Ethereum Security and the DAO Solidity Attack

The basics of Ethereum are described in the Gavin Wood paper. A list of keywords in Solidity are described in this file from its source, which includes “address”, “contract”, “event”, “mapping” and “wei” ( 1 Eth= 10^18 Wei). This list does not include “gas”, which is a mechanism described in Wood’s paper to combat abuse. Interestingly the paper says “The first example of utilising the proof-of-work as a strong economic signal to secure a currency was by Vishnumurthy et al [2003]”, aka the Karma paper.

The karma paper talks about solving a cryptographic puzzle as enabling one to join the network and be assigned a bank set: “the node certifies that it completed this computation by encrypting challenges provided by its bank-set nodes with its private key. Thus each node is assigned an id beyond its immediate control, and acquires a public-private key pair that can be used in later stages of the protocol without having to rely on a public-key infrastructure”. Crypto puzzles for diverse problems have been proposed before, a survey and comparison is at

The DAO attack had 3 components, a hacker, a malicious contract and a vulnerable contract. The malicious contract is used to withdraw funds from the vulnerable contract so that it does not get a chance to decrement its balance. Oddly enough the gas mechanism which is supposed to limit computation did not kick in to stop this repeated remittance.

A few weeks before the DAO attack someone had pointed out to me that security of solidity was a bit of an open problem. My feeling was contracts should be layered above the value exchange mechanism, not built into it. Bitcoin based protocols with the simpler OP_RETURN semantics appeared more solid. Later around October’16 at an Ethereum meetup, Fred Ehrsam made the comment that most new projects would be using Ethereum instead of bitcoin. But Bitcoin meetups had more real-world use cases being discussed. The technical limitations exist, which are being addressed by forks such as SegWit2x this November. Today saw a number of interesting proposals with Ethereum, including Dharma, DataWaller and BloomIDs. Security would be a continuing  concern with the expanding scope of such projects.

Linux Bridges and Firewalls for Containers

Hubs and Repeaters work on bits on Layer1. Bridges and Switches work on frames, the protocol data unit for Layer2. A Hub supports a star configuration – like a splitter it simply forwards traffic received on one segment to all the others. A Repeater amplifies signal for longer distances. A Bridge is a smarter type of Hub, usually implemented in software, which looks at the destination MAC address of the packet before forwarding to the segment with MAC matching that destination, thereby avoiding unnecessary network traffic on all the other segments. It does not look at IP addresses. A Switch is a reincarnation of a Bridge in hardware with several ports.  The basic operation of a Bridge/Switch is L2 forwarding, which includes ARL (Address Resolution Logic) table learning and ageing.  An L3 Router switches packets based on the IP address instead of the frame address.

A linux bridge from a VM to the Host, does not need an IP address, but it is a good idea if you want to talk to the host for management or set firewall rules on it.

To create a bridge br0 between two interfaces eth0 and eth1, one runs commands: brctl addbr br0, brctl stp br0 on, brctl addif br0 eth0, brctl addif br0 eth1.

Since a bridge is an L2 construct, how can one filter IP addresses with it. A quote: An Ethernet interface cannot usefully have an IP address if it is also attached to a bridge. However it is possible to achieve the same effect by binding an address to the bridge itself:

ifconfig br0

An example of firewall configuration on a bridge.

Troubleshooting bridges:

brctl showmacs br0

Docker creates a few bridges by default. A primer on Docker Container networking and the bridges it creates are here and  here. Here are three common ones.

bridge: docker0 bridge . This is the default, allowing all containers to talk to each other

host: container listens to host, without any isolation.

Some problems with host iptables rules reveal themselves in working with docker.

A look at firewalld, ebtables, ipset, nftables at . The command line tools talk to the kernel over a netlink socket.

A more extensible approach is with Berkeley Packet Filter (BPF), the classical one now called cBPF, the extended successor called eBPF.   A very basic problem is filtering HTTP or DNS packets based on the destination hostname, rather than the IP address.  This is discussed in an example by Cloudflare here.

NVidia Tiny Linux Kernel and TrustZone

The NVidia Tiny Linux Kernel (TLK), is 23K lines of BSD licensed code, which supports multi-threading, IPC and thread scheduling and implements TrustZone features of a Trusted Execution Environment (TEE). It is based on the Little Kernel for embedded devices.

The TEE is an isolated environment that runs in parallel with an operating system, providing security. It is more secure than an OS and offers a higher level of functionality than a SE, using a hybrid approach that utilizes both hardware and software to protect data. Trusted applications running in a TEE have access to the full power of a device’s main processor and memory, while hardware isolation protects these from user installed apps running in a main operating system. Software and cryptographic isolation inside the TEE protect the trusted applications contained within from each other. A paper describing the separation with alternatives for virtualizing the TEE appeared at

TrustZone was developed by Trusted Foundations Software which was acquired by Gemalto. Giesecke & Devrient developed a rival implementation named Mobicore. In April 2012 ARM, Gemalto and Giesecke & Devrient combined their TrustZone portfolios into a joint venture Trustonic, which was the first to qualify a GlobalPlatform-compliant TEE product in 2013.

A comparison with other hardware based security technologies is found here. Whereas a TPM is exclusively for security functions and does not have access to the CPU,  the TEE does have such access.

Attacks against TrustZone on Android are described in this blackhat talk. With a TEE exploit,  “avc_has_perm” can be modified to bypass SELinux for Android. By the way, Access Vectors in SELinux are described in this wonderful link. “avc_has_perm” is a function to check the AccessVectors allows permission.

Neural Network learning

Notes from Autonomous Driving Udacity course.




Embedded Neural Nets

A key problem for embedded neural networks is reduction of size and power consumption.

The hardware on which the neural net runs on can be a dedicated chip, an FPGA, a GPU or a CPU. Each of these consumes about 10x the power of the previous choice. But in terms of upfront cost, the dedicated chip costs the highest, the CPU the lowest. An NVidia whitepaper compares GPU with CPU on speed and power consumption. (It discusses key  neural networks like AlexNet. The AlexNet was a breakthrough in 2012 showing a neural network to be superior to other image recognition approaches by a wide margin).

Reducing the size of the neural network also reduces its power consumption. For NN size reduction, pruning of the weak connections in the net was proposed in “Learning both Weights and Connections for Efficient Neural Networks” by Song Han and team at NVidia and Stanford. This achieved a roughly 10x reduction in network size without loss of accuracy. Further work in “Deep Compression” achieved a 35x reduction.

Today I attended a talk on SqueezeNet by Forrest Iandola. His team at Berkeley modified (squeezed) the original architecture, then applied the Deep Compression technique above to achieve a 461x size reduction over the original, to 0.5Mb. This makes it feasible for mobile applications. This paper also references the V.Badrinarayan’s work on SegNet – a different NN architecture, discussed in a talk earlier this year.

The Nervana acquisition by Intel earlier this year was for a low power GPU like SOC chip with very high memory bandwidth.