Routing attacks continue to play an effective role on the Internet. These attacks send malicious packets to routers that exchange routing information, causing them to exploit vulnerabilities. The protection against these attacks is challenging. Not only are countermeasures hard to deploy but they also fail to provide protection guarantees.
Relay networks are no exception to routing attacks, which decrease their performance. Particularly vulnerable to these attacks are Blockchain systems. Even the most widely-used cryptocurrency Bitcoin is at risk. It appears as though the infamous cryptocurrency can be split in half by “any AS-level adversary using BGP hijacking”. Hence, a couple of networks have emerged, designed to disseminate Bitcoin blocks. These include Falcon network, FIBRE (Fast Internet Bitcoin Relay Engine), and this article’s focus; SABRE.
“A relay network is a broad class of network topology commonly used in wireless networks, where the source and destination are interconnected by means of some nodes.” In other words, a relay network is used to send information between two devices – known as nodes. Because in wireless networks the distance between the source and destination is greater than the transmission range allows, disallowing them to communicate with each other directly, intermediate nodes are needed.
Therefore, the network relays the information from the source to different nodes. These pass on the information to the destination. How else would the Internet work?
A Secure Relay Network For Bitcoin
SABRE is a secure and scalable relay network running alongside the Bitcoin network. Through a set of connections that are resilient to routing attacks, SABRE protects Bitcoin clients worldwide. Crypto Heroes caught up with Maria Apostolaki, a Ph.D student and Microsoft researcher who wrote SABRE’s whitepaper with three other colleagues from the Swiss University ETH in Zurich.
What threats does SABRE aim to eliminate?
“SABRE is mostly designed to deal with routing attacks in Bitcoin but it’s generalized to any cryptocurrency or any blockchain application that uses the Internet. The exact attack that it protects against is the partitioning attack in which the attacker splits the network into two disjoint components by hijacking traffic.”
When asked how people can use SABRE, Apostolaki says that they need to have an extension installed in their Bitcoin nodes and a UDP connection to the SABRE node.
How is SABRE going to be implemented from the technical perspective?
“Two kinds of deployment are possible, namely a private SABRE and a public one. A private SABRE can be implemented by a single mining pool or a group of collaborated mining pools. This means, only a few predefined nodes (the gateways of those pools) will need to use SABRE. This type of deployment is easier as SABRE nodes need not be protected against DDoS attacks. As a result, SABRE nodes would be regular Bitcoin clients.
Apostolaki adds; “For the public deployment, we proposed a software/hardware co-design. This would allow a SABRE network of as few as six nodes to protect almost all Bitcoin clients and also sustain maliciously crafted requests. This deployment is harder because it requires specific hardware and would thus be more expensive. Even so in our paper, we discuss ways to bootstrap it as well as ways to reduce costs”
Would this be considered a hardfork?
“No, it wouldn’t be considered a hard fork. If SABRE is deployed and certain nodes want to use it, they would need to have one UDP connection to the relay node. But they would still be able to connect to as many regular Bitcoin clients as they would want.
“The changes to the software have already been implemented. Now each Bitcoin client can independently choose whether or not they want to be connected to SABRE.”
Apostolaki adds that the UDP connection doesn’t require a lot of resources from the node. Hence, there will be no problems with respect to how the extra connection scales.
Will other projects be able to make use of SABRE or is it only designed for the Bitcoin network?
“Our design – especially the network design – is very general. In fact, any blockchain technology can use it. Any consensus protocol is vulnerable to partition attacks and partition attacks are practical due to BGP hijacks. Moreover, a consensus protocol tends to be communication-heavy as opposed to computation-heavy, making the use of hardware beneficial.”
In their whitepaper, Apostolaki and her colleagues state that relaying blocks is communication-heavy and not computation-heavy, enabling them to offload most of the relay operations to programmable network hardware.
During the interview, she mentions that other networks, such as the aforementioned Falcon network or FIBRE, could benefit from the routing attack protection that SABRE offers by adopting the network design and building their own networks accordingly.
What was the feedback you received after publishing the paper?
“The first paper we wrote revealed the attack on Bitcoin and received a lot of attention. People wanted to understand the concept better and find out how vulnerable we all are to the attack. Above all, they wanted to know how we could protect ourselves.
“Several mining pools and individuals contacted us, expecting to hear how one could deal with the attacks we’ve discussed. In that sense, it was interesting to voice our professional opinion and to be heard.”
Why do you believe that SABRE is the right choice?
“Routing attacks are very dangerous and practical at the same time. It is high time we deal with them. SABRE can protect all blockchain systems from such attacks while being partially deployable. Also, I am not aware of any other system that does that.“
The whitepaper adds that only “6 SABRE nodes are enough to protect 80% of the clients from 96% of the AS-level adversaries (assuming worst case scenario for SABRE).”
Do you believe in SABRE? After all, their aim is protection against relay attacks. Share your thoughts in the comment section.