ON–046: DeFi (Part 2)

• Half of the fees generated goes to NXM stakers in staking rewards. Below is a breakdown of staking rewards per contract. Ren stakers have earned the most with $68k in rewards, followed by Set Protocol and Curve. (Source)

• Nexus added a proof-of-loss exclusion to covers, which means that cover purchasers need to prove they lost money in a hack in order for claims to be paid out. Previously, Nexus cover acted more like credit default swaps rather than insurance cover, so the proof-of-loss exclusion will prevent people from buying covers just to speculate on projects getting hacked.

• The uncle rate is an important metric often directly related to the health of the Ethereum network: the smaller the uncle rate, the better the network performs. For Ethereum mining pools, the uncle rate is seen as the key measure to indicate the mining pool’s infrastructural capability and, ultimately, how efficient it is at mining new blocks that are included in the longest chain. Figure 1 shows Ethereum’s uncle rate over the past year (Source).

A higher gas limit means pools can produce larger blocks, which some have worried would lead to an increase in pools’, and thus the network’s, uncle rate. However, we have shown in a previous post that there was no statistically significant increase in the uncle rate after the gas limit was increased by 25%, implying that a much higher gas limit can be safely deployed. At the same time, our observation was that despite the deployment of an advanced Blockchain Distribution Network (BDN) infrastructure, the uncle rate wasn’t diminishing towards zero. This motivated us to dive deeper and further our understanding of the problem.

We initially conducted an experiment over a 2-week period, which included 3,727 blocks. Figure 2 summarizes the results. We used the block-to-BDN entry point to measure because it clearly indicates scenarios where BDN helps and where it is less effective. In particular, if two blocks of the same height have the same entry point, then the BDN could help a lot. On the contrary, if the delta is larger, then the BDN’s effect is limited.

Figure 2 shows that, for example, an entry in the Bin of 1500 means that the uncle block entered the BDN between 1 – 1.5 seconds before the other block in the fork. As can be seen, in half the cases (Bin 500), block propagation would have a huge impact. However, for 12.5% of the cases, uncle blocks were propagated a full two seconds or more before/after the same height block. This is a clear indication that the network (connecting mining pools) is not the only reason for the uncle rate. It is the network infrastructure within a mining pool that also counts.

We were able to prove the above hypothesis. In collaboration with a mining pool (i.e., F2Pool). In particular, the pool modified their Geth image so that they could send newly-mined blocks as soon as possible to bloXroute’s BDN. As described above, and as verified below, this had a significant impact on reducing the uncle rate, as shown in Figure 3.

• In addition, we deployed two other changes: (a) the collaborating pool took full advantage of BDN-hosted gateways, which allows users to always connect to the latest version of the BDN, i.e., by adding hosted gateways as trusted peers for all of its ETH mining nodes. (b) BDN placed their hosted gateways in locations with lower latency to pool nodes, while the pool relocated their nodes according to the location of existing hosted gateways. The results, measured both by the BDN (Figure 3), and by the pool (Figure 4), show that the uncle rate decreased by 50%.

• We demonstrated that end-to-end performance, which both includes network between a client and a mining pool (covered by BDN), and between a mining pool and miners (covered by a mining pool), are essential for reducing the uncle rate. Reduced uncle rate makes the entire Ethereum network more scalable and sustainable, while at the same time provides the necessary stability and revenue increase for the mining pools and miners alike.