• Merged mining refers to the use of the work done for one blockchain (i.e., parent blockchain) on other smaller child blockchains, using Auxiliary Proof of Work (AuxPoW). Historical examples include Namecoin (with Bitcoin), Dogecoin (with Litecoin) and Myriadcoin (with both Litecoin & Bitcoin).

• Created in December 2013, Dogecoin (DOGE) adopted the merged mining model in August 2014. Within one month, Dogecoin’s hashrate/mining difficulty increased by +1500% as large mining pools widened their operations.

• Since then, Dogecoin’s hashrate has exhibited an extremely strong and positive correlation (0.95) with Litecoin’s hashrate. As of July 2019, nearly 90% of Dogecoin’s total hashrate comes from large Litecoin mining pools, with its blockchain processing around 30,000 transactions per day.

• While greater network security is the main purpose for implementing merged-mining, there are also potential disadvantages and risks to consider for both miners and the blockchain itself.

• Specifically, the value proposition of the underlying cryptoasset needs to be compelling enough to attract a critical mass of miners to switch their current mining set-up to a new one that supports merged mining.

• Given that block mining rewards are halved every four years for both Litecoin and Bitcoin, merged mining could potentially become a solution to maintain network security in the long-run as newer cryptoassets, with higher block rewards, could be merge-mined within the same pools.

Dogecoin (DOGE) and Litecoin (LTC) have been two of the longest-standing cryptocurrencies, each registering around 30,000 daily on-chain transactions.1 Dogecoin, which launched on December 6th, 2013 as a meme currency, is a fork of Litecoin, hence it relies on the same hashing function (Scrypt).

In 2014, Dogecoin’s developers, along with the community, decided to switch to an Auxiliary Proof of Work (AuxPoW) algorithm in order to maintain network security through reliance on Litecoin’s mining workforce. As a result, Litecoin miners could now obtain an additional cryptocurrency for their efforts while providing the same amount of hashpower as before.

Given that Litecoin block mining rewards will halve again in August 20192, this report will attempt to assess the impact of this upcoming catalyst on Dogecoin by discussing the advantages/disadvantages of merged mining from both the miner and project team perspectives, based on an analysis of Dogecoin’s merged mining policy established in 2014.

1. What is merged mining?

1.1 Description of merged mining

Blockchains rely on different consensus mechanisms along with specific procedures to decide who validate each new block. Blockchains for cryptoassets such as Bitcoin, Litecoin, Monero or Dogecoin all rely on Proof of Work (PoW) along with specific consensus protocols like the Nakamoto consensus for Bitcoin.

Miner competition is critical to the security of PoW blockchains; miners have finite resources and must efficiently allocate hashpower to blockchains with the highest expected revenue3.

As a result, merged mining was introduced as early as 2011 with many different definitions; here are two of the clearest definitions:

“Merged mining refers to the act of mining two or more cryptocurrencies at the same time, without sacrificing overall mining performance. Essentially, a miner can use their computational power to mine blocks on multiple chains concurrently through the use of what is known as Auxiliary Proof of Work (AuxPoW).”

Source: Binance Academy

“The act of using work done on one blockchain on more than one chain, using Auxiliary Proof of Work (AuxPoW).”

Source: Bitcoin Wiki

In summary, merged mining relies on a reputable set of miners from a relatively more established coin in order to secure other (smaller) blockchains, at no added explicit cost to the larger chain’s miners.

Conversely, Auxiliary Proof of Work (AuxPoW) can be defined as:

“The relationship between two blockchains for one to trust the other's work as their own and accept AuxPOW blocks.”

Source: Bitcoin Wiki

In the next section, we will dig into historical examples of merged mining cryptocurrencies.

1.2 Historical examples

The first cryptocurrency to adopt a merged mining model was Namecoin (NAME) which switched over from PoW in 2012. Namecoin aimed to solve issues related to DNS (Domain Name System) owing to potential censorship from centralized registration systems that could shut domains down4. Launched in April 2011, its merge-mining implementation occurred at block 19,200 (October 2011) which was quite innovative for what is commonly referred to as “the first alt-coin”.

Since then, other cryptocurrencies have adopted merged mining. Some cryptocurrencies even use multiple hashing functions, such as Myriadcoin (XMY), which support 5 different algorithms (SHA256, Scrypt, Myr-Groestl, Argon2d, and Yescrypt).

However, the most popular example of merged mining is still Dogecoin (DOGE), created in December 2013. Often referred to as a “meme currency” in its early days, it quickly attracted significant attention from market participants. Specifically, on-chain transactions were fairly cheap and the “meme component” made it very popular on Bitcointalk and crypto-communities on Reddit. Unlike Namecoin whose merge-mining protocol was an expected feature from the start, Dogecoin adopted merged mining at block 317,337 (July 2014).

It is worth noting that no single cryptoasset has ever shifted away from AuxPoW since it adopted a merged mining protocol.

Recent examples of cryptoassets such as Elastos (mined with Bitcoin) serve as a reminder that merged mining as a concept is not obsolete and could be further investigated for existing and future PoW blockchains. As the crypto-industry focus has shifted to PoS and sharding for scalability, PoW merged-mined child-chain blocks might still potentially serve scalability functions in the future.

2. Tale of a Doge: Dogecoin and Litecoin merged mining

Table 1 - Comparison between Bitcoin, Litecoin and Dogecoin (July 7th 2019)

6

2.1 Correlations in mining difficulty/hashrate

Chart 1 - Dogecoin and Litecoin hashrate (THz/sec)

Sources: Binance Research, BitInfoCharts

In September 2014, Dogecoin’s hashrate increased by more than +1500% as miners forked over their mining operations to start including DOGE.

Since then, Litecoin and Dogecoin’s hashrates have moved in line with each other. The monthly percentage change in hashrates of Litecoin and Dogecoin have been highly correlated with an extremely high correlation coefficient of 0.95.

It is also worth noting that a sharp increase in the spread between the two hashrates occurred on December 2017, owing to new mining pools that started to mine only Litecoin.

Table 2 - Correlation between monthly changes (%) in hashrate

In comparison, the strength of the relationship between Bitcoin hashrate and Dogecoin/Litecoin hashrate was lower - yet still statistically significant - with a correlation coefficient around 0.30-0.35. This could potentially signal that factors such as the overall market cap of the industry play an important role in the hashpower dedicated to mining. More studies should be conducted in order to analyze whether this is consistent across other major PoW cryptoassets.

Chart 2 - Dogecoin and Litecoin mining difficulty (MHz)

Sources: Binance Research, BitInfoCharts

Unsurprisingly, Litecoin’s mining difficulty is higher than Dogecoin’s, with the ratio between the two tightly related to the difference in blocktime between Litecoin and Dogecoin.

Since September 2014, the following equation fairly approximates the mining difficulty of Dogecoin to Litecoin’s.

This relationship is explored further in chart 3, as seen below.

Chart 3 - Hashrate and difficulty ratios between Litecoin and Dogecoin

Sources: Binance Research, BitInfoCharts

As illustrated above, both hashrate and difficulty ratios, defined by Litecoin hashrate/difficulty divided by Doge hashrate/difficulty, have remained fairly constant since September 2014.

Interestingly, the hashpower dedicated to Dogecoin was sharply increasing before the change became effective on July 2014. After Dogecoin’s hard-fork, existing miners stopped mining, which led to easiness to mine before Litecoin pools took over, between late August 2014 and September 2019. This led to a drastic sharp in hashrate and difficulty ratios.

Chart 4 - Dogecoin and Litecoin daily transactions

Sources: Binance Research, BitInfoCharts

Interestingly, Dogecoin displayed more daily on-chain transactions than Litecoin until June 2017. From June 2017 to January 2018, Litecoin displayed more transactions than Dogecoin. Since then, Dogecoin and Litecoin have exhibited a similar amount of daily on-chain transactions.

Though Dogecoin transaction fees are relatively lower (which could lead to more transactions than on the Litecoin blockchain, in absolute numbers), a cryptocurrency can only adopt a merged mining model (with AuxPoW mechanism) if enough miners find value in undergoing required technical and maintenance work to upgrade their mining protocol.

2.2 Overview of existing Litecoin and Dogecoin mining pools

In this section, Litecoin and Dogecoin mining pools will be analyzed to see what pools are involved in merged mining and to check whether Dogecoin’s mining activities more concentrated than Litecoin’s.

On July 8th 2019, Litecoin’s estimated hashrate was around 454 TH/s. Specifically, the four largest mining pools accounted for more than 51% of the network, as illustrated by the below table.

Table 3 - Largest mining pools for Litecoin as of July 8th 2019

In comparison, Dogecoin’s estimated total hashrate stood at around 415 TH/s on July 8th 2019, representing 91% of the current hashrate of Litecoin.

Table 4 - Dogecoin mining pools as of July 8th 2019

For the purpose of analysis, we will focus on these key pools, as if all the other LTC miners (not included in table 3) were also mining DOGE, their cumulative hashpower would be lower than the “Others” hashrate from table 47. As a result, around 8.5% of the hashpower for DOGE is exclusively allocated to this single blockchain.

Despite being merged mined, Dogecoin mining is actually less “concentrated” than Litecoin, with some miners actually solely mining Dogecoin and not Litecoin.

2.3 Economics and risks of merged mining protocols

2.3.1 Shortcomings from a miner’s perspective

From a miner’s perspective, there are several reasons why Dogecoin might not be included in some of Litecoin’s mining pools. Here are the three key reasons:

• Maintenance work: In order to support new child blockchains, mining pools would need to adjust their mining setups, wallet management, and disbursal processes.

• Anticipated child-chain reward value expected to decline: If miners perceive the proceeds from the child chain as worthless, they will have no incentive in securing the auxiliary network.

• Lack of awareness: some (potentially newer) pools may not be aware of merged mining opportunities and would not be aware of the potential extra rewards.

In summary, if a mining pool decides that the maintenance work (i.e., costs) associated with supporting child blockchains is too expensive relative to potential rewards, these pools will likely avoid hard-forking their operations in order to switch to a different mining protocol.

2.3.2 Potential future economics of merged mining

As described in the previous section, there are a few miners that mine DOGE without mining Litecoin, meaning that for some Dogecoin stakeholders, its block rewards have a higher expected future value than Litecoin rewards. Other options would include either a lack of information from miners or the absence of rationality from Dogecoin miners.

Merged mining could also play a decisive role in how child chains support decreasing mining rewards of parent chains such as Bitcoin and Litecoin. Every four years, both Litecoin and Bitcoin are scheduled to halve their mining rewards8. With increased competition from the proliferation and development of layer-2 solutions, Bitcoin and Litecoin may eventually provide lower incentives for miners to continue operations, as their revenue streams from transaction volume could potentially decrease, with Bitcoin becoming even further used as a store of value rather than a medium of exchange.

Conversely, merged mining may also potentially serve as a solution to maintain such legacy blockchains (e.g., Litecoin, Bitcoin). If new child blockchains were to be added through merged mining, these could potentially bring an extra stream of revenue for miners, significant enough for miners to maintain their operations for mining both (and all) chains.

For instance, let’s assume prices to be constant with Dogecoin being equal to USD0.0035 and Litecoin being equal to USD121. At current prices in USD terms, Dogecoin’s mining rewards would outweigh Litecoin’s mining rewards by around 2039-2040.

Chart 5 - Breakdown of mining rewards until 2050 for Dogecoin/Litecoin merged-miners

However, this scenario is quite unlikely, as both prices will not remain constant over time. Specifically, rational market participants would expect Dogecoin price to “depreciate” relative to Litecoin owing to the differential of inflation rates among the two cryptocurrencies. By being a deflationary currency by design, Litecoin is a better store of value than Dogecoin.

On July 5th 2019, Binance listed Dogecoin9, thus providing a more liquid market for miners to sell their coins to cover their fiat operating costs. However, it remains to be seen whether this recent listing could incentivize more Litecoin mining pools to also hard-fork their pooled mining protocols into a merged mining one that includes Dogecoin. With this new listing, Dogecoin market depth is also likely to increase, making it easier for miners to realize the value of their mined DOGE, especially with listings against USD stablecoin pairs10.

Shortcomings from a project team’s perspective on merged mining

From the perspective of any project team working on a PoW cryptoasset, several shortcomings must be considered when looking at the adoption of a merged mining protocol. Some key disadvantages11 include:

• Concentration risk: regardless of whether a cryptoasset provides a value proposition compelling enough for miners to change their mining protocols, a cryptoasset adopting a merged mining model would expose itself to a threat of a 51% attack if not enough large miners from the existing parent chain start mining “child cryptoasset(s)”. For instance, if one single miner from the parent blockchain started mining the child blockchain, it could easily reach up to 51%12 of the required hashpower to conduct an attack. However, this is similar to the overall problem about how to secure a PoW blockchain in its early stages where the value is not yet established, which is why many new chains choose unique hashing algorithms to level the playing field for miners upon launch.

• Potential new attack vectors: as the merged-mined cryptoasset is essentially mined “for free” (i.e., without additional hashpower required), malicious miners could be more tempted to launch an attack (e.g., selfish attack) on the child blockchain. However, malicious mining pools could potentially be penalized by its participants which would transition to other mining pools, reducing pool-specific economies of scale as a malicious attack on a child blockchain would be interpreted as a “negative signal”.

• Dependency on the parent blockchain: if a cryptoasset wanted to change its hashing function down the road, it would need to hard-fork and move away from the parent chain, leading to a subsequent drop in hashrate and difficulty for the network. Afterward, the value proposition of the blockchain would need to be convincing enough for miners to drop mining on the parent blockchain and switch over to exclusively mine ex-child blockchain. As a result, child cryptocurrencies could end up being hamstrung in their development and adoption, with little possibility to opt-out from the established merged mining model. Furthermore, as mining child chains is essentially free, they could keep mining the child chain with empty blocks, making the transition to the updated ex-child blockchain misleading for users. With the lack of empirical cases, it remains to be seen how a transition would occur.

Whereas these potential disadvantages must be considered by project stakeholders about whether merged mining should be implemented, no attack has ever been conducted on child blockchains which may indicate that the industry might have a majority of benevolent miners.

3. Conclusion

Merged mining, despite its introduction in 2012 by Namecoin developers, has not been extensively been used by PoW blockchains. Dogecoin has been the most successful example of merged mining, operating securely for almost six years. However, other past examples such as Namecoin and Myriadcoin serve as a reminder that any blockchain, if not used, has little value. In light of these cases, it is all the more impressive that the Dogecoin hashrate increased by +1500% after forking over to AuxPoW, retaining a large majority (91%) of Litecoin’s hashrate to this day.

Proof of Work (PoW) has recently lost popularity in favor of Proof of Stake (PoS) consensus mechanisms for a variety of reasons, including but not limited to high mining costs and environmental impacts. Other smaller chains could potentially move to AuxPoW in the coming years in order to maintain greater network security while reducing the need for a seperate miner set.

However, merged mining could potentially lead to unfair concentration of market participants if only a few large mining pools were to actually switch to a merged mining protocol, ultimately leading to smaller blockchains being even more exposed to the risk of a 51% attack. Furthermore, other risks such as dependency on parent blockchains or new attack vectors are to be considered from the perspective of coin-holders and project teams interested in merged mining. As a result, Auxiliary Proof of Work (AuxPoW) would only be useful if a chain has an actual user-base and a value proposition compelling enough to bring miners to update their mining protocols.

With future block reward halvings scheduled for both Litecoin and Bitcoin, merged mining could potentially provide an opportunity to increase incentives for miners to continue mining these “legacy”, high-value, PoW cryptocurrencies. This could be achieved by setting new or existing child chains whose high-block rewards are larger than the parent chains through merged mining. While it is hard to define whether Dogecoin, an inflationary currency, may bring any direct value to the Litecoin network, it is evident that Dogecoin has in some way supported the popularity of Litecoin, with a number of Litecoin mining pools changing their setups to include support for Dogecoin through merged mining.

References

• Bitcoin Wiki. Merged mining specification. https://en.bitcoin.it/wiki/Merged_mining_specification

• Bitcoin StackExchange. Thread: “How does merged-mining work”(2016). https://bitcoin.stackexchange.com/questions/273/how-does-merged-mining-work

• Butterin, V. Bitcoin Magazine (2014). https://bitcoinmagazine.com/articles/side-chains-challenges-potential

• Judmayer, Stifter, Voyiatzis, Zamyatin. Merged Mining: Curse or Cure? (2017) https://eprint.iacr.org/2017/791.pdf

• Namecoin GitHub. https://github.com/namecoin/wiki/blob/master/Merged-Mining.mediawiki

• Namecoin Website. https://namecoin.org/

• Zamyatin, A.Taking a Deep-Dive into Merged Mining (2018). https://medium.com/@alexeiZamyatin/a-technical-deep-dive-into-merged-mining-5b67706e1a19

1. Data as of June 2019.↩

2. See our report from April 2019 about Litecoin’s halving. https://research.binance.com/analysis/litecoin-halving↩

3. ASIC mining led to further professionalization of PoW mining and as a result, “mining allocation” cannot be moved from one coin to another (e.g., Litecoin miners cannot use same ASICs to mine Bitcoin).↩

4. https://medium.com/@bdratings/namecoin-the-first-altcoin-8e9b8eb9160b↩

5. https://arstechnica.com/information-technology/2014/02/dogecoin-to-allow-annual-inflation-of-5-billion-coins-each-year-forever/↩

6. ASIC-resistance is discussed in one of our previous reports named “Examining the Implications Of Monero's Latest Fork”. https://research.binance.com/analysis/monero-hard-fork↩

7. The remaining hashrate from other participants is inferior to 2 TH/s. It is lower than the hashrate from “Others” section in the table 4.↩

8. See our report about Litecoin’s halving scenarios. https://research.binance.com/analysis/litecoin-halving↩

9. https://www.binance.com/en/support/articles/360030488211↩

10. Dogecoin, owing to its small price, is not suitable to trade against BTC as its minimum spread is around 3.50% as the precision of a price level in BTC pairs must be at least equal to 1 satoshi.↩

11. Most of these ideas are discussed in the research paper named “Merged Mining: Curse or Cure?”.↩

12. Other attacks, such as the selfish mining attack, require less hashpower.↩