Marlin’s P2P ‘SuperHighway’ can Solve Blockchain’s ‘Slow Speed’ Problem: Nilotpal Mukherjee

It has become more like a redundant quote to speak about how reports keep popping up about the Indian Authorities’ aversion to Cryptocurrencies, while at the same time, Indians are successfully working towards building Blockchain solutions.

After a successful fund raise under Binance’s new IEO program by Matic, yet another Blockchain company with Indians at helm have bagged the Binance support, albeit this time it is from Binance Labs, the investment arm of the exchange.

It is none other than Marlin Protocol. We connected with Nilotpal Mukherjee, the chief business officer at Marlin to learn more about how the protocol works and what innovation is it bringing to the table. As it always is with Coin Crunch, we bring you the most exclusive stories.

In an email interview, Mukherjee answered all our questions regarding Marlin. Some answers below have been edited to make them suitable for a short read.

CCI: Tell us something about the Marlin Protocol

Nilotpal: “The scaling challenges of blockchains and some solutions to solve them such as sharding, state channels and side chains are well known. However, these techniques are pretty complex and difficult to implement securely or compromise on decentralisation. Marlin’s P2P superhighway offers a unique approach to overcome the throughput bottleneck of blockchains by providing a lightning-fast alternative to the existing peer-to-peer packet propagation model.”

“Enhanced networking allows nodes in a P2P network to communicate faster and transmit greater amount of information in a shorter duration thus increasing the system’s throughput. In comparison to the layer-1 and layer-2 scaling solutions mentioned earlier, Marlin’s layer-0 scaling solution is blockchain-agnostic which means that unlike Lightning or Raiden neither are we tied to any particular blockchain nor do we create new ecosystems like Matic or Celer. Instead, we can scale all existing and upcoming blockchains irrespective of whether they use Proof of Work, Proof of Stake or any other consensus mechanism.”

CCI: Tell us about the Key People in Marlin

Nilotpal: “Marlin has a small but productive team with rich experience in building and shipping code as well as in business. Most of our core tech team graduated from IIT Bombay. Our CEO Siddhartha worked in the Cloud division of Microsoft before moving on to become one of two core devs at Zilliqa back in 2017. Pratyaksh and Prateesh who lead R&D did their masters and PhD from Stanford and MIT respectively. They both have a strong research expertise in the area of networking with Pratyaksh also having worked at Facebook and Pinterest. I, on the other hand, lead business development and marketing along with my colleague Scott. While I have been part of the consulting and strategy teams at Accenture and Deutsche Bank, Scott got into Ethereum pre-ICO and was most recently Director of Global BD at Wanchain.

A good team is incomplete without great mentorship. Thus, we are proud to be advised by the industry’s finest – Rogelio Choy, the CEO of Bittorrent, Hari Balakrishnan, a Chair Professor at MIT and CTO of Cambridge Telematics, Ethan Fast, the cofounder of Nash Exchange and City of Zion and Prabhakar Reddy, former VC at Accel Partners.”

CCI: The whitepaper has a lot of complex equations involved, can you simplify it for us?

Nilotpal: “Blockchain networks can be represented on a simple architecture of 3 planes — Network plane, Consensus plane and Ledger plane. The network plane is where the P2P propagation of messages between nodes happen, the consensus plane consists of the mechanism through which a set of untrusted nodes arrive at an agreement on a set of valid transactions while the Ledger plane is where the blocks finally get stored once they are mined. 

Now, blockchain throughput is measured by the number of transactions per second that is

Throughput = Transactions/Block * Blocks/second (inverse of block interval),

where Transactions/Block depends on the ratio of block size to an average transaction’s size.

Thus, to increase throughput one could either increase the transactions per block or the blocks produced and transmitted per second like BCH or LTC respectively.

Marlin’s decentralised overlay network offers a performant alternative that eliminates the very network bottleneck. The enhanced networking reduces multicast latency by dynamically choosing better paths optimised for latency, congestion, packet drop rate and bandwidth. Path conditioning and dynamic traffic steering along with cut-through routing, multi-path routing, FEC etc provide blockchains with private network like performance over public networks allowing them to scale without compromising on security or decentralisation.

Putting it simply, most of the equations in the paper are an attempt to explain the mathematical models around this construct and validate the network optimizations that Marlin intends to achieve.”

CCI: Once again, barring the math, can you explain how the Marlin protocol will work, including how blocks are created, managed, audited and miners are incentivised?

Nilotpal: “The working of the Marlin network can be broken down into four components:

Creation:

A node first needs to register with the Marlin network, then either create its own relay network or join an existing relay network from the marketplace.

Management:

Once a relay network is created using a set of available Marlin nodes, it is important that the network be smoothly managed so that poorly performing individual nodes do not harm the network as a whole. Miners are subscribed to all relay networks serving the blockchain they mine on and randomly pick a few relay networks to publish blocks. These randomly chosen relay networks are expected to notify and deliver blocks to all subscribers with pre-negotiated performance guarantees.

Packet propagation:

SDK: Miners, wallets and other users that use the Marlin Relay network integrate the Marlin SDK which is responsible for interaction with Marlin nodes. Protocol developers can write plugins to customize pre- and post-processing functions and execute custom code for packets related to their protocol at Marlin nodes.

Block Entry: A verifiable unbiased source of randomness R is used to choose n relay networks from those advertising the capability to serve the platform that a miner producing the block should send the block to. R and n can be customized by the protocol designers if required as per their security considerations.

Block Traversal: The nodes in a Marlin relay network form a mesh network with multicast paths optimized depending on the source. The packet loss rates, bandwidth and latency of the different paths are constantly monitored to ensure that multicast delivery of blocks using cut-through routing can be completed in as little time as possible.

Block Exit: As soon as a Marlin node receives a new block it sends its hash to all its subscribers, which further passes it forward to its own subscribers. Subscribers hearing about the hash for the first time proceed to request the block.

Auditing:

A number of anonymous auditor nodes monitor the working of Marlin nodes by subscribing to them as any other subscriber. A Schelling scheme is used to incentivize the auditors to report results correctly. Our goal is to keep the auditing scheme independent of the relay network and allow third-party reporting schemes to be plugged in as and when it is mutually acceptable by both relay networks and its users.

Incentivization: Since the Marlin network is comprised of public nodes it is crucial to incentivize them for the resources expended. This is accomplished through fees and staking rewards. In addition to serving as incentivization mechanism, fees serve as a deterrent against DoS attacks by users while staking ensures that byzantine nodes can be penalized for malicious actions.”

CCI: What are the advantages of a Decentralised Relay network such as marlin over other scalable solutions like sharding, tangle, hash graphs?

Nilotpal: “The last few years have seen projects primarily emphasizing on the consensus plane. Both on-chain and off-chain scaling solutions have been widely explored with on-chain projects like Zilliqa, Dfinity, EOS trying to modify the consensus algorithm to achieve faster throughput while side-chain and off-chain projects like Plasma, Lightning Network, Raiden Network etc achieve scale by computing off the chain or on parallel chains and storing the final output on-chain. But most of these have either failed to meet the needs of security and/or decentralization or have fragmented the network enough causing concerns of adoption. Moreover, solutions such as sharding have an inter-shard communication problem themselves while several secure PoS consensus algorithms make assumptions of partially or fully synchronous networks placing requirements that can’t be fulfilled by Internet today and thus limiting the blockchains to run on a single datacenter.

Marlin forms a versatile peer-to-peer networking infrastructure that’s interoperable with any kind of blockchain platform and can become a plug and play solution for scaling Blockchains without the need to create new chains or new ecosystems. Thus Marlin is complementary to all the other solutions mentioned above and can effectively offer a scaling factor to most of the existing scaling projects in Blockchain Ecosystem.

Given Marlin lies at the base of both Layer 1 and Layer 2 and all such Layer 1 and Layer 2 solutions too use some form of P2P messaging for node to node communication, Marlin becomes a plug and play solution for scaling Blockchains without the need to create new chains or new ecosystems. To put in other words, instead of creating new vertical models of scaling, Marlin runs across horizontally as a scaling solution for all Blockchains and makes no compromise on decentralization.

Thus, Marlin is complementary to all the other solutions mentioned above and can effectively offer a scaling factor to most of the currently existing scaling projects in Blockchain Ecosystem. The existing projects can have their scaling components and multiplying them by scaling factor of Marlin present below at the network level to further speed up their models.”

CCI: How can existing networks be brought in to work with Marlin, I mean to ask are they interoperable?

Nilotpal: “Marlin’s interoperability with any kind of blockchain platform is one of its major strengths. Not only is it agnostic about the Layer-1 consensus mechanism, but also integrates with several Layer-2 scaling and privacy solutions. The APIs provided are similar to those provided by existing P2P networking stacks making integration smooth and simple for developers as well as miners. “

CCI: Who do you consider a competition in the space?

Nilotpal: “Marlin is the only decentralized layer-0 scaling solution being built. Developing a network-layer scaling solution is considerably hard but assembling the right team with rich research experience in P2P networks as well blockchains proved even harder. We are proud to have assembled a team of successful veterans in these areas uniquely positioned to solve the scaling problem in this novel way. Nonetheless, there exist alternate scaling solutions such as those being developed by Celer, Starkware, Coda, Zilliqa, Prysmatic, Lightning Labs etc.

However, due to our blockchain-agnostic nature and the experience of our founder at Zilliqa, we are aware of challenges with several existing techniques and are rather complementary than a competitor to several such solutions with the ability to augment their scaling initiatives.”

CCI: Is LIN token up for an ICO/IEO soon?

Nilotpal: “Due to legal reasons we cannot speak about the same.”

CCI: I did not see a roadmap on the white paper, is there one? If there is, can you share some important timelines?

Nilotpal: “The project is currently undergoing heavy development. The goal is to open-source the code-base in the coming months with libp2p like interfaces so that developers building on Tendermint and Substrate can easily get started with leveraging the benefits that come on using Marlin.”

CCI: Tell us more about your partnership with WandX and Murmur? Also, if MurMur is on EOS, how will it work with Marlin?

Nilotpal: “Marlin and Murmur are engaging and exploring efficient data, content and notification delivery mechanisms to build a completely decentralized high-performance social networking platform. With WandX, we are exploring network-layer optimizations that will enable faster order display and lower confirmation times necessary to build a high-performance DEX. 

Marlin is blockchain-agnostic; applications running on blockchains other than ETH can still use Marlin for their high-performance networking needs.”

---

If you enjoyed reading this story, donate to support us so we can bring more of these to you and remove advertising from our website.