Launching the first iteration of our marketplace app to a handful of testers confirmed the maxim: ‘No battle plan ever survives first contact with the enemy.’ One of the key issues we discovered is that listing items on the blockchain makes no sense, due to the gas costs associated with the Ethereum network. The good news is that, like many other projects that have encountered the same issue, we can instead use IPFS to store item listing information.
Last week, the BitBoost team deployed the first version of our app on mainnet to a small group of users.
We’ve always taken a cautious approach to making releases, because any cutting-edge tech app will need to adapt to the real world. One of our major concerns was what gas costs would be on mainnet. We always expected some issues around this, since just about every project operating on Ethereum has faced the same gas crisis and have had to adapt in one way or another. However, we needed real data for our specific use case. In this respect, testnet is not the same as mainnet and it is not until the app is deployed properly that we can gain a realistic idea of the costs.
After a week of testing, it became clear that gas costs were not only high but unpredictable. There are a number of factors that impact the cost of gas, including the price of ETH, network activity and, of course, the nature of the transaction you are undertaking. (Storing data on the blockchain is typically expensive.) The upshot of this fact-finding process has been that we established the clear need to find an alternative approach to storing item listing information. Moreover, we know that the lower the gas costs for using the marketplace, the more attractive it will be for customers and businesses.
Introducing the Inter-Planetary File System
The need is for a reliable decentralised system for content storage that doesn’t impose high costs on the user. A number of blockchain initiatives have run up against this problem, especially those that need to store images and documents in a way that is trustworthy and transparent. The storage and throughput limitations of blockchains mean they are simply not suited to this application, even if they can fulfill that purpose in principle. BitBoost is currently addressing the problem of image storage for products by using the free hosting service Imgur. However, it turns out that even the limited amount of information entailed in listing an item can be too expensive to hold on the blockchain.
Like other projects that have encountered similar issues, we have decided to use the Inter-Planetary File System (IPFS), a distributed database that allows users to connect and upload data on a peer-to-peer basis. Unlike with the traditional web, which relies on individual servers and companies, there is no single point of failure. We consider this the ideal compromise between the need to maintain the decentralised nature of BitBoost Marketplace, and the ultimately unsuitable nature of the Ethereum blockchain for this purpose. (We are always looking for ways to improve our service, and in the future may well switch to using IPFS for images, too. For the immediate release, though, Imgur is adequate, and this content is not sensitive enough to require decentralised storage.)
All financial operations will continue to take place on the Ethereum blockchain, including all the smart contract functionality we have built. Payments, escrow and arbitration will remain unchanged. However, it is clear that if we want to minimise costs for users, we must limit the number of interactions with the blockchain to those that are most important. At the same time as integrating support for IPFS, we are also working with some experts from the Ethereum community on the ways to optimise gas use to ensure that fees are as low as possible.
These changes will necessarily delay our launch, since we must implement support for IPFS before we can expect mainstream users to make the switch from existing e-commerce solutions. Integrating IPFS does not affect our value proposition – in fact, it improves it and enables us to deliver a peer-to-peer marketplace that is fit to compete with traditional e-commerce platforms. We will keep the community updated about how long the process is likely to take and our revised timeline, as soon as we have more information.
It’s in the nature of any tech that is evolving as fast as blockchain that any solution will need to adapt and iterate when faced with the realities of launch. We will deal with this issue as creatively and persistently as we have dealt with previous obstacles, and will no doubt need to deal with further obstacles in the future! In the meantime, thank you for bearing with us as we make these updates.