Storage of Critical Master Cell Bank Files on Symbol Blockchain using NFT-Drive
Ken-ichi Tezuka
Gifu Shizui Project, Gifu University Graduate School of Medicine
Short Summary
This article reports a trial of storing critical data of Gifu Shizui Project using NFT-Drive system. Fourteen PDF files with cell bank master data were successfully stored in Symbol testnet without problems. On-chain storage of such compact and critical data of cell bank can become pivotal to ensure the feasibility of blockchain in medical traceability.
Introduction
Cells isolated from donors are cultivated and subjected to various manipulation including genetic modifications before applied to patients. In this long sequence of manipulations recording of each process is critical for safe medical procedures. Gifu Shizui Project has been developing a cell traceability system called ‘ShizuiNet’. ShiziNet is aiming at traceability from donors to patients by recording all the processes of transfer and manipulation providing safe and reliable regenerative medications.
However, until now the miscellaneous data connected to public cell banks are usually stored in local database. Such local storage of data might cause problems. For example, Gifu Shizui Project collected cells from approximately 300 donors and made frozen stocks of them. The number of tubes is more than 1000. The list of these cells are stored in forms of paper files. Until now, we tried to convert them into PDF files several times, however, each researcher copied and modified them without strict synchronization. As a result, we can not determine which one contains the current master data, and, eventually come back to the original paper file. Those paper files always have risks of degradation and can be lost in longterm storage.
Methods
We selected data created 3 years ago just before sending a part of our cell collection to Riken BioResourse Research Center, one of the biggest bank for various biological resources.
The 14 original paper materials were scanned and stored in JPEG format. The size of each file was approximately 500kB. To reduce the size of files, we used OCR to convert JPEG files into Excel files. The contents were verified and converted again into PDF files. The size of each file was reduced to 20kB.
Because, the size of data that can be stored on-chain is limited ,we paid much attention to reduce the size of each file as small as possible.
Results
NFT-Drive servers provided functions to upload PDF files to Symbol testnet. Before the processing of data files, transaction fees were sent from wallets to the server. Then, each PDF file was divided into 40 to 50 transaction messages and bonded into single aggregate transaction. The server returned Mosaic-ID (NFT) and address of an account that signed the aggregate transaction. The private key of this account was discarded to prevent malicious users can generate identical transaction with the same key.
Total of 822 XYMs were spent to save these 14 PDF files into Symbol testnet.
Next, the file stored in the testnet was extracted using the same NFT-Drive server. File hashes (SHA256) were compared with the original files as a proof that all the files recovered from Symbol blockchain were identical with original ones.
During the test, the activity of node (ibone73) was monitored for CPU load and memory consumption. The monitor showed normal level of CPU and memory usage throughout the process.
Symbol testnet explorer caught peaks of transactions generated by NFT-Drive server. Compared to normal cluster of transaction peaks generated by XYM transfer transactions (blue), the NFT-Drive peaks were constantly low suggesting the load to the network was well controlled and evenly distributed.
Discussion
This time we tried to store 14 critical master cell bank data into Symbol testnet. By reducing the data size as small as possible, the load and fees used for data storage were made minimal. If we can use ready-made mosaics in place of unique NFT mosaics, we could save the mosaic generation fees (50 XYMs/mosaic), possibly. If people are going to use NFT-Drive to store large digital data, such as pictures and movies, the cost will become much larger. In such cases, the load given to each node should also be considered. Previously, stress test was held by NFT-Drive team. Even though testnet nodes were not interrupted by heavy load generating 2GB of increase of storage size in each node in a short time, an issue was raised to cast a caution that continuous dense blocks generated by NFT-Drive and/or other transactions can cause slow down of synchronization process of nodes with limited resources. This issue should be verified and fixed, before the major release of NFT-Drive on mainnet.
Conclusion
Master data of our cell bank has long been stored as paper files. This time we succeeded to store them into Symbol blockchain, and, these data will never be modified nor lost. Users can add traceability data connected to these master data. Storing critical data on-chain will play pivotal role for traceability of cell banks and possibly applied to other medical procedures.
Acknowledgement
We appreciate NFT-Drive team for technical discussions and kind assistance.
