Why is Blockchain Storage Different?

The SNIA Ethernet Storage Forum (ESF), specifically ESF Vice Chair, Alex McDonald, spent Halloween explaining storage requirements for modern transactions in our webcast, “Transactional Models & Their Storage Requirements.” Starting with the fascinating history of the first transactional system in a bakery in 1951 (really!), to a discussion on Bitcoin, it was an insightful look at the changing role of storage amid modern transactions. If you missed it, you can watch it on-demand at your convenience. We received some great questions during the live event. Here are answers to them all:

Q. How many nodes are typical in the blockchain ledger?

A. As many as are required to ensure that a single node or small number of nodes can’t crack the hard problem that gives you the right to add your blockchain as the next. There are estimated to be more than 11,000 Bitcoin nodes right now (see https://bitnodes.earn.com/), but not all blockchain systems have as many nodes (or such a hard problem to crack!)

Q. A traditional DB like Oracle…how does it fit into CAP? What two features would it have?

A. A traditional database doesn’t have an unreliable geographically spread network connecting parts of the database (or at least, it shouldn’t). The CAP theorem (Consistency, Availability & Partition tolerance – pick any two from three) doesn’t apply to traditional databases like Oracle. These kinds of systems provide ACIDity; Wikipedia has an excellent article on the subject.

Q. Is block same as node? I thought each node has a copy of each block.

A. That’s correct. Ledger transactions and blocks get distributed to all the nodes, and every node has a copy of the entire blockchain.

Q. Hey! Bitcoin isn’t just limited to criminals! C’mon!

A. Blockchain is more than cryptocurrencies, and is increasingly important for legit businesses too. For instance, just this week American Express Is Getting Into Blockchain-Based Payments With Ripple. Other applications across a broad range are being introduced; in healthcare for instance, for digital identity, smart contracts and digital voting to name a few.

Q. Why is blockchain storage different from more regular storage we might have today?

A. Let’s take the example of a cryptocurrency like Bitcoin. Blockchain ledgers are transactional in nature. So if I have 200 transactions in my wallet, they are spread across all the blockchains where they are recorded, and there isn’t a central value of what’s in my wallet. To get that, I need to add up all the wallet transactions recorded in every blockchain; and the blockchains can be pretty big. The Bitcoin blockchain is more than 140GB as of November 2017, and growing. That demands fast (both high bandwidth and low latency) storage; in memory as much as possible, with Flash based SSD for persistence. Other blockchain technologies will have equivalent or even more demanding requirements.

If you have questions about this topic, please comment on this blog below. And if you’re interested in more vendor-neutral education on storage topics, check out the full library of SNIA ESF webcasts where we cover all thing related to Ethernet storage, including block, file and object storage, RDMA and NVMe over Fabrics, storage performance benchmarking, containers and our very popular 101 series “Everything You Wanted To Know About Storage But Were Too Proud To Ask.” Happy viewing!

Leave a Reply

Your email address will not be published. Required fields are marked *