New Alpha Release, Mining Slow Start


#1

here on the Zcash blog: https://z.cash/blog/new-alpha-release-mining-slow-start.html


#2

fetch-params.sh refers to a z4-proving.key but I don't see it listed in https://z.cash/downloads/

All the same, it's good to see the progress that's been made.


#3

So the slow start will take 34 days if i am correct?


#4

Well, despite my previous comment, I've just updated and I notice that the signing key is now down to 1.1 GB. Nice to see.


#5

Wouldn't using a slow start make Zcash vulnerable to malicious actors creating forks, and/or causing a denial of service attacks very easy at the start of mining? This would be due to the fact that there is effectively zero economic incentive to mine in the beginning, so there will likely be much less hashing power making this attack more likely. Haven't some alt-coins been attacked previously at launch due to low initial hashing power? If a slow start is still desired would it not make sense to start at a non-zero value for the first block like 10 or 25 rather than zero?


#6

There's zero economic incentive currently but the difficulty already needs to be high enough that I've only managed to mine one block so far. My guess is that mainnet will attract even more participants.


#7

Is there any way to determine the current network hashrate? Ultimately, the fact that you can't find many blocks doesn't mean much. The question that needs to be asked, is what could someone do with $10k - $100k USD with some cloud computing time. You need to understand the economic cost of conducting an attack versus the potential economic gain. If Zcash is forked at the beginning of mining, it would help cause loss of confidence and create an opportunity to offer a fork of the code, which doesn't suffer from the "bug" of not having economic incentive to mine.


#8

Two guys from the UK and a US IT professor spent $250k in 2014 when Monero launched on Amazon cloud instances. They made a fortune. I expect at least a million usd will be spent on cloud computing when zcash goes live. Add a couple million zombie machines run by 4-5 botnet operators and what you get is a completely unfair environment.

Botnets are going to dominate this coin and I wouldn't be surprised if 80% of hash rate / block rewards go their way...

If you think you will mine anything with a single or a couple of late gen i7 machines - i guarantee you won't.


#9

Spot instance generally cost a few cents apiece per hour. For example, a spot instance I recently used cost $.04. If a person were to spend $250k on spot instances for a month, at $.04 that would mean they are running about 9000 spot instances. At 2.5 minutes a block, there should be about 17280 blocks per month. Thus, if the mining world was a person with one machine (equivalent to a spot instance) and a 9000 spot instance cluster, the person with one machine might still expect to mine one or two blocks per month.

So it's not completely hopeless, but they would be better off joining a mining pool, when someone gets around to making one.


#10

Do the math again to account for a 1 million zombie botnet. Trust me, 1 million is a small number compared to what is out there...


#11

Well obviously you'd never have a shot if you are competing against a 1 million computer strong botnet.

But then again, a person spending a million dollars a month on AWS would only have 35,000 instances running--that is about 3.5% of the million computer botnet. That person spending a million dollars might get around 600 blocks, or 6000 coins. So against a million computer botnet, a person spending a million dollars would need each coin to be worth $167 to break even! In short, a million computer botnet would screw all miners. :sweat:


#12

That's what I am saying... Botnets will dominate mining just like they dominate Monero mining. Keep in mind Monero can be mined on high end machines only, the number on low-end zombie machines outnumber high-end zombie machines by a factor of 1000.


#13

The size of the block reward is zero at the start of mining, therefore no rational economic actor would participate in a non-malicious way at the start of mining. No one would use bot nets or AWS because there is no reward for mining.


#14

Exactly, they will all wait until the block reward is at maximum. ZCash better not use bitcoin's difficulty retarget algo because once the botnets start mining way more blocks will be mined than expected before the next diff. retarget.


#15

We want to prevent specialized hardware for mining to make the algorithm more equitable to encourage distribution of the hashrate globally. Necessarily that will involve chosing an algorithm that is CPU/GPU friendly and ASIC unfriendly. I don't see how you can have that and prevent botnet operators from mining,

I don't find the threat from botnet operators that compelling; if our algorithm consumes a ton of memory, it may pose a threat to the stability or stealthiness of their zombies. I also doubt it will be as common as you think.


#16

Is difficulty the only factor to prevent a quick succession of blocks? What would be the effect of the network insisting that at least x seconds have transpired before accepting a new block?

Would this help reduce (near)empty blocks from occurring (and other issues such as pending transactions that remain unprocessed)?

btw If ZCash has done away with nodes, do miners maintain their own memory-pools? Where does the network store in-coming transactions?


#17

The problem with this kind of metric is that timestamps in block headers are untrusted. Any miner can put whatever timestamp they want in that field (within reason - blocks more than two hours ahead of the local node's understanding of the current time are rejected). The reason we can't trust the timestamp is because there is no such thing as a universal clock. Different systems will lose time at different rates; some systems may not be getting their clocks updated properly; updates may be coming from different locations with their own offsets... Invalid local clocks are one of the main contributors to "SSL Certificate Invalid" browser errors (per data collected by Chrome, presented at RWC 2016).

Zcash has not done away with nodes. It is identical to Bitcoin in many respects, and certainly with regard to the network architecture. What has been altered (wrt this thread) is the proof-of-work algorithm, which influences what types of nodes are more likely to participate (or more precisely, what hardware the mining will run on). Changing the PoW doesn't change the way that transactions propagate through the network or are stored in mempools.


#18

Aha. That's what I thought when I started this thread - https://forum.zcashcommunity.com/t/should-node-operators-be-rewarded/668

Who are the current ZCash node operators?


#19

Hmm, there aren't "nodes" in the sense of nodes that participate in the creation of transaction privacy. There are nodes in the same sense as Bitcoin has full nodes, that participate in the blockchain. So at this stage, anyone running zcashd is running a Zcash node. The only notable one I guess is the bootstrapping node listed in the Public Alpha Guide that we run.


#21

Is that what you're hanging around for?