EOS Gets Congested, CPU Doubled – Trustnodes

EOS Gets Congested, CPU Doubled


Eos has been facing constraints on resources with their network becoming highly congested just months after they launched. An eos user said:

“I have about 80 eos in one account with 37 staked for CPU and 36.5 staked for net and 8 liquid to play with apps. I’ve only played about 10 rolls of dice and did two days of mining grandpa coins. Does it make sense that I’m using 100% of CPU and 5452 of 5474 bytes of ram? Net is using nothing.”

Net being bandwidth. The eoser thus has locked $430 worth of eos to move around just $43 worth, hitting a CPU capacity limit. A Block Producer (BP) says:

“As soon as a block hits 10% of the total 200ms of CPU time allowed per block, the CPU allocation algorithm switches to the congestion mode…

This proposal aims to change the current value for target_block_cpu_usage_pct from 1000 (10%) to 2000 (20%).”

In other word, the eoser can now play around with $86 worth of eos, rather than just $43. If that is correct, there’s a max ceiling here whereby one has to lock $400 worth of eos in order to use another $400 worth.

That suggests, in a very simplified manner, that if one has $800, then only half of it can be used even if the CPU capacity is increased by BPs to 100%.

EOS stats, October 2018.

That would explain the very high level of “staking” in eos whereby more than half of the current 906 million total eos is being “staked.”

The “stakers” receive no “interest” or “dividends.” Only the colluding 21 BP admin nodes receive more eos for staked eos. The rest just get “free” transactions at the cost of being unable to use currently some 80% of their funds.

Eos’ circulating supply, therefore, is more about 350 million eos, than nearly a billion.

These are apparently leading to transaction numbers of more than two million a day and according to the stats on the right, the network is currently processing 40 transactions a second.

Eth, eos transaction numbers, October 2018.

According to the above data from Coinmetrics, eth has seen somewhat stable transaction volumes for much of this year.

For eos, they quickly jumped in July, just a month after it launched. Then they spike to nine million, which is obviously presumably fake, with it diving recently now to almost below eth.

When we look at other stats, however, there’s a very different story. Eos transfers measured in USD, for example, amount to just about 5% of eth’s:

Eth, eos usd transfers, October 2018.

What is far more striking is perhaps the best indicative stat, active addresses. There, eos’ network usage is at just 1.4% of eth:

Eth, eos, active addresses, October 2018.

As you can see, there were just 4,000 active addresses in eos this Wednesday, while eth has 283,000. Eth’s active usage, moreover, is somewhat stable while for eos it wildly fluctuates.

Meaning that eos is obviously a far smaller network compared to eth, with it apparently having a 1MB blocksize limit.

We tried to find out the size in bytes of a current block, but they don’t show it. The idea in eos however is to have big datacentre nodes with users not able to validate the network for themselves, so having to rely on the colluding BP admin nodes to tell them just how much is the actual supply, if there is an invalid block and so on.

If they print out a billion or a trillion eos, it wouldn’t be easy to find out until it is far too late through indicators like price tanking.

That’s because unlike say in ethereum where miners can’t change the rules because there are some 16,000 eth admin nodes, in eos the BPs can do what they want. They can be voted out by “stakers,” but most of them keep their eos on Huobi or Bitfinex, which are BPs, so can outvote them.

Despite this almost complete tradeoff, eos is not even able to function without congestion now and then. First ram, now CPU, and who knows what next.

Copyrights Trustnodes.com


Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>