Author Topic: Witness Report for Verbaltech2  (Read 4851 times)

0 Members and 1 Guest are viewing this topic.

Offline Thom

Re: Witness Report for Verbaltech2
« Reply #15 on: June 03, 2019, 07:38:34 pm »
I have published my quarterly witness report on the  Whaleshares platform:

I am the text here as well, since some may not have access to that website:

Welcome to my quarterly witness report.

In this report I'll cover:

    infrastructure / technical topics
    assessment of trends / activities / BitShares politics
    crypto market prices
    status of all nodes

Infrastructure & Technical
All nodes were updated recently to address a vulnerability that could potentially halt the chain. This patch was announced by @abitmore and was applied to all verbaltech2 nodes the same day. Prior to that in this past quarter the new 3.0.0 release was deployed into production as was mentioned in the last update.

I have observed a significant increase in the amount of network activity as reflected in the number of connections reported by each server. This causes an increase of logs, requiring either changing logging rotation parameters or more space. I have added additional drive space to accommodate this increased traffic.

The testnet has also received 2 updates in this past quarter.

It is anticipated all nodes will soon require another update, this time to version 3.1.x.
Assessment of Trends / Activities / BitShares Politics

This last quarter has also seen the launch of BEOS. With the migration of BTS onto that platform to obtain the BEOS rainfall, a substantial amount of voting power has gone with it. This represents a concentration of voting power in fewer hands. Whether you agree with what those hands are doing is separate from the issue of centralization, however as in most everything in life balance is important.

Anyone that understands the premise of PoS and DPoS concensus knows that the power of influence is ultimately determined by how many BTS an entity has, either directly or via a proxy. Control or influence over the direction the platform evolves is not as decentralized as the number of accounts. If that were the case AND those accounts participated in voting there would be far less centralization. Voter apathy and non-participation pushed the developers to implement proxy voting to mitigate that, so the platform could evolve and move forward, but at the expense of centralization.

Until BEOS, the strongest influence has been with proxies from the eastern cultures. Now it has swung back to those in the west who were instrumental in setting the vision and operating principles of BitShares, based on free markets and Austrian economics. I see this as a balancing action. I don't necessarily see it as more centralized, but simply a shift in the balance of power.

Price Feeds
I commented in the last report that the MCR bug fix has finally been released, thus it is now possible to implement algorithms to dynamically change the MCR. When BSIP42 was put into production the implementors chose to manipulate the feed price rather than wait for the MCR fix. BSIP42 was a disaster as a result, and split the community due to it's effects, and caused a black swan on BitUSD.

Despite the availability of the MCR fix, I have seen no effort to refine the BSIP42 algorithm to eliminate the price manipulation and replace it with dynamically changing MCR as a control mechanism to tighten BitAsset pegs. Why not? If you are aware of any please let me know in the comments.

Instead we see complaints against individual witnesses, that claim their feed is causing a significant (negative) influence on the price of CNY. That is simply not the case. The "official" price recognized published for a BitAsset in BitShares is the median price price published from all active witnesses. Using the median price guarantees no fringe or outlier price affects the median price. Verbaltech2 monitors his feed prices with RoelandP's witness log, which is a wonderful contribution that provides quality feedback about witnesses, including price feeds. It is the most objective tool available to evaluate witnesses.

I looked into exactly how the "median" calculation is made, and it is far simpler than I previously thought. Since the number of witnesses is always odd, it's easy to calculate (Note: not all witnesses publish feeds for all BitAssets, so the 2 middle values are averaged to define the median when the number of prices are an even number).

The median or "middle" is based on ordering all the values from low to high. The median value is the one with an equal number of values less that it and greater than it. In other words, in the middle of all values literally. Changing the magnitude of the sample at each end of the scale has zero effect on the median value, none, zilch.

There are many factors that affect a witness' feed price.Among them are which sources does the witness use, how often are they sampled, and the volatility of the market. All sources of prices are not available to all witnesses. Diversity is important in arriving at price that truly represents the market for that asset.

It appears to me that a lynch mob mentality pervades the assessment of witness price feeds, and ignores the careful design intentions built into the platform. This mentality doesn't serve the diverse needs of a world market but rather coerces conformity based on biases and agendas that don't acknowledge free trade and the difficulty imposed on witnesses to produce a feed with the sources he has access to.

I for one am glad for the BEOS voting influence (now that they have stabilized their voting) that recognizes these considerations and counter-acts this witch hunt mentality that targets witnesses based on only one narrow definition (I haven't even mentioned other very important criteria a witness should be measured by) of performance, that being price feeds.

That's all for now, but wow! This quarter has seen the end of the current bear market. Whew who! Up up and away! Here is the market summary since last update (all prices via CMC): BTC=$8,584.53, BTS=$0.067, STEEM=$0.399, EOS=$7.26, PPY=$0.904. I'll list WLS when CMC does.

Amsterdam


Gemany, node 1

Germany node 2

Singapore

And the testnet server for BTS:

Your vote for witness
Verbaltech2
is greatly appreciated! Thanks for your time and attention

Injustice anywhere is a threat to justice everywhere - MLK |  Verbaltech2 Witness Reports: https://bitsharestalk.org/index.php/topic,23902.0.html

Offline Thom

Re: Witness Report for Verbaltech2
« Reply #16 on: September 03, 2019, 06:02:57 pm »
Welcome to my quarterly witness report. For node performance graphs please see them posted on whaleshares.

In this report I'll cover:
- infrastructure / technical topics
- assessment of trends / activities / BitShares politics
- crypto market prices
x status of all nodes (please see whaleshares link above for performance graphs)

Infrastructure & Technical

In July I started to update one of the nodes to the 3.2.0 release, however that was delayed until after 3.2.1 was released and no other witnesses reported issues. The testnet was also updated in this past quarter. I just now noticed version 3.3.0 was released 10 hours ago, so I will be updating all of the nodes again in the next couple of days.

The amount of network traffic continues at it's elevated level as I last reported, though I haven't seen a significant increase since the elevated levels were first noticed.

I have also disabled bts_tools price feeds and I now use Zapata's price feed script based on Xeroc's "BitShares Pricefeed". More on this topic below.

Assessment of Trends / Activities / BitShares Politics

This last quarter saw the conclusion of BEOS "rainfall" and the decline of the BEOS voting strength as new depositors declined and others withdrew their BTS. I couldn't possibly quantify the extent of the decline, but I did see it reflected in the voting. The next phase of BEOS is called the "Manna" phase, and is yet another form of sharedrop or incentive to participate. [See Stan's blog](https://www.google.com/url?client=internal-uds-cse&cx=011445933062884569876:-jmpkfnkicw&q=https://steemit.com/bitshares/%40stan/beos-spacedrop-manna-from-heaven&sa=U&ved=2ahUKEwiAvIqSxLPkAhUhnq0KHWTOAIEQFjACegQIDhAB&usg=AOvVaw1GBIpyCrMf-SRq1zXqmZhb) for details.

A very serious and disturbing trend in voting is occurring right now. It looks like a major effort is underway to either 1) destroy the BitShares platform by crashing it's price, or 2) dramatically drop the price so more BTS can be bought cheap. Almost all workers have been voted out, core being an exception thankfully, and the refund400K worker suddenly voted up to increase the worker funding threshold thus cutting off workers. Even the UI worker was voted out. This just shows how far short-term thinking has come with these whales that don't care about BitShares beyond what short-term profits can be squeezed out of it. How can BitShares possibly remain a viable project if funding for work is cut off?

Until BEOS, the strongest influence has been with proxies from the eastern cultures. With BEOS power  briefly swung back to those in the west who were instrumental in setting the vision and operating principles of BitShares, based on free markets and Austrian economics. Unfortunately that balancing influence is not as effective as it once was.

Price Feeds

I stopped feeding BitCNY and BitUSD (but continued all other feeds) in protest to the price manipulation and continual reduction of collateral requirements, at a time when they should be increased due to higher levels of market volatility. At the same time I began to dig into the bts_tools feed code, and made many improvements. However I also looked into Zapata's feed script and made the judgment it was superior to the feeds code in bts_tools, for several reasons, the main one being Zapata continues to maintain it. I still use bts_tools to provide a web interface and monitoring of all my nodes.

I need to investigate how to eliminate Zapata's script hanging, waiting for user confirmation when the market causes the price to swing to much since last publication. He has 3 settings that affect that but I never want the script to wait for user input. Rather, it should just skip publishing the feed if the markets are so volatile, no reason to stop all price feeds b/c 1 has changed too much.

I wrote a small Python script to run Zapata's script 3 times, 1) main set of assets and currencies, 2) HERO price feed, 3) HERTZ price feed. I pass the config.yaml file specific to each run as a command line option. The Python program that runs each feed is just a simple scheduler that waits a specific period of time between each run. In retrospect cron might have been a better choice to schedule feed runs, b/c if one instance hangs it wouldn't prevent subsequent runs, well at least until memory is exhaused by a bunch of zombie processes waiting for input that never comes.

I also run the feed script on 2 different servers with a 3rd feed server in reserve if necessary. This approach provides fault tolerance and redundancy as well as provide frequent price updates and spacing around the clock. There was a time when witness verbaltech2 had the most operations, or at least in the top 3 accounts with the highest number of operations, and this was due to how often I updated my feeds. I will monitor the operations count to see if that still holds true using Zapata's feed script. If it does that reflects it takes lots of operations to publish feeds. If not could be Zapata found a more efficient way to get the job done.

Despite those at the governance helm, BitShares has a shit-ton going for it. It is a very mature project with a huge list of features, including HTLC. I was surprised when I read this review of BItShares, not so much because of specific content, but rather due to how positive, comprehensive and accurate the report was. Can't find the link in Telegram right now, but I believe Alfredo has added a new HTLC tab to his explorer. Keep an eye open for it. There's a lot more HTLC activity going on right now than I thought there would be for a feature without a GUI yet. The activity I observed may be only testers, I don't know.

Cryto Market Price Update
Full stop. So much for the "end of the bear market" I mentioned last month. It was a short lived upspike, and the bear continues its rampage. Here is the market summary since last update (all prices via CMC): BTC=$10,364.82, BTS=$0.0334, STEEM=$0.165, EOS=$3.33, PPY=$0.822. I'll list WLS when CMC does.
Injustice anywhere is a threat to justice everywhere - MLK |  Verbaltech2 Witness Reports: https://bitsharestalk.org/index.php/topic,23902.0.html

Offline R

  • Hero Member
  • *****
  • Posts: 733
    • View Profile
Re: Witness Report for Verbaltech2
« Reply #17 on: September 03, 2019, 06:11:52 pm »
Thanks for providing price feeds for Hertz 👍