How to revive an old HP Proliant G6 server for Data Mining, Seed Node and Witness node.

Hi there, Steem!

It all had begun during the dotcom boom, when young and ambitious start-ups realized that to create an additional cost on top of quasi-decentralized internet, it is necessary to provide services that will collect or receive this data, and also store and give access to it.

Not for free, silly you, for money, of course. Your money!

Well, we've faced a major breakthrough technology of the first web hosting services - accessibility and massiveness had to come. Now you do not need to buy a server, it does not take up space anymore, it does not make any noise in your house. And then came a legion of similar services, and the world wide web and its data made the first step towards centralization.

Services did not stand still, they continued to grow and expand, and gladly accepted the data into their centralized hands. All this began to form a giant data warehouses of diverse data: Google, Amazon, Facebook, Instagram and the others. All requests from you go to their servers, which in one way or another are centralized. And it's very bad. I mean - very bad.

Why is it bad,duh, you may ask? I use these services every day, they are so cool and useful bifidobacterium, I can post my photos, chat with friends...

Hell, no! Data centralization is not good for at least 3 reasons:

1. We are giving them too much power ===> CENTRALIZATION!

   Your favorite Instagram or Facebook servers are offline due to unpredictable hacker attack. Yeah, suddenly you may feel emptiness inside. lol

2. Monopolization of the market ===> NO COMPETITION!

   It's difficult to be a competitor to giants, and competition is good, it works. That is, this market becomes very difficult to monetize and earn on. Did you ever think to run a cloud service in the states? How about a hosting service? Maybe distributed or decentralized ones? lol xD

3. Dirty deeds done dirt cheap ===> WE NEED ALL YOUR DATA (AND WE WON'T PAY YA FOR THAT!)

Have you ever wondered what information about you is stored by the IT-giants?
Why everyone keeps quite about Uber's Godmode? Same as Google location logging))) Photos, music preferences, any activity ... In general, this is not good, if you think about it. They collect such tremendous amounts of data, and in fact ready-to-use data sets, and pass through their top notch neural networks for absolutely implicit purposes.

Evil corp knows that the Data is the new gold!

Data creation is exploding! Selfies and useless files people refuse to delete in the cloud... 92% of the worlds data was created in the last 2 years alone. At the current rate the world's data storage capacity will be overtaken by next spring. This will be catastrophe. Data shortages, data black markets. Someone's compression algorithm will save the world from "Datageddon", and it sure is gonna be Nucleus and not Goddamn Pied Piper. I don't know how about you, people, but I don't want to live in a world where someone else makes the world a better place better than we do.

Gavin f*****ing Belson, CEO of Hooli, (sharp quote from Silicon Valley show).

---

Witnessing preparations

INTRO

Ok now, STEEM is facing a tremendous user base growth. There is registration bottleneck and process is broken which is kinda sad, but I'll keep believing and seeing bright future.

say that there is going to be some kind of account mining. The suggestion is to burn STEEM (and remove it from the supply) during account creation and credit it with permanent bandwidth.

STEEM technology is a very disruptive one and IT WORKS RIGHT HERE AND RIGHT NOW! It's not flawless, we have a long road to go! Not like some shitty ICO's. Give us ETH, we'll give you back some tokens. Not now, silly you, in a year!

**IMO, STEEM-based apps have more future that STEEMIT. No one really needs another one social network (and steemit is kinda quasi-network more of a blogging network), but I think that @dtube and @dlive have the brightest one, as a author-to-audience content delivery platforms with no intermediaries. Youtube does a good job to demonetize creators and they eventually move to the STEEM. I'll keep my finger crossed. **

BUT LETS KEEP IT REAL, STEEM IS NOT READY FOR GROWTH!!!!

So, STEEM, it looks like we are in desperate need of stable infrastructure system to prepare for this parabolic growth curve.

---

WHY THE HELL SOMEONE NEEDS A DEDICATED SERVER IN HIS HOUSE? DATA-MINING & WITNESSING!

Me and my friend been hanging around since school times and recently decided to run a small business in developing automated trading software and trading algorithms. One the main features of our product is an AI-based market analytical system/asset management, which analyzes chart data, social networks sentiment, influencers - all types of the data related to a specific market.

Despite our main market to work with is the commodities market, it is easy to implement all the stuff to the crypto sphere (kudos to exchange APIs developers). And especially, kudos to Google for TensorFlow, this is a game-changer library.

Recurrent neural networks suit best for chart data but sentiment analysis is the most complicated part so far, because neural networks to analyze texts require a lot of testing and preparation. I mean:

Did you realize how complicated is to explain to the machine what the sentiment is? What is good or bad, what is relevant or irrelevant?

Now our MVP is on the testing phase so we need to gather a lot of data to pack into the right and trustful datasets to train our models.

- AI models training is like this! lol

Recently, we decided to purchase an old enough, but not yet completely outdated HP Proliant DL360G6 server. This obsolete but not yet outdated beast was a backup server in a Russian bank. Its previous owner didn't even manage to erase all the data from hard drives. Thanks to the Lord it was encrypted, and we were too busy to decrypt all the stuff (password could be bruteforced), but still it's weird to sell hardware with bank data on it.

We use VMware to run numerous virtual machines where our robots test strategies and neural nets

Server specs:

Core	2xIntel Xeon 5520 2,3GHz, 8mb cache, 8 cores in total, 16 threads
RAM	64GB DDR3 (192GB max)
Storage	1x500Gb SSD + 2x300 HP HDD

---

Some photos of this beast!

---

Seed node had kick-started easily, but there were some issues to keep it online for a long time simultaneo. But we'll make one - it's just the matter of time and overall we were busy setting up witnessing node.

Despite, we ain't no noobs in Linux Server administration but it was tough, some stuff is really implicit, and inforamtion about it is completely outdated!

STEEM-IN-A-BOX by @someguy123 is outdated AF, so it's necessary to say it doesn't work out-of-the-box. It's ok! We love free education - there is always where to expand your skills! Kudos for @jerrybanfield for aggregating even tiny spits of the valuable guides!

When we'll finish our job - I'll make a very comprehensive guide how to setup a witness node. I promise - someone is got to do it!