/biz/ - Business & Finance

[Anonymous No.12211154]

>https://www.pivotaltracker.com/n/projects/2129823/stories/162537930

They are trying to implement Postgres as a backend for chainlink nodes. This is nice as they had some toy go-lang specific data store before and it made redundancy practically impossible to implement.

Question is, can't they use something like ECTD? This is what kubernetes uses to track it's state in the cluster and it works beautifully.

[Anonymous No.12211168]

>>12211154
>wearing pants

[Anonymous No.12211211]

I dunno I just put my life savings in it

[Anonymous No.12211225]

>>12211154
So are we going to handle sharding using chainlink?

[Anonymous No.12211241]

>>12211168
Based

[Anonymous No.12211283]

>>12211168
Absolute unit

[Anonymous No.12211291]

Postgres is really nice, though, built for supporting redundancy. Between the two, it mostly depends on how the data is structured, and I'm not sure how well etcd would handle the relational type of data that you're plumbing through with Chainlink. (Obviously, you *could* make etcd work, but it's a matter of efficiency.)

>This is what kubernetes uses to track it's state in the cluster
It tracks more slow-moving state, though, doesn't it? Stuff like config management, location/version of running services, etc.

[Anonymous No.12211311]

I dunno, maybe the extraction of Symbian Logical expropriation will integrate veracity for the codilogical aspects of the ChainLink Ecosphere.

However if post-Nubian articulacies are used, that will present MAJOR issues in the pivotal tracker.

Staying tuned.

[Anonymous No.12211339]

>>12211311
O.. O....... One...... One kay eoy

[Anonymous No.12211376]

>>12211291
Seems pretty performant
>https://github.com/etcd-io/etcd/blob/master/Documentation/op-guide/performance.md
>etcd provides stable, sustained high performance. Two factors define performance: latency and throughput. Latency is the time taken to complete an operation. Throughput is the total operations completed within some time period. Usually average latency increases as the overall throughput increases when etcd accepts concurrent client requests. In common cloud environments, like a standard n-4 on Google Compute Engine (GCE) or a comparable machine type on AWS, a three member etcd cluster finishes a request in less than one millisecond under light load, and can complete more than 30,000 requests per second under heavy load.

Why would a chainlink node require SQL?

[Anonymous No.12211382]

>>12211168
based and nopants

[Anonymous No.12211433]

>>12211311
kek'd

[Anonymous No.12211541]

>>12211154
>Question is, can't they use something like ECTD? This is what kubernetes uses to track it's state in the cluster and it works beautifully.


Sorry retarded here, what is ECTD?

[Anonymous No.12211575]

The pivotal discussion centers about problems with locks it seems.
>https://coreos.com/blog/etcd-3.2-announcement
>One advantage of a consistent distributed key value store like etcd is it can coordinate and synchronize distributed systems. Typical primitives for this coordination tend to be distributed shared locks and leadership elections. In etcd 3.2, both distributed shared locks and elections are exported as RPC services, greatly simplifying distributed coordination while also improving performance in high latency environments.

[Anonymous No.12211997]

>>12211575
>>12211376
I'll admit I'm not familiar with Chainlink's technical design, but my thoughts on SQL as opposed to a straight KV store, is that I imagine the data of concern is structured in a multi-dimensional way (referenced by a tuple of data source, address, possibly currency, etc.). While you can always store these as key-value pairs, it's less than optimal for subsets of those dimensions. Now, we're also talking about working with a blockchain, so your local DB throughput is likely not going to be your bottleneck, so "less than optimal" could very well be good enough. As for the issue of locking, again I don't really have the context on how the DB is being used/accessed, but you're right to point out that etcd is pretty much a standard at this point for managing distributed locks and/or leadership.

[Anonymous No.12212079]

>>12211154
This will not be hitting $1000 by EOY

[Anonymous No.12212108]

>>12211283
You have to go back

[Anonymous No.12212190]

>>12211997
Their previous data store they were working with was BoltDB. a embedded key/value database
>https://github.com/boltdb/bolt

Guess they were ok with key/value stores. Perhaps some assumptions changed...

[Anonymous No.12212439]

>>12211154

what a piece of shit. i sold. i have been a software dev now for over 10+ years specialising in blockchain and financial technology (fintech) and i can categorically say that the pivotal tracker is clearly broken. i ran it locally in vmware and the boolean operators are not responding to the actionscripts in the agile enviroment they are set up in (why do it in agile?) the json lookup tables are not parsing and to cap it off, the oracle deference operators are not multi-pass compiling. this is nowhere near completion and judging by the state of the code it wont be until they hire more than 2 devs (that actually know what they are doing). i guess this is what you get when you have a philosophy grad attempting to code software. embarrassing.

[Anonymous No.12212463]

>>12211311

[Anonymous No.12212493]

>>12212190
Ah, I wasn't aware of that. Yeah, if you've been running BoltDB and just want something more "natively" distributed, etcd seems like a natural choice. Who knows what the reason would be, then.

[Anonymous No.12212530]

>>12212493
>Who knows what the reason would be, then.

because they have no idea what they are doing.

[Anonymous No.12212577]

>>12211154
>>12211997
ask them https://gitter.im/smartcontractkit-chainlink/Lobby

[Anonymous No.12212603]

i cant understand a single post here

[Anonymous No.12212642]

they have no idea what they're doing hence why link is written in Go and not something as Rust or C++ and instead of using LMDB the fastest key value store which is exactly what they need they use Postgres now. bunch of morons, dump this shitcoin at $1

[Anonymous No.12212656]

>>12212577
>Last post by thomas at Dec 20 14:20

[Anonymous No.12212696]

>>12211283
Can you say man rocket?

[Anonymous No.12212700]

>>12212642
yes the most demanded devs have no idea what they are doing youre right lets go guys its time we quit

[Anonymous No.12212705]

>>12211154

As far as I know there is no mining in Chainlink tokens, unless you mean SHA256?

[Anonymous No.12212719]

>>12212705
Not talking about mining, just persistent data stores for nodes.

[Anonymous No.12212730]

>>12212719

Yeah I disagree with that usecase

[Anonymous No.12212740]

Well, ORMs cant be programmed as an open enclave so it would have zero advantage of using them but as i can see they are trying to push up and migrate it to JSONB with the parallelism and this is the only way to do it.

Ctrl+f "we could be using a multitude of layers to reduce the amount of code we have to write"

It seems that they are trying to implement closed sourced GORM's as a second layer just to save some weeks of work... and thats not very bullish because they will add a second, unverified code as a layer that can be a huge security breach over unknown code implementation to their core JSONB dynamic extension...

What that mean ? You wont have a huge privacy running your own smartcontract on chainlink network because you still need to provide GORM's code, which use zero encryption as in not very secure end-to-end.

[Anonymous No.12212748]

>>12212740

I agree but I have to disagree with your thesis. It just seems shallow and pedantic. Thanks

[Anonymous No.12212769]

lol link is already displaying technical flaws, link 0.01 eoy.

[Anonymous No.12212898]

>>12211154
They are migrating from BoltDb to Postgres due to this:
>Start-up delay of 2-3min~ due to DB size
https://github.com/smartcontractkit/chainlink/issues/752
and due to BoltDB inability to handle multiple connections.

[Anonymous No.12212954]

>>12212898
>inability to handle multiple connections
lol link already btfo

[Anonymous No.12213045]

I've always wondered how much of Link's supply /biz/ controls. 10-20%? What do you think?

[Anonymous No.12213272]

>>12213045
10% at most i would guess, probably less.
t. 70k linker

[Anonymous No.12213278]

>>12211211
Checked and sane

[Anonymous No.12213287]

>>12213272
damn 70k must feel comfy? are you still accumulating?

[Anonymous No.12213292]

>>12211154
>>12211291
>>12211311
>>12211376
>>12211575
>>12211997
>>12212190
>>12212740
>>12212898
>>12212954
WHAT ARE YOU TALKING ABOUT? Please explain for brainlets...

>>12211168
>>12211211
>>12211283
>>12211382
>>12212603
Look at how many brainlets are here who just want to join in your big-brained conversation. We may not understand but we do care... Please tell me what you're talking about you smart fucks and tell me if it means I really will make it...

[Anonymous No.12213311]

>>12213287
You always want more. I have a little more than 70 and im eyeing 100k. Once i hit 100 ill be eying 150k. I can already feel it. 6 months ago assblaster had 350k.

[Anonymous No.12213366]

>>12213311
check'd

i was thinking of dumping my ZRX stack to add more LINK to hit 70k, but I think ill just start accumulating with fiat and try to hit 100k next
year.

[Anonymous No.12213538]

>>12213292
A chainlink node is an application. Applications usually require a place where one can safely store and fetch their data.
As of now, Chainlink used a store that was very fast and neat, but worked correctly if only one application on one computer used it. If one wanted to have multiple computers serving as a chainlink node to ensure that if one computer shits itself, the whole node would still be able to serve requests, preserving stacked link.
Another data store is currently being proposed that can serve as a store for multiple computers - PostgresSQL. Pivotal tracker suggests that there's a variety of problems with it for the task that it would be serving in a chainlink node.
OP proposed a alternative option, something called ETCD, which is commonly used in similar big boy architectures.

Hope that helps.

[Anonymous No.12213544]

>>12211168
Based and THICC pilled

[Anonymous No.12213581]

>>12213287
yeah i'm still buying as much as possible

[Anonymous No.12213622]

Looks like Thomas is all over etcd now

[Anonymous No.12213647]

>>12213538
based and redpilled

[Anonymous No.12213735]

>>12213538
Based brainlet translator. Fully get it now. Question, for a 1000% brainlet who won't be able to run nodes like (Me) is linkpool the only option? Do you speculate that linkpool will have competitors? Will some brainlet proof platform be made where running a node would be as simple as drag and drop? Alot of questions I know. Please accept this as payment (Pic related). Thanks

[Anonymous No.12213811]

>>12213538
Thank you friend.
Does this mean mainnet is near?

[Anonymous No.12213827]

>>12213811
Given that something like this is a major reimplementation, I would indicate that mainnet isn't anywhere near.

[Anonymous No.12213844]

>>12213735
Thank you for the pic, but I don't know the answers to your questions.
Personally, I don't plan to stake at all.

[Anonymous No.12213847]

>>12213827
Thanks Thomas

[Anonymous No.12213861]

>>12213827
>major reimplementation
well Sergey said it's a matter of months so hopefully by Q2 2019....
still time to accumulate

[Anonymous No.12213862]

>>12213538
these are the kind of posts biz used to have. i miss the old biz.

[Anonymous No.12213926]

From what I gathered from some high level folks who study this as a full time job, they expect 2-3 months left of work.

For perspective, Sergey has been doing this for over 5 years now.

[Anonymous No.12213951]

>>12213844
So you won't be running a node as well? Won't you need to stake link in your nose to get paying contracts to use your node? I know you can run a node with out staking but would that be as profitable?
And what's your reason for not wanting to stake at all? Just curious about your decision.
Thanks

[Anonymous No.12214021]

>>12213735
the whole point of the network is independent nodes being financially motivated to participate in a decentralized oracle network to guarantee the security and fidelity of the data that reaches blockchains; independence of disparate nodes is highest (centralization lowest) when random nobodies can run a node with ease, it seems very likely to me that linkpool will have either many competitors or it will eventually be very easy for brainlets to run nodes without needing to join a pool

[Anonymous No.12214054]

>>12213538
Thank you very much.

[Anonymous No.12214064]

>>12213951
>And what's your reason for not wanting to stake at all?
Don't want more uncertainty about the fate of my link.

[Anonymous No.12214136]

>>12214064
Why not stake a small amount apart from your main stack

[Anonymous No.12214142]

>>12214021
>>12214021
I'm prepared to stop being a brainlet and learn but the fear of fucking up stops me from doing so well because I'm a brainlet.
But hopefully fren because I'd run multiple nodes if based Sergey made it easy to do so

[Anonymous No.12214150]

Project has more devs than ever but I still feel like it's lacking:

Programmers:

Steve Ellis (CTO)
Demetri Roche (software engineer)
Thomas Hodges (integration engineer)
Jonny Huxtable (shit kicker)
John Barker (software engineer)
Alex Kwiatkowski (software engineer)

CEO:
Sergey Nazarov (eating burgers)

[Anonymous No.12214404]

>>12214136
Always an option, but I like the sweet, round number of linkies I have right now.
Also if I start, I'll get obsessive and I'd rather focus on other areas of my life.

[Anonymous No.12214501]

>>12213272
What do guys like you even see here? All this shit is above my head. What’s your 6 month forecast for this?

[Anonymous No.12214607]

>>12214404
How many link tokens do you have?

[Anonymous No.12214624]

>>12214501
Link, if successful, will probably moon around 2020/2021 when bitcoin halves again, ignoring the memes, link was always seen as a 2+ year hold.

[Anonymous No.12214626]

>>12211154
thomas scheduled this suggestion for discussion in pivotaltracker. please tell me this was posted to reddit or sent via some official line of communication, and the team aren't actually browsing biz

[Anonymous No.12214813]

>>12214626
See >>12212577

[Anonymous No.12214973]

>>12214626
Of course they browse. If you were working on world changing economy shifting technology, and a small group of sometimes funny autists kept finding breadcrumbs and holes with your company and team and associates, you too would be intrigued to follow along or check in from time to time.
I’m sure they have experienced everything from entertainment and laughs to shock and disgust, but ultimately find respect for what goes on here over time.

[Anonymous No.12215004]

>>12213735
>>12213735
Binance will almost certainly run nodes. They hold 100 million LINK of their own and of customers. I can't imagine they would not monetize their stack

[Anonymous No.12215013]

>>12214973
i hope we haven't made sergey self conscious whenever he orders a few big macs

[Anonymous No.12215108]

>>12213538
thank you fren

[Anonymous No.12215209]

>>12214021
>>12214142
i'm a brainlet, ready to learn how to run a node. but i'm honestly scared to losing my link somehow. I throw it in a wrong address, or make a stupid mistake that costs me the 5k that i hope to stake.

ill obv start staking 20, 50 or 100 first.

[Anonymous No.12215371]

>>12213287
Yeah he is bec you need 100k link to have a shot at making it

[Anonymous No.12215453]

>>12215371
Maybe, but i think less will do. I'm going for elite status mate

[Anonymous No.12215528]

Bump

[Anonymous No.12215693]

Deep Dive: etcd
>https://www.youtube.com/watch?v=GJqO1TYzVDE