Why not using the "Previous generation" instances with local ephemeral storage for the blockchain?

m1.small has 160GB that is doubly free (no per GB charges and no per IOPS charges)
m1.medium has 410GB doubly-free

They are both cheaper than m3.medium.

So what is the point of forcing everyone to double pay for EBS (per GB and per IOPS) to primarily store a copy of the shared blockchain which will be exactly the same for everyone (if everything works right)?

I'm really curious what is the technical point of all this rigmarole, aside from shilling for AWS EC2 and earning sales commissions?

Edit:

I just did a quick refresh of my AWS knowledge, and it seems like the my long-time observations still hold:

1) Bitcoin blockchain should be a shared, no-per-user charge, dataset hosted on S3 (one copy per region, 3 to cover USA and 10 to cover the whole world). You'll need to come up with some schedule on how to update them, e.g. one region update per day. This way the blockchain backup will be no more than 10 days old (if you cover the whole world) or 3 days old (if you just target USA).

2) rapidly restore the blocks and chainstate directory from that shared set before starting the bitcoind for the first time in the particular instance.

3) This should save money for the users, I don't know what will be the effect for the commission for the shill.

4) One can recognize good Amazon shills by how they provide a really differentiated, high-level service instead of relying on brute force multiplication of charges for storing duplicate data in each client's private storage.

I think hdtune only does read test, not the read-modify-write tests.

For CPU & RAM testing the best is "Torture test" in the Mersenne primes search program prime95.exe.

I also forgot the standard Windows database killers: online antivirus needs to be disabled.

Bitcoin-Core during initial synchronization is quite well optimized and really stresses the machine more than 95% of programs we've tested. So your "rock solid" doesn't mean much unless you've run something in the upper 5% of optimization & efficiency.
 

You have bad hardware: either SSD, SATA controller or DRAM memory.

Go get a Bittorrent client and download a bunch of multi-gigabyte files, maybe even the Bitcoin torrent. After download force the recheck of downloaded files multiple times in your Bittorrent client.

If you are moderately knowledgeable Windows user download some big files from Microsoft (e.g. Windows evaluation disk images) and multiple times verify SHA1 checksums that Microsoft supplies.

I'm betting on bad SSD, Bitcoin-Core uses write-ahead-logging databases that are pessimal case for SSD usage: lots of very short writes and no reads of the written logging data, which causes humongous write amplification.

Is any other database engine working correctly on your machine, e.g. MS-SQL or Oracle?

I'm sorry I don't have time to write a longer expose.

Any voting method that doesn't explicitly preserve privacy of the individual vote and doesn't preserve the requirement that the vote is cast publicly while observers can look will be subverted by forcing voters to cast the votes against their will and conscience.

Fascists (and ultra-nationalists) were already doing this in Europe's history.

Even currently underground white-power/neo-Nazi groups are privately running campaigns to force susceptible voters to use absentee ballots to vote for them. Their ideal voting setup will involve voting with something similar to the TV decryption card, where each voter will be doing the voting publicly in the pubs/bars frequented by their peers or the device will be simply left at the party headquarters where the appropriate functionaries will cast the votes along the party line.

I found it very worrisome that the current public's love affair with electronic gadgets make them completely forget the lessons from the past:

1) voting has to be done in public where anyone can see that the voter isn't forced
2) voting has to be secret in the sense that the observers cannot tell each individual's vote, only the aggregate statistic.

If you really don't understand the above, you are either:

1) in collusion with some vote-rigging organization
2) an idiot in the original Greek meaning of that word: person who isn't concerned with public affairs, only the private ones.

Regretfully the love of gadgetry runs very deeply in some Bitcoiners, here's the thread from 2013 that similarly proposed Trezor to be used in easily rigged voting.

https://bitcointalk.org/index.php?topic=138174.0

You should stop yourself.

Your idea is easily subvertable and will produce fascist dictatorship if implemented.

Fascists always propose open, visible ticket elections to exert the pressure on voters to vote "right, patriotic". With your in-app voting they will have even easier task: the bosses will ask every employee individually into their room and will watch them while they "correctly" vote or give the employee a pink slip.

Did you even pass a simple test at the end of a civic class? Do you know why elections provide privacy booths for vote casting/marking?

Or are you some retarded kid who flunked their civics? I remember seeing a curious 2nd or 3rd grader ask the teacher "why there are curtains around the desks? don't they make it very easy to cheat?" when seeing voting booth in the school.

So, are you a kid younger than a normal voting age?

Speedtest (and other similar sites) are primarily marketing tools. They are indeed applicable if somebody does non-p2p things like:

• streaming video/audio
• downloading/uploading files
• multiplayer gaming on a centralized server

If you really want to advance Bitcoin (or other peer-2-peer technologies) you'll have to learn to filter the marketing misinformation and simplifications.

In particular, I understand that speedtest was explicitly designed using Flash application such that the user experience precludes you from triggering out of powerboost. It intentionally slows down your interaction by displaying complex screens and prompting you to inform others about your results. It makes it really hard to reliably repeat the test until stable results are achieved.

The only true peer-2-peer results that I have ever seen were collected using Bittorrent. In the past I used to do a lot of torrenting, now I just occasionally help people with their problems.

The true information about the performance of consumer broadband is extremely proprietary. Even the ISP's employees don't have access to the detailed technical information. There are two reasons:

• obvious one: it is a source of competitive advantage and allows them to play various games with customers and competitors
• non-obvious one: modern high-performance hardware-accelerated routers used by ISPs are actually layer-3 switches that are dynamically reprogrammable per packet flow. This is rather fiendishly complicated technology, even the professional networking engineers are surprised by their limitations

See for yourself: lookup on the web discussions about Cisco IOS commands e.g. "show ip cache flow" or "show ip cef" and see when you start losing the plot.

I've recently moved and I'm no longer directly involved in infrastructure details. But from few years ago I remember having a discussion with some students using Bittorrent and other P2P tools. Quite common observation was that overall stable bandwidth can be significantly increased by using a SOCKS proxy server, or trivial (non-encrypting) VPN connection (PPTP, L2TP, etc.) to a machine in some data center or even a dorm room. The consumer ISP then sees only a single pair of packet flows which they can handle more efficiently on their equipment.

The data are far from "excellent". They are mostly bullshit numbers measured using short-term bursts. I checked several markets that I'm very familiar with and those all were successfully gamed by the DOCSIS cable providers using the equipment with "powerboost" (or similar marketing names).

"Powerboost" is a ultra-short-term (few to few-teen or sometimes few-ty seconds) bandwidth increase made available to the modems of the customers that haven't maxed out their bandwidth in previous minutes.

The configuration details are highly proprietary and vary by market and by time of day&week. But the overall effect is that that the DOCSIS modem seriously approaches 100Mbps LAN performance for a few packets in bursts.

On some markets that I know the VDSL2 competitors (that optimize average bandwidth over periods of weeks) didn't even rank on the "TOP ISPS". In reality of the non-burst-y loads the VDSL2 providers outperform DOCSIS providers, especially on the upload side as the VDSL2 technology is a fundamentally symmetric technology that's being sold as asymmetric only for market-segmenting reasons.

I'm not even going to delve into further restrictions on consumer broadband where the providers explicitly limit number of packet flows that can be handled by the customer's equipment. OOKLA (and almost everyone else) measures a 2-flow single-tcp connections, which have really nothing in common with peer-2-peer technologies like Bitcoin or Bittorrent.

Executive summary:

Bullshit marketing numbers, divide by 3-5-10 to get real number achievable with P2P technologies and continuous operation.

Well, I proposed and published the idea of "digital prospectus" and its "amendments" here on this forum late in 2011.

https://bitcointalk.org/index.php?topic=54382.0

I wore this in my forum signature since then.

As one would expect, the developers actually don't want to be held to their word of "not changing" or "changing only for better" when there are means to actually set those promises in the "cryptographic stone" as opposed to the present state-of-the-art "pinky swear".

It is up to investors to force the developers to behave, not vice-versa.

I guess I miscommunicated.

Speculators aren't users. In fact they are somewhat anti-users. They only hold, never spend. In the classical capital flow diagram they would be the ones who have exchange accounts that are not set up "for delivery" only "for contract settlement".

In the market where speculation accounts for anywhere between 90% and 99% of the economic activity omitting that reservoir of capital is either pointless or an attempt to deceive/falsify. Trying to track the fees (which are less than a fraction of percent or ever per mille) and disregarding the futures market seems like a weird joke or deep misunderstanding.

You could include the  https://en.wikipedia.org/wiki/Black%E2%80%93Scholes_model somewhere in your diagram. It could make your model both more powerful and more dissertable.

I also can't happen to note that you are shy of your alma mater and seem to be unclear what it means to defend your thesis in public. It is not the same as being defensive.

Your model omits the speculators and the speculation on the future prices of Bitcoin, on this forum known as HODL-ers, from the misspelling of HOLD.

Therefore I would say it is nearly worthless.

Which school approved this as a master thesis in a computing-related field? This looks like some sort of rather weak economic sciences thesis.

I think he's angling to get the "I trashed Satoshi's resume" credential. Who would settle for somebody with C++ & Windows skillset, when there are people with "(Ruby [on Rails] or Java) for web. Preferably both." ?

Sorry, man! I'm really not up to speed on the SSD sales lingo. It used to be that you wanted SLC instead of MLC, but now they sell eMLC (a.k.a. enterprise MLC) and TLC (8-level cell, don't ask me why they call it "triple").

You want to look for benchmarks showing performance with database workloads involving write-ahead-logging.

The corporation I work for experienced such a large failure rate of the "enterprise class" SSD drives that they kissed up and pulled back from retirement the old, multi-spindle, Fiber-Channel-attached RAID/JBOD boxes with battery-backed caches. They use SSD only for the developer's laptops.

I'm sorry. I'm having problems posting reliably now.

The goal shouldn't be "impossible to make an ASIC". The goal should be "an ASIC should be a substantial fraction of a general-purpose CPU", therefore the cost of developing a competitive ASIC should be a substantial fraction of developing a competitive CPU.

It isn't trivial but it also isn't a completely uncharted territory of research.

The general technical methodology of semi-randomly changing algorithms to optimize certain figure of merit is called "genetic programming".

The point is not to make it impossible to develop an ASIC. The goal needs to be that an ASIC should be a substantial fraction of a general-purpose CPU, such that the cost of developing a competitive implementation should be a substantial fraction of developing a competitive CPU.

I know it isn't trivial, but I also know that it isn't impossible.

The required technical framework is called "genetic programming". It provides for the development of nearly-randomly evolving of algorithms to optimize some figure of merit.

If somebody is really interested in application of genetic programming for the evolution of algorithm professor John Koza from Stanford has a really good textbooks on that subject.

If somebody has approx. 9GB of server space available I could upload the ISO DVD images of his 4 lectures on this subject from around the turn of the century. This was a "sales brochure" for his textbooks, I don't think it is copyrighted in any restrictive way. I have very limited Internet connectivity at the moment, I couldn't even seed a torrent reliably. Just please promise that you'll seed it properly afterwards.

The variable portion of the PoW algorithm doesn't have to be secure, one can sandwich the variable portion between a pair SHA-2 for strong cryptographic security. The anti-ASIC meat in between needs to be only sufficiently close-to-bijective and reasonably chaotic. No need to attempt a complete proof of that, just a decent test of in a subspace that there are no fixed-points.

https://en.wikipedia.org/wiki/Fixed-point_theorem

Get a better, non-cheating, SSD drive.

Bitcoin (like most database-heavy) application is nearly exactly pessimal access pattern for a flash disk: it uses write-ahead-logging. This means that it forces relatively short writes (512B or 4kB) to sync up to the media (the erase block on the SSD is probably 256kB) and then never reads back the written log (because the log is only for crashes and rollbacks). This gives you horrible write amplification factors (512 or 64) that really stress up the flash drive (both the storage array and the controller) and almost no reads because you have large RAM cache.

I would kinda bet (a little) that the similar freezes would occur if you run any classical database engine (like Oracle, MSSQL, etc.) with a similar update-heavy workload. I'm not up to speed on SSD drives, but in the past my friends with similar workloads threw away OCZ and purchased Intel.

If you are a senior (over 65 y.o. or over 55 y.o, don't recall) then yes. They have special department/programs to deal with fraud against elderly.

Thanks. I guess it pays to ask the WU agent in person. Such a low fees are nowhere to be found on the Western Union's web site.

So I guess the premium charged by the USPS goes partially to their Postal Inspectors police force that makes it easier to put pressure on the non-responsive people who accepted the money.

Where?

I used the respective websites to estimate $750.00 send from&to California and USPS quoted me $1.65 and WU quoted $12.50 (for cash in-person service) to $75.00 (debit via mobile app). I know that WU has negotiated lower rates with some agents, so can you tell us (generally) who's giving you lower fees.

Lines at the Post Office? Where? Every one I've seen is nearly empty and it is more like lines of underemployed postal clerks standing in front of the office and trying to attract someone to come inside and mail something.

Unfortunately I know nearly nothing about Western Union besides the fact that they are by far the most expensive.

Did you send them postal money order or a private company money order? Because if you used USPS paper then all the "mail fraud statues" are on your side and it is sufficient to contact US Postal Inspectors to initiate investigation against them. No need to go through stupid arbitration or other customer service bullshit.

Only USPS paper has this enhancement, not Western Union, MoneyGram, or other private providers. That's why all the fraudsters encourage people to use FedEx, UPS, etc.