Our tale begins June 2017, when Sia developers announce their intent to ship Obelisk: an ASIC miner of their own production for their network, Siacoin (SC). Two days later lead Sia developer David Vorick (taek) publishes an excellent post outlining why this is a necessary step for the network and their company. After a mixed reaction from the community, taek addresses concerns in a Reddit post.
The plan at this stage is for a one week presale for miners shipping on/by June 30, 2018 (one year from then). The team believes this timeline is conservative, and realistically hope for early 2018. Many are (rightfully) skeptical of Obelisk as crowdfunded ASIC production has historically been riddled with issues (see Butterfly Labs). Presale is extended due to lower than anticipated sales, and ultimately runs to late November.
Obelisk expects to be the first and only ASIC to the Sia network, though they plan for the worst and quietly add the ability to run a proprietary and slightly tweaked version of the blake2b hashing algorithm that the network operates on (often called the ‘knife’).
November 2017 a slide leaks from a Bitmain presentation listing SIA (a common though incorrect abbreviation for Sia or SC) among coins they have market share in. Some in the community are alarmed, but most dismiss the slide or don’t recognize the significance. With the benefit of hindsight it was here that the Obelisk team should have announced the existence of the knife.
January 2018 (and shortly after SC is pumped to an all time high 10 cents) brings news of a new Bitmain Antminer A3 Siacoin ASIC miner, shipping days later. Sia devs are quick to respond to this news, and outline several concerns:
Bitmain is known for overselling their miners, flooding the mining scene and ensuring little profit for anyone but themselves.
Bitmain has historically shown no interest in cooperating with the wishes of coin developers or the community.
Though only glossed over by taek in the post, Obelisk buyers are quick to realize that this puts their yet to be delivered miners in great danger of ever paying themselves off. taek also announces the existence of the knife, though this would require a fork to implement for the Sia network.
The community is now torn between those who think the highly disruptive Bitmain doesn’t belong on the network, and those who think forking them away at this stage would be uncalled for. Within a week, the pro-fork camp publishes an article detailing why they think a fork is the right move, and receive support from many reputable community members.
After two long days, the Sia developers publish their official response reaffirming that unless they are attacked, they don’t support a fork to invalidate Bitmain machines. Tension in the community remains fairly high even after this official response, with some lasting rift becoming apparent. Work quietly begins by community member toaster (Mark) on a fork of Sia.
Though the significance would not be apparent for many months more, it is also around this time that a relatively unknown ASIC manufacturer Halong Mining claims to be close to shipping a Siacoin miner of their own that leaves the Bitmain A3 in the dust.
This unit (branded the DragonMint B52 by Halong) begins shipping early April, however would ultimately appear to be a miner of Innosilicon design (branded the S11 SiaMaster). The relationship between Halong/DragonMint and Innosilicon remains unclear, but the two units are identical in every way except name. I personally run a B52 with the firmware of an S11 without issue.
Revisiting the Obelisk scene, shipping deadlines continue to slide through 2018, even past the advertised June 30th ship date. With mining rewards continuing to shrink and still no solid idea when their miners might arrive, morale in the Obelisk community is very low and talks of law suit are common.
Mark drops the toaster moniker and gives up his initial intent for a launch coinciding Obelisk deliveries. The project announcement drops June 28, and ‘Hyperspace (XSC)’ plans for a late July launch. Hyperspace hopes to focus on marketing and adoption (historically weak points for Sia), and makes several significant changes from the Sia codebase. One such change is the removal of the SiaFund mechanism (a revenue model for the Sia team). The Sia team takes offence to this move, and believe that Hyperspace has shown itself to be unfriendly despite benefiting from several years of the Sia team’s work. It is also worth noting Hyperspace gifted a significant (though timelocked) pre-mine to itself and several community members (primarily from the Sia community) that showed interest in working with Hyperspace. I was one such recipient.
As the summer progresses Sia’s network hashrate continues to grow, with one particular address mining a significant fraction of the whole network. It is believed this mining farm is operated by Innosilicon itself, given that the address being mined to matches the default address on the S11 miner firmware and appears in an early Halong demo (which also further mystifies the relationship between Innosilicon and Halong). This address becomes important later.
Late July Hyperspace launches, and Sia community member hakkane (also author of SiaStats.info) publishes an excellent investigation in to Bitmain’s early A3 operation, proving that Bitmain self mined enough blocks to cover a significant chunk of the production cost for the A3. At this stage it is thought that B52s/S11s make up a vast majority of the total hashrate, given that one S11 (4.3 TH/s) is equivalent to five A3s (0.815 TH/s) despite using the same power (~1350 W) and that margins for even the greatly more efficient S11 are becoming thin.
August 2018 user FaustianAGI, a prominent voice in the January fork war, authors the ‘Community Proposal’ where he suggests a series of changes that are aimed at placing the Sia community and team in a stronger position after a year of continual beatings. This includes utilizing the knife to fork to an Obelisk only chain for at least some period of time (and ideally absolving Obelisk of pending law suits which could place the Sia team in serious financial danger), and a developer fee on the block reward. The proposal receives strong community support.
It is also mid-late August that in small numbers the first Obelisk miners begin shipping.
And now we can cover arguably the most significant decision made since Sia’s launch four years ago. Referring to the alarmingly high fraction of the total hashpower that single Innosilicon address holds (now nearing 50% of the network), on October 1st, 2018 the Sia team announced that they intended to hardfork at block 179,000 (anticipated October 31st) to an Obelisk only chain. Though citing FaustianAGI as inspiration for the fork, Faustian would later express his dissatisfaction with the move, feeling that the fork the Sia team ultimately settled on missed the point of his original Community Proposal.
October 13th ‘SiaClassic (SCC/SCA)’ emerges as a contentious fork, with immediate support from now controversial Innosilicon. It is quickly apparent to the Sia community that no one they recognize is involved with SiaClassic, and it fails to gain any amount of organic support.
Now with two adversarial forks competing with Sia, and Sia still lacking the strength Faustian hoped it would take from the Community Proposal, Faustian and several prominent Sia community members announce ‘SiaPrime (SCP)’ on October 15. Aimed at addressing many of the issues (marketing, real world adoption) that the competing forks are focused on, SiaPrime hopes to lead by example in maintaining a good relationship with the Sia team. taek expresses his initial approval, though many in the community criticize the further splintering of Sia and generally remain skeptical.
October 31st comes and goes with only a relatively minor hiccup, and its largely back to grinding as people either accept the fork or move to other projects.
SiaClassic becomes more obvious as an astroturfed effort with a mission to create chaos and allow for Innosilicon and Bitmain to sell what remaining blake2b miner stock they have. Hyperspace continues until Christmas when development ceases, and the project leaders almost completely disappear. Cursed with being last to the party SiaPrime faces growing pains and skepticism from Sia OGs, but remains the only fork continuing development as of June 2019.
So where are we now?
Sia remains on its mission, and is presently hiring, which can only be a sign of strength; network utilization has also nearly doubled in the last two months alone. Many in the community feel Sia lost precious time because of the ASIC wars, but the popular sentiment is just being glad its over. There remains uncertainty over the future of Sia’s mining ecosystem; with the block reward now decreasing rapidly (over 1% a week and rising) its becoming less economically feasible for a manufacturer to produce a new generation of ASICs for Sia, regardless of community support.
Though I’ve been on the other side of some of these debates, I commend taek and the Sia guys for sticking to their guns and making decisions tougher than I could imagine. I look forward to the future of the Sia ecosystem.
Thanks for reading; I’m open to writing more about these exciting past 24 months, so let me know what you’d like to see! Cheers, — MasterHW#3493
Disclosure: the author holds SC, XSC, SCP, and is actively involved in the SiaPrime project.
