Rollup Sequencers Are Centralized — And That’s Fine

Cen­tral­iza­tion: the hideous ene­my of free­dom and progress in the realm of dis­trib­uted ledger tech­nol­o­gy. It often rears its head as soon as devel­op­ers encounter scal­ing challenges.

In decen­tral­ized pro­to­cols, the fastest way to move from point A to point B often iron­i­cal­ly involves resort­ing to some sort of cen­tral­ized mech­a­nism. For­get ideals like cen­sor­ship-resis­tance and inde­pen­dence, devs might cry out, we just want this thing to be quick and cheap!

The quest for fur­ther decen­tral­iza­tion in the blockchain space con­tin­ues, but for some ele­ments, Stephane Gos­selin says, cen­tral­iza­tion might not be such a bad thing, after all.

The for­mer co-founder and chief archi­tect of Flash­bots and founder of Fron­tier Research spoke to Block­works on the Bell Curve pod­cast about layer‑2 rollups and how cen­tral­ized sequencers might not be the prob­lem that many fear.

All rollup sequencers are centralized

Let’s get one fact out of the way to begin with: All layer‑2 rollups on Ethereum — every sin­gle one of them — use cen­tral­ized sequencers. 

The sequencer’s job is to process and order trans­ac­tions into blocks to be added to the chain. It’s cheap­er, faster and eas­i­er for rollup providers to main­tain their own pro­pri­etary cen­tral­ized sequencer sys­tem than to farm out the job.

“I’m still not con­vinced that that’s a bad thing,” says Gos­selin, “I don’t think it’s a done deal to say that, actu­al­ly, first in, first out sequencers on a layer‑2 are a bad thing.”

The usu­al argu­ment against rollup cen­tral­iza­tion, Gos­selin says, is that it cre­ates a “laten­cy game” that draws cen­tral­iza­tion toward a spe­cif­ic geo­graph­ic region. Being con­cen­trat­ed in a par­tic­u­lar place leaves a rollup sus­cep­ti­ble to cen­sor­ship and oppres­sive reg­u­la­tion wher­ev­er the rollup is deployed, Gos­selin says.

“But, still, the ques­tion is like, is that actu­al­ly bad?”

Ethereum has been designed, Gos­selin says, as a max­i­mal­ly decen­tral­ized layer‑1 with rel­a­tive­ly lit­tle eco­nom­ic activ­i­ty on the base lay­er. Its goal is to set­tle data with­out what he describes as “con­tention”   — demand for set­tling on a spe­cif­ic posi­tion — which takes place inside lay­er-2s instead.

“If you have an archi­tec­ture where the layer‑1 is only set­tling data blobs and there isn’t con­tention, and you have all the activ­i­ty inside of lay­er-2s, it reduces the cen­tral­iza­tion pres­sure on the layer‑1 significantly.”

Cross-chain messaging to the rescue

Cross-chain mes­sag­ing could save the day, Gos­selin says, pro­vid­ing cen­sor­ship-resis­tance between lay­ers when need­ed. “You have some way to push mes­sages out from the layer‑2 back into the layer‑1, or maybe for it to be inter­pret­ed by some oth­er spin-up of that layer‑2 some­where else.”

Through a mes­sag­ing mechan­ic like IBC, Gos­selin says lay­er-2s would remain cen­sor­ship-resis­tant and non-cus­to­di­al because indi­vid­ual rollup par­tic­i­pants can “exit their state and bridge it over to some oth­er roll-up in some oth­er jurisdiction.”

Host Mike Ippoli­to points out that in such a sit­u­a­tion, users would expe­ri­ence sig­nif­i­cant “mar­ket disruption.”

“There would be a peri­od of time where we’d have to migrate the assets and every­thing down into the main chain and back up onto anoth­er rollup.”

The loom­ing threat of dis­rup­tion, Ippoli­to says, could “pre­vent TVL and activ­i­ty from migrat­ing up to the rollups as much as they oth­er­wise would.”

Gos­selin con­curs, not­ing, “the oth­er argu­ment is, well, if you have some way for the state to be able to exit back to the layer‑1,” he says, “then you have a lot of con­tention on the layer‑1.”

“And so you have all the same cen­tral­iza­tion pres­sure on the layer‑1,” he says.

“I don’t think, by any means, it’s per­fect­ly solved.”

“At the end of the day, yes, you’re going to have trade-offs in these dif­fer­ent exe­cu­tion envi­ron­ments,” Gos­selin admits, but ulti­mate­ly, app devel­op­ers just want an inter­face for con­nect­ing and auto­mat­i­cal­ly deploy­ing their services.

“These shared sequencers or decen­tral­ized block builders, cross-chain bridges, are all in the same game of try­ing to build and pro­vide those ser­vices,” he says.

“There’s so many dif­fer­ent ways to build these things — and it’s not clear to me where it’s going to go.”

Get the day’s top cryp­to news and insights deliv­ered to your email every evening. Sub­scribe to Block­works’ free newslet­ter now.

Want alpha sent direct­ly to your inbox? Get degen trade ideas, gov­er­nance updates, token per­for­mance, can’t‑miss tweets and more from Block­works Research’s Dai­ly Debrief.

Can’t wait? Get our news the fastest way pos­si­ble. Join us on Telegram and fol­low us on Google News.