Does EIP-8105 add latency to ETH?
ETH Tbilisi 2.1 · April 2026 · 2:40
Luis Bezzenberger and Loring Harkness of Shutter Network debate whether encrypting the mempool adds latency or other problems for Ethereum transactions. The short clip addresses performance concerns around EIP-8105. Recorded at ETH Tbilisi 2.1. It follows the main EIP-8105 session on encrypting the mempool and keeps to the engineering trade-offs.
Speakers
Topics
From ETH Tbilisi 2.1, hosted by ETHTbilisi.
Transcript
And the next question would be to Louis. Uh I just want to know will the encrypted mem pools add any noticeable latency to the transactions? This one I want one thing to the question before um because the we're also working on this other protocol version of the Ethereum pool with Prime app, so um I think that's an opportunity to get it in earlier um and it depends a little more on adoption from supply chain makers like builders. So, it's a little more of an apple apple battle if we go this this route, but um at least there will be the option for people to use it. Might take a little longer to to um for a transaction to go in, but that that will
be an option and hopefully very soon. Um but yeah, um will there be a latency impact? Um uh yes. Um so, essentially the way we think about it is it it depends on the EIP and depends on the exact approach. But ours and Louis's EIP essentially we would say that if you do a transaction in this block, then it'll then if you don't send it through the encrypted mem pool, it would have the chance to land in this block, right? In the the one n, right? Whereas if you use the encrypted mem pool, it'll be the earliest of it that'll be included would be at the top of the block in n plus
one. So, so depending on that it yeah, it means like kind of like it'll land in the next block. So, you you you you don't have if you do it now, you don't have the chance that it lands in this block. It'll always be the next block. It's kind of like the way to think about it. Um but but like yeah, I I I think like in the grand scheme of things latency doesn't for Ethereum doesn't really matter because you have kind of this 12 seconds or in the future maybe 10 or maybe 10 seconds or maybe in the future maybe have 6 seconds. Um and I would say all the encryption technologies that we're looking into and all everything should kind of be well below that, right? I think we can get threshold encryption also kind of into the
to the order of 100 milliseconds to decrypt. So, I think on the time scales we're talking about in Ethereum, it the latency and this like time to decrypt doesn't really isn't really that big of a deal. Might may might be a bigger issue for like Solana. Because there's all about like super low latency or like mega e, right? Where you where it's everything about low latency. That then maybe the decryption step and all of that stuff adds more latency. That's not a Yeah.