Rollup sequencers are specialized nodes that are perhaps the most important part of a rollup system. They perform the most crucial functions of a rollup, such as processing and batching together transactions, and submitting those batches to the Ethereum mainnet.
The following paragraphs offer a closer look at this important technology and its different implementations. To start off, let’s briefly go over the sequencing process.
Rollup sequencers 101
We can think of rollup sequencers as the equivalent of validators in rollup networks, but with additional specialized functions owing to the fact that they operate within a Layer 2 network. So a sequencer is responsible for picking up transactions from a Layer 2 mempool and then processing them to be in a state that is appropriate for submission to the Ethereum mainnet. That process involves verifying and ordering the transactions, then bundling them together, which allows them to be sent to Ethereum via a single transaction. Where applicable (read in ZK rollups) the sequencer also the validity proof for those transactions to Ethereum.
var gCaptchaSibWidget;
var onloadSibCallbackInvisible = function () {
var element = document.getElementsByClassName(‘sib-default-btn’);
var countInvisible = 0;
var indexArray = [];
jQuery(‘.sib-default-btn’).each(function (index, el) {
if ((jQuery(el).attr(‘id’) == “invisible”)) {
indexArray[countInvisible] = index;
countInvisible++
}
});
jQuery(‘.invi-recaptcha’).each(function (index, el) {
grecaptcha.render(element[indexArray[index]], {
‘sitekey’: jQuery(el).attr(‘data-sitekey’),
‘callback’: sibVerifyCallback,
});
});
};
form#sib_signup_form_1 .sib-default-btn { color: #fff; text-align: center; cursor: pointer; vertical-align: middle; -webkit-user-select: none; -moz-user-select: none; -ms-user-select: none; user-select: none; white-space: normal; border-radius: 0px 5px 5px 0px; width: 30%; height:45px; padding: 0; margin: 0; font-size: 16px; font-weight: 700; } input.sib-email-area { border-radius: 5px 0px 0px 5px; border: solid 1px #D4DCDC; background: #fff; width: 70%; height:45px; font-size: 14px; padding-left: 20px; padding-right: 20px; } .sib-email-area { display: flex; }form#sib_signup_form_2 p.sib-alert-message {
padding: 6px 12px;
margin-bottom: 20px;
border: 1px solid transparent;
border-radius: 4px;
-webkit-box-sizing: border-box;
-moz-box-sizing: border-box;
box-sizing: border-box;
}
form#sib_signup_form_2 p.sib-alert-message-error {
background-color: #f2dede;
border-color: #ebccd1;
color: #a94442;
}
form#sib_signup_form_2 p.sib-alert-message-success {
background-color: #dff0d8;
border-color: #d6e9c6;
color: #3c763d;
}
form#sib_signup_form_2 p.sib-alert-message-warning {
background-color: #fcf8e3;
border-color: #faebcc;
color: #8a6d3b;
}
Types of rollup sequencers
While rollup sequencers all perform similar tasks, they differ in terms of their structure and level of decentralization. Today, we have predominantly centralized rollup sequencers, but there are some notable projects that employ decentralized sequencers. More recently, we’ve had two other types joining the party – shared sequencers and based rollups. In the following paragraphs, we’ll be taking a closer look at all of these categories.
Centralized sequencers
Decentralization is the ultimate goal of any Web3 system, but the centralized approach has its advantages. Typically those include gains in efficiency. So it’s not surprising that efficiency is the main selling point of centralized rollup sequencers.
Decentralized sequencers
A decentralized sequencer is a network of nodes tasked with taking on the sequencing responsibilities. Typically, a node from the network is selected randomly to handle sequencing duties for a set period of time. When its mandate expires, a new node is selected.
Clearly, this introduces an additional level of complexity to the sequencing process, which hinders efficiency and increases the time that it takes to complete it. But this is a familiar tradeoff.
On the flipside, decentralized sequencers offer a level of stability that their centralized counterparts simply cannot match. If a centralized sequencer suffers an outage, the entire rollup network that relies on it practically shuts down until the issue is resolved. With a decentralized rollup you can simply select another node to handle the sequencing.
Shared sequencers
The easiest way to have a decentralized sequencer is by utilizing the rollup’s native network of validators. But there is also another way that you can use a decentralized network for sequencing and that is where shared sequencers come in.
A shared sequencer is essentially a separate decentralized network of nodes that is not tied to a particular rollup, but can be utilized by multiple rollups, instead. It’s a ‘sequencing-as-a-service’ solution of sorts that could become increasingly popular in the future, if the trajectory of the ‘aaS’ model in other industries is any indicator.
Based rollups
Shared rollup sequencers sound like a cool tech and can be a very good alternative to building a dedicated sequencer. But can we take the concept even further? Well, based rollups are kind of doing exactly that.
The whole idea behind based rollups is that they use the validators of their underlying blockchain for sequencing. It’s a really straightforward concept, but one that has some interesting advantages. For one, it eliminates a lot of the hassle associated with building a dedicated rollup sequencer. It can also tap into the inherent decentralization and liveliness of an Layer 1 network, but without sacrificing the sovereignty of the rollup. Based rollups can also be more cost-efficient than the L2s that use dedicated rollup sequencers. To learn more about the technology check out our detailed “What are based rollups?” piece.
Currently, Taiko is the only major rollup platform that uses the ‘based’ model.
Conclusion
Rollup sequencers are an important part of rollup infrastructure and, by extension, a crucial component of the expanding Web3 space. Most of the big names in the rollup space today are running centralized sequencers, but decentralized sequencers are actively being explored, while the emergence of the ‘sequencing-as-a-service’ approach is contributing to the multitude of options available out there.
