DDIA: Chp 5. Replication (Part 1)

-

Chapter 5 of the Designing Data Intensive Applications (DDIA) book discusses strategies and challenges in replicating data across distributed systems.

Replication is critical for ensuring data availability, fault tolerance, and scalability. One of the key challenges in replication is maintaining consistency across multiple replicas. Leader-based replication, the most common method, involves a single leader node accepting all writes, while follower nodes replicate the leader's changes. While single leader replication does not guarantee consistency readily (due to leader failover cornercases), it gives us a fighting chance.

One primary advantage of leader-based replication is its simplicity: all write requests are serialized through the leader, ensuring a consistent order of operations. Followers receive a stream of changes (i.e., the replication log) from the leader and apply them in sequence, ensuring eventual consistency. However, replication lag, the delay between when a write is committed/acknowledged by the leader and when it is reflected in the followers, can lead to inconsistencies for reads. For example, a user might read outdated information from a follower if the follower hasn't caught up with the leader.

To mitigate this, several strategies are employed. For instance, when ensuring read-after-write consistency (read-your-writes), it is recommended to serve reads from the leader if the data has been recently modified by the user. This prevents the user from seeing stale data after making updates. Additionally, the system can track the timestamp of the user's most recent write and ensure that subsequent reads reflect updates until that timestamp. If a follower is not sufficiently up to date, the system can wait for it to catch up or route the query to a different replica.

Synchronous replication alleviates the replication lag problem because it ensures that followers are always in sync with the leader before the leader acknowledges a write. While this approach provides stronger consistency guarantees, it can degrade performance, as the leader must wait for followers to confirm receipt of each change. In practice, a hybrid approach, often called semi-synchronous replication, is commonly used. In this setup, one follower is kept synchronous to ensure availability in case the leader fails, while other followers replicate asynchronously to optimize performance. The quorum approach, used in Paxos, ABD, and variants, provide more flexibility by not requiring the synchronous replica to be predetermined, and change between updates. Note that synchronous replication is still prone to replication lag problem and need to employ the above strategies to ensure consistency for reading from the replicas. This is because of inherent asymmetry of view/information in distributed systems: the leader commits/acks first and the followers then learn of the commit/ack with a delay. Synchronous replication is still very useful for preventing replicas to diverge too far from the leader's state. With pipelined replication, it is still possible to pull good throughput from synchronous replication.

Asynchronous replication is used for relaxing the tight synchronization requirements and alleviating the latency of write-ack penalties associated with synchronous replication. it allows followers to operate bit more independently of the leader, and could be suitable for scenarios with heavy read workloads and light write activity especially in WAN environments. However, asynchronous replication carries the risk of inconsistency since there is no guarantee of how quickly followers will replicate changes from the leader.

Let's get back to the issue of replication inconsistency cornercases even when using single leader synchronous replication. Failover, the process of promoting a follower to leader when the leader fails, introduces additional complexity for consistency. The new leader may not have all the latest writes from the failed leader, leading to potential data loss or inconsistency. Or even worse,  in certain fault scenarios (involving lost messages and asynchrony), it could happen that two nodes both believe that they are the leader. This situation is called split brain, and it is dangerous: if both leaders accept writes, and there is no process for resolving conflicts, data is likely to be lost or corrupted. Properly handling thesse scenarios are hard and involves principled distributed consensus algorithm design. This is why Paxos and later implementations including Raft are so popular in distributed systems deployments.


How: Implementation of Replication Logs

There are three main log replication methods.

Statement-based replication logs every write statement (e.g., SQL INSERT, UPDATE, DELETE) executed by the leader and sends them to followers, who re-execute them. While compact, this method is error-prone due to nondeterminism (e.g., throught function call execution at the leader versus at the replicas). Though still used in some cases, it has largely been replaced by more reliable techniques.

Write-ahead log (WAL) shipping involves logging low-level changes to disk blocks (e.g., B-tree, Postgres). The leader sends its WAL to followers, enabling them to replicate exact disk state changes. This method is tightly coupled to the storage engine, limiting flexibility in database versioning, and requiring downtime during upgrades.

Logical (row-based) log replication decouples replication from storage and logs the changes at the row level. Each log record details the new or deleted row data, and transactions generate multiple records. This approach allows leader and follower nodes to run different database versions, facilitating zero-downtime upgrades. This also enables external systems to consume the log for tasks like change data capture. Among many other systems, MySQL's binlog (when configured to use row-based replication) uses this approach.


Multi-leader replication


Finally, multi-leader replication (aka master-master or active-active replication) is an alternative to leader-based replication.  It allows multiple nodes to accept writes simultaneously. Each node then forwards changes to the other leaders, which also act as followers. This setup can be useful for geographically distributed systems, but it introduces challenges in resolving write conflicts when concurrent writes occur across multiple leaders. Handling these conflicts requires sophisticated techniques to ensure consistency across nodes.


Some related paper summaries from my blog

These two papers provide a good overview of replicated data consistency models/definitions in distributed systems.

The many faces of consistency (2016)

Replicated Data Consistency Explained Through Baseball (2011)

There is also causal consistency model, which is very useful for read-your-write. An example is Don’t Settle for Eventual: Scalable Causal Consistency for Wide-Area Storage with COPS.

I recently reviewed these consistency papers from MongoDB.

Checking Causal Consistency of MongoDB

Implementation of Cluster-wide Logical Clock and Causal Consistency in MongoDB

Tunable Consistency in MongoDB

Clocks and hybrid logical clocks are becoming more popular in enforcing consistency.

On the use of Clocks to Enforce Consistency in the Cloud


On the asynchronous replication side of things, the "optimistic replication" survey from 2005 provides a great coverage of asynchronous replication literature. 

Probabilistically bounded staleness (PBS) paper deserve special mention as it added a new perspective when studying/analyzing asynchronous replication.

You can think of this as a practical followup to PBS: Measuring and Understanding Consistency at Facebook

And of course CRDTs. 

Conflict-free Replicated Data Types

Keep CALM and CRDT on

Comments

Post a Comment

Popular posts from this blog

Hints for Distributed Systems Design

-
This is with apologies to Butler Lampson, who published the " Hints for computer system design " paper 40 years ago in SOSP'83. I don't claim to match that work of course. I just thought I could draft this post to organize my thinking about designing distributed systems and get feedback from others. I start with the same  disclaimer Lampson gave. These hints are not novel, not foolproof recipes, not laws of design, not precisely formulated, and not always appropriate. They are just hints.  They are context dependent, and some of them may be controversial. That being said, I have seen these hints successfully applied in distributed systems design throughout my 25 years in the field, starting from the theory of distributed systems (98-01), immersing into the practice of wireless sensor networks (01-11), and working on cloud computing systems both in the academia and industry ever since. These heuristic principles have been applied knowingly or unknowingly and has proven
Read more >>

Learning about distributed systems: where to start?

-
This is definitely not a "learn distributed systems in 21 days" post. I recommend a principled, from the foundations-up, studying of distributed systems, which will take a good three months in the first pass, and many more months to build competence after that. If you are practical and coding oriented you may not like my advice much. You may object saying, "Shouldn't I learn distributed systems with coding and hands on? Why can I not get started by deploying a Hadoop cluster, or studying the Raft code." I think that is the wrong way to go about learning distributed systems, because seeing similar code and programming language constructs will make you think this is familiar territory, and will give you a false sense of security. But, nothing can be further from the truth. Distributed systems need radically different software than centralized systems do.  --A. Tannenbaum This quotation is literally the first sentence in my distributed systems syllabus. Inst
Read more >>

Making database systems usable

-
Image
C. J. Date's Sigmod 1983 keynote, "Database Usability", was prescient. Usability is the most important thing to the customers. They care less about impressive benchmarks or clever algorithms, and more about whether they can operate and use a database efficiently to query, update, analyze, and persist their data with minimal headache. (BTW, does anyone have a link to the contents of this Sigmod'83 talk? There is no transcript around, except for this short abstract .) The paper we cover today is from Sigmod 2007. It takes on the database usability problem raised in that 1983 keynote head-on, and calls out that the king is still naked.  Let's give some context for the year 2007. Yes, XML format was still popular then. The use-case in the paper is XQuery. The paper does not contain any  reference to json. MongoDB would be released in 2009 with the document model; and that seems to be great timing for some of the usability pains mentioned in the paper! Web 2.0 was in
Read more >>

Looming Liability Machines (LLMs)

-
As part of our zoom reading group ( wow, 4.5 years old now ), we discussed a paper that uses LLMs for automatic root cause analysis (RCA) for cloud incidents. This was a pretty straightforward application of LLMs. The proposed system employs an LLM to match incoming incidents to incident handlers based on their alert types, predicts the incident's root cause category, and provides an explanatory narrative. The only customization is through prompt-engineering. Since this is a custom domain, I think a more principled and custom-designed  machine learning system would be more appropriate rather than adopting LLMs. Anyways, the use of LLMs for RCAs spooked me vicerally. I couldn't find the exact words during the paper discussion, but I can articulate this better now. Let me explain. RCA is serious business Root cause analysis (RCA) is the process of identifying the underlying causes of a problem/incident, rather than just addressing its symptoms. One RCA heuristic is asking 5 Why&#
Read more >>

Foundational distributed systems papers

-
I talked about the importance of reading foundational papers last week. To followup, here is my compilation of foundational papers in the distributed systems area. (I focused on the core distributed systems area, and did not cover networking, security, distributed ledgers, verification work etc. I even left out distributed transactions, I hope to cover them at a later date.)  I classified the papers by subject, and listed them in chronological order. I also listed expository papers and blog posts at the end of each section. Time and State in Distributed Systems Time, Clocks, and the Ordering of Events in a Distributed System. Leslie Lamport, Commn. of the ACM,  1978. Distributed Snapshots: Determining Global States of a Distributed System. K. Mani Chandy Leslie Lamport, ACM Transactions on Computer Systems, 1985. Virtual Time and Global States of Distributed Systems.  Mattern, F. 1988. Practical uses of synchronized clocks in distributed systems. B. Liskov, 1991. Expository papers
Read more >>

Advice to the young

-
Image
I notice I haven't written any advice posts recently. Here is a collection of my advice posts pre 2020. I've been feeling all this elderly wisdom pent up in me, ready to pour at any moment. So here it goes. Get ready to quench your thirst from my fount of wisdom. No man, think for yourself, only get what works for you. It is called foundations, not theory Foundations of computer science (or rather any field of study) are the most important topics you can learn. These lay down the frame of thinking/perspective for that area of study. Yet, I am saddened to hear these called as "theory", and labeled as "unpractical". This couldn't be farther from the truth. Take a look at how I recommend studying distributed systems . Don't you dare call this "theory" and "unpractical". This lays the bedrock that you build your practice on. Don't skimp on the foundations. Don't build your home on quicksand. Keep your hands dirty, your mind cl
Read more >>

Linearizability: A Correctness Condition for Concurrent Objects

-
Image
This paper is from Herlihy and Wing appeared in ACM Transactions on Programming Languages and Systems 1990. This is the canonical reference for the linearizability definition. I had not read this paper in detail before, so I thought it would be good to go to the source to see if there are additional delightful surprises in the original text. Hence, this post. I will dive into a technical analysis of the paper first, and then discuss some of my takes toward the end. I had written an accessible explanation of linearizability earlier; you may want to read that first. I will assume an understanding of linearizability to keep this review at reasonable length. Introduction I love how the old papers just barge in with the model, without bothered by pleasantries such as motivation of the problem. These are the first two sentences of the introduction. "A concurrent system consists of a collection of sequential processes that communicate through shared typed objects . This model encompass
Read more >>

Scalable OLTP in the Cloud: What’s the BIG DEAL?

-
Image
This paper is from Pat Helland, the apostate philosopher of database systems, overall a superb person, and a good friend of mine. The paper appeared this week at CIDR'24. (Check out the program for other interesting papers). The motivating question behind this work is: " What are the asymptotic limits to scale for cloud OLTP (OnLine Transaction Processing) systems? " Pat says that the CIDR 2023 paper "Is Scalable OLTP in the Cloud a Solved Problem?" prompted this question.  The answer to the question? Pat says that the answer lies in the joint responsibility of database and the application. If you know of Pat's work, which I have summarized several in this blog , you would know that Pat has been advocating along these lines before. But this paper provides a very crisp, specific, concrete answer. Read on for my summary of the paper. Disclaimer: This is a wisdom and technical information/detail packed 13-page paper, so I will try my best to summarize the sa
Read more >>

Understanding the Performance Implications of Storage-Disaggregated Databases

-
Image
Storage-compute disaggregation in databases has emerged as a pivotal architecture in cloud environments, as evidenced by Amazon ( Aurora ), Microsoft ( Socrates ), Google (AlloyDB), Alibaba ( PolarDB ), and Huawei (Taurus). This approach decouples compute from storage, allowing for independent and elastic scaling of compute and storage resources. It provides fault-tolerance at the storage level. You can then share the storage for other services, such as adding read-only replicas for the databases. You can even use the storage level for easier sharding of your database. Finally, you can also use this for exporting a changelog asynchronously to feed into peripheral cloud services, such as analytics. Disaggregated architecture was the topic of Sigmod 23 panel . I think this quote summarizes the industry's thinking on the topic. "Disaggregated architecture is here, and is not going anywhere. In a disaggregated architecture, storage is fungible, and computing scales independently.
Read more >>

Designing Data Intensive Applications (DDIA) Book

-
Image
We started reading this book as part of Alex Petrov's book club . We just got started, so you can join us, by joining the discord channel above. We meet Wednesday's 11am Eastern Time.  Previously we had read transaction processing book by Grey and Reuters. This page links to my summaries of that book. Chp 1. Reliable, Scalable, and Maintainable Applications I love the diagrams opening each chapter. Beautiful! The first chapter consists of warm up stuff. It talks about the definitions of reliabilty, scalability, and maintainability. It is still engaging, because  it is written with an educator and technical blogger voice, rather than a dry academic voice. This book came out on 2017. Martin is working on the new version. So if you have comments for things to focus on for the new version, it would be helpful to collect them in a document and email it to Martin. For example, I am curious about how the below paragraph from the Preface will get revised with 8 more years of hindsight:
Read more >>