Building fault-tolerant applications on AWS

-

Below is my summary of the Amazon Web Services (AWS) whitepaper released on Oct 2011 by Jeff Barr, Attila Narin, and Jinesh Varia.

AWS aims to simplify the task of building and maintaining fault-tolerant distributed systems/services for its customers. For this, AWS prescribes the customers to embrace the following philosophy when they are building their applications on AWS.

1. Make computing nodes disposable/easily replaceable

AWS (as with any cloud provider actually) employs a level of indirection over the physical computer, called virtual machine (VM), to make computing nodes easily replaceable. You then need a template to define your service instance over a VM, and this is called Amazon Machine Image (AMI). The first step towards building fault-tolerant applications on AWS is to create your own AMIs. Starting your application then is simply a matter of launching VM instances on Amazon Elastic Compute Cloud (EC2) using your AMI. Once you have created an AMI, replacing a failing instance is very simple; you can just launch a replacement instance that uses the same AMI as its template. This can be done programmatically through an API invocation.

In short, AWS wants you to embrace an instance as the smallest unit of failure and make it easily replaceable. AWS helps you to automate and make this process more transparent by providing elastic IP addresses and elastic load balancing. To minimize downtime, you may keep a spare instance running, and easily fail over to this hot instance by quickly remapping your elastic IP address to this new instance. Elastic load balancing further facilitates this process by detecting unhealthy instances within its pool of Amazon EC2 instances and automatically rerouting traffic to healthy instances, until the unhealthy instances have been restored.

2. Make state/storage persistent

When you launch replacement service instances on EC2 as above, you also need to provide persistent state/data that these instances have access to. Amazon Elastic Block Store (EBS) provides block level storage volumes for use with Amazon EC2 instances. Amazon EBS volumes are off-instance storage that persists independently from the life of an instance. Any data that needs to persist should be stored on Amazon EBS volumes, not on the local hard-disk associated with an Amazon EC2 instance because that disappears when the instance die. If the Amazon EC2 instance fails and needs to be replaced, the Amazon EBS volume can simply be attached to the new Amazon EC2 instance. EBS volumes store data redundantly, making them more durable than a typical hard drive. To further mitigate the possibility of a failure, backups of these volumes can be created using a feature called snapshots.

Of course this begs the question of "what is the sweet-point in storing to EBS" as it comes with a significant penalty over EC2 RAM, and a slight penalty over EC2 disk. I guess this depends on how you can stretch the definition of "the data that needs to persist" in your application.

3. Rejuvenate your system by replacing instances

If you follow the first two principles, you can (and should) rejuvenate your system by periodically replacing old instances with new server instances transparently. This ensures that any potential degradation (software memory leaks, resource leaks, hardware degradation, filesystem fragmentation, etc.) does not adversely affect your system as a whole.

4. Use georeplication to achieve disaster tolerance

Amazon Web Services are available in 8 geographic "regions". Regions consist of one or more Availability Zones (AZ), are geographically dispersed, and are in separate geographic areas or countries. The Amazon EC2 service level agreement commitment is 99.95% availability for each Amazon EC2 Region. But in order to achieve the same availability in your application, you should deploy your application over multiple availability zones, for example by maintaining a fail-over site in another AZ as in the figure.

5. Leverage other Amazon Web Services as fault-tolerant building blocks

Amazon Web Services offers a number of other services (Amazon Simple Queue Service, Amazon Simple Storage Service, Amazon SimpleDB, and Amazon Relational Database Service.) that can be incorporated into your application development. These services are fault-tolerant, so by using them, you will be increasing the fault tolerance of your own applications.

Let's take the Amazon Simple Queue Service (SQS) example. SQS is a highly reliable distributed messaging system that can serve as the backbone of your fault-tolerant application. Once a message has been pulled by an instance from an SQS queue, it becomes invisible to other instances (consumers) for a configurable time interval known as a visibility timeout. After the consumer has processed the message, it must delete the message from the queue. If the time interval specified by the visibility timeout has passed, but the message isn't deleted, it is once again visible in the queue and another consumer will be able to pull and process it. This two-phase model ensures that no queue items are lost if the consuming application fails while it is processing a message. Even in an extreme case where all of the worker processes have failed, Amazon SQS will simply store the messages for up to four days.

Conclusions:

Netflix (one of the largest and most prominent customers of AWS) embraced all of the above points. In fact Netflix has developed a chaos monkey tool to constantly and unpredictably kill some of its service instances in an effort to enforce that their services are built in a fault-tolerant and resilient manner and expose and resolve hidden problems with their services.

OK, after reading this, you can say that at some level the cloud computing fault-tolerance is boring: the prescribed fault correction action is to just replace the failed instance with a new instance. And if you say this, this will make the AWS folks happy, because this is the goal that they try to attain. They want to make faults uninteresting, and automatically dealt with. Unfortunately, not all faults can be isolated at the instance level, the real world isn't that simple. There are many different types of faults, such as misconfigurations, unanticipated faults, application-level heisenbugs, and bohrbugs, that won't fit into this mold. I intend to investigate these remaining nontrivial types of faults, and how to deal with them. I am particularly interested in exploring what role can self-stabilization play here. Another point, that didn't get coverage in this whitepaper is about how to detect faults and unhealthy instances. I would be interested to learn what techniques are employed in practice by AWS applications for this.

Comments

Email archiving software said…
This comment has been removed by a blog administrator.
January 29, 2013 at 7:29 AM
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 >>