<description>&lt;p&gt;&lt;br&gt;&lt;/p&gt;&lt;p&gt;Page tables provide the mapping between virtual memory and physical memory for each process. This means it needs to be as efficient and as fast as possible. I explore the inner workings of page tables in this episode.&lt;/p&gt;&lt;p&gt;&lt;br&gt;&lt;/p&gt;&lt;ul&gt;&lt;li&gt;0:00 Intro&lt;/li&gt;&lt;li&gt;2:00 Virtual Memory&lt;/li&gt;&lt;/ul&gt;&lt;p&gt;	⁃		⁃	8:00 MMU&lt;/p&gt;&lt;ul&gt;&lt;li&gt;10:00 Page Tables&lt;/li&gt;&lt;/ul&gt;&lt;p&gt;	⁃		⁃		⁃		⁃		⁃		⁃		⁃	11:30 Single Table Byte Addressability  &lt;/p&gt;&lt;p&gt;	⁃		⁃		⁃		⁃		⁃		⁃		⁃		⁃	16:00 Single Table Page addressability &lt;/p&gt;&lt;p&gt;	⁃		⁃		⁃		⁃		⁃	19:00 Multi-level Paging (Radix tree)&lt;/p&gt;&lt;p&gt;	⁃		⁃	31:00 Huge Tables&lt;/p&gt;&lt;p&gt;	⁃		⁃	33:00 TLB&lt;/p&gt;&lt;p&gt;	⁃		⁃	Summary&lt;/p&gt;&lt;p&gt;&lt;br&gt;&lt;/p&gt;
</description>

Page tables provide the mapping between virtual memory and physical memory for each process. This means it needs to be as efficient and as fast as possible. I explore the inner workings of page tables in this episode. <ul><li>0:00 Intro</li><li>2:00 Virtual Memory</li></ul>	⁃		⁃	8:00 MMU<ul><li>10:00 Page Tables</li></ul>	⁃		⁃		⁃		⁃		⁃		⁃		⁃	11:30 Single Table Byte Addressability  	⁃		⁃		⁃		⁃		⁃		⁃		⁃		⁃	16:00 Single Table Page addressability 	⁃		⁃		⁃		⁃		⁃	19:00 Multi-level Paging (Radix tree)	⁃		⁃	31:00 Huge Tables	⁃		⁃	33:00 TLB	⁃		⁃	Summary

Page Tables

<description>&lt;p&gt;Page faults occurs when the process tries to access a memory that isn’t backed by a physical page kernel raises a fault which loads a page. It happens on first access, stack expansion, COW, swap and much more. However it comes with a cost. &lt;/p&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;&lt;p&gt;In this episode of the backend engineering show I dissect the need and the cost page faults in the kernel. &lt;/p&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;&lt;ul&gt;&lt;li&gt;0:00 Intro &lt;/li&gt;&lt;li&gt;4:00 Virtual memory&lt;/li&gt;&lt;ul&gt;&lt;li&gt;Abstraction of physical memory&lt;/li&gt;&lt;li&gt;Memory sharing&lt;/li&gt;&lt;li&gt;Allow more processes to run , unused go to disk&lt;/li&gt;&lt;li&gt;Numa, kernel can place memory near the cpu&lt;/li&gt;&lt;/ul&gt;&lt;li&gt;12:00 VMA areas&lt;/li&gt;&lt;ul&gt;&lt;li&gt;Text/code &lt;/li&gt;&lt;li&gt;Data &lt;/li&gt;&lt;li&gt;BSS&lt;/li&gt;&lt;li&gt;Heap&lt;/li&gt;&lt;li&gt;Stack&lt;/li&gt;&lt;/ul&gt;&lt;li&gt;19:50 Kernel mode&lt;/li&gt;&lt;li&gt;25:30 What is a Page fault?&lt;/li&gt;&lt;li&gt;30:30 First access page fault&lt;/li&gt;&lt;li&gt;33:00 Stack Expansion page fault&lt;/li&gt;&lt;li&gt;34:30 CoW page fault&lt;/li&gt;&lt;li&gt;38:00 Swap page fault&lt;/li&gt;&lt;li&gt;39:39 File backed page fault&lt;/li&gt;&lt;li&gt;40:29 Permission page fault &lt;/li&gt;&lt;li&gt;&lt;strong&gt;45:30 Summary&lt;/strong&gt;&lt;/li&gt;&lt;/ul&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;
</description>

Page faults occurs when the process tries to access a memory that isn’t backed by a physical page kernel raises a fault which loads a page. It happens on first access, stack expansion, COW, swap and much more. However it comes with a cost.  In this episode of the backend engineering show I dissect the need and the cost page faults in the kernel.  <ul><li>0:00 Intro </li><li>4:00 Virtual memory</li><ul><li>Abstraction of physical memory</li><li>Memory sharing</li><li>Allow more processes to run , unused go to disk</li><li>Numa, kernel can place memory near the cpu</li></ul><li>12:00 VMA areas</li><ul><li>Text/code </li><li>Data </li><li>BSS</li><li>Heap</li><li>Stack</li></ul><li>19:50 Kernel mode</li><li>25:30 What is a Page fault?</li><li>30:30 First access page fault</li><li>33:00 Stack Expansion page fault</li><li>34:30 CoW page fault</li><li>38:00 Swap page fault</li><li>39:39 File backed page fault</li><li>40:29 Permission page fault </li><li>45:30 Summary</li></ul>

CPU and Kernel Page Faults

<description>&lt;p&gt;On October 19 2025 AWS experienced an outage that lasted over a day, 10 days later we finally got the root cause analysis and we know exactly what caused the DNS to fail0:00 Summary 5:30 How did Dynamo lost its DNS?13:41 EC2 Errors 16:16 Network Load Balancer ErrorsRCA here https://aws.amazon.com/message/101925/&lt;/p&gt;
</description>

On October 19 2025 AWS experienced an outage that lasted over a day, 10 days later we finally got the root cause analysis and we know exactly what caused the DNS to fail0:00 Summary 5:30 How did Dynamo lost its DNS?13:41 EC2 Errors 16:16 Network Load Balancer ErrorsRCA here https://aws.amazon.com/message/101925/

Amazon US-EAST-1 Outage in Details

<description>&lt;p&gt;&lt;br /&gt;&lt;/p&gt;&lt;p&gt;There are cases where the backend may need to close the connection to prevent unexpected situations, prevent bad actors or simply just free up resources. Closing a connection gracefully allows clients and backends to clean up and finish any pending requests. &lt;/p&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;&lt;p&gt;In this episode of the backend engineering show I discuss graceful connections in both HTTP/1.1 via the connection header and HTTP/2 via the GOAWAY frame. &lt;/p&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;&lt;ul&gt;&lt;li&gt;0:00 Intro&lt;/li&gt;&lt;li&gt;4:58 Why shutdown connection? &lt;/li&gt;&lt;li&gt;6:46 HTTP/1.1 Graceful shutdown&lt;/li&gt;&lt;li&gt;12:26 Cost of HTTP/2 &lt;/li&gt;&lt;li&gt;17:40 HTTP/2 GoAWAY frame&lt;/li&gt;&lt;li&gt;23:40 Summary&lt;/li&gt;&lt;/ul&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;&lt;p&gt;Links&lt;/p&gt;&lt;p&gt;&lt;a href="https://www.youtube.com/watch?v=fVKPrDrEwTI&amp;amp;t=1s" rel="ugc noopener noreferrer" target="_blank"&gt;https://www.youtube.com/watch?v=fVKPrDrEwTI&amp;amp;t=1s&lt;/a&gt; &lt;/p&gt;&lt;p&gt;&lt;a href="https://chromium.googlesource.com/chromium/src/net/%2B/master/socket/client_socket_pool_manager.cc#76" rel="ugc noopener noreferrer" target="_blank"&gt;https://chromium.googlesource.com/chromium/src/net/%2B/master/socket/client_socket_pool_manager.cc#76&lt;/a&gt;&lt;/p&gt;&lt;p&gt;&lt;a href="https://issues.chromium.org/issues/40555364" rel="ugc noopener noreferrer" target="_blank"&gt;https://issues.chromium.org/issues/40555364&lt;/a&gt;&lt;/p&gt;&lt;p&gt;&lt;a href="https://issues.chromium.org/issues/40501721" rel="ugc noopener noreferrer" target="_blank"&gt;https://issues.chromium.org/issues/40501721&lt;/a&gt;&lt;/p&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;
</description>

There are cases where the backend may need to close the connection to prevent unexpected situations, prevent bad actors or simply just free up resources. Closing a connection gracefully allows clients and backends to clean up and finish any pending requests.  In this episode of the backend engineering show I discuss graceful connections in both HTTP/1.1 via the connection header and HTTP/2 via the GOAWAY frame.  <ul><li>0:00 Intro</li><li>4:58 Why shutdown connection? </li><li>6:46 HTTP/1.1 Graceful shutdown</li><li>12:26 Cost of HTTP/2 </li><li>17:40 HTTP/2 GoAWAY frame</li><li>23:40 Summary</li></ul> Links<a href="https://www.youtube.com/watch?v=fVKPrDrEwTI&amp;t=1s" rel="ugc noopener noreferrer" target="_blank">https://www.youtube.com/watch?v=fVKPrDrEwTI&amp;t=1s</a> <a href="https://chromium.googlesource.com/chromium/src/net/%2B/master/socket/client_socket_pool_manager.cc#76" rel="ugc noopener noreferrer" target="_blank">https://chromium.googlesource.com/chromium/src/net/%2B/master/socket/client_socket_pool_manager.cc#76</a><a href="https://issues.chromium.org/issues/40555364" rel="ugc noopener noreferrer" target="_blank">https://issues.chromium.org/issues/40555364</a><a href="https://issues.chromium.org/issues/40501721" rel="ugc noopener noreferrer" target="_blank">https://issues.chromium.org/issues/40501721</a>

Graceful shutdown in HTTP

<description>&lt;p&gt;Postgres 18 has been released with many exciting features such as UUIDv7, Over explain module, composite index skip scans, and the most anticipated asynchronous IO with worker and io_uring mode which I uncover in this show. Hope you enjoy it&lt;/p&gt;&lt;p&gt;&lt;br /&gt;&lt;/p&gt;&lt;p&gt;&lt;br /&gt;0:00 Intro&lt;/p&gt;&lt;p&gt;1:30 Synchronous vs Asynchronous calls&lt;/p&gt;&lt;p&gt;3:00 Synchronous IO&lt;/p&gt;&lt;p&gt;6:30 Asynchronous IO&lt;/p&gt;&lt;p&gt;10:00 Postgres 17 synchronous io &lt;/p&gt;&lt;p&gt;17:20 The challenge of Async IO in Postgres 18&lt;/p&gt;&lt;p&gt;20:00 io_method worker23:00&lt;/p&gt;&lt;p&gt; io_method io_uring&lt;/p&gt;&lt;p&gt;29:30 io_method sync &lt;/p&gt;&lt;p&gt;31:08 Async IO isn’t done! 3&lt;/p&gt;&lt;p&gt;1:30 Support for backend writers&lt;/p&gt;&lt;p&gt;32:36 Improve worker io_method&lt;/p&gt;&lt;p&gt;33:00 direct io support &lt;/p&gt;&lt;p&gt;37:00 Summary &lt;br /&gt;&lt;/p&gt;
</description>

Postgres 18 has been released with many exciting features such as UUIDv7, Over explain module, composite index skip scans, and the most anticipated asynchronous IO with worker and io_uring mode which I uncover in this show. Hope you enjoy it 0:00 Intro1:30 Synchronous vs Asynchronous calls3:00 Synchronous IO6:30 Asynchronous IO10:00 Postgres 17 synchronous io 17:20 The challenge of Async IO in Postgres 1820:00 io_method worker23:00 io_method io_uring29:30 io_method sync 31:08 Async IO isn’t done! 31:30 Support for backend writers32:36 Improve worker io_method33:00 direct io support 37:00 Summary

Postgres 18 gets Async IO

Welcome to the Backend Engineering Show podcast with your host Hussein Nasser. If you like software engineering you’ve come to the right place. I discuss all sorts of software engineering technologies and news with specific focus on the backend. All opinions are my own. 

Most of my content in the podcast is an audio version of videos I post on my youtube channel here http://www.youtube.com/c/HusseinNasser-software-engineering 

Buy me a coffee 
https://www.buymeacoffee.com/hnasr

🧑‍🏫 Courses I Teach
https://husseinnasser.com/courses 


The Backend Engineering Show with Hussein Nasser

Details

Recent Episodes