ACM Sigarch

APRIL 26, 2018

APU: Accelerated Processing Unit is the AMD’s Fusion architecture that integrates both CPU and GPU on the same die. BPU: Brain Processing Unit is the design of the AI chips by Horizon Robotics. They unveiled their first two embedded AI chips fabricated with TSMC 40nm process in December 2017: “Journey 1.0

Hardware 99

Hardware Artificial Intelligence Architecture Google 99

Brendan Gregg

FEBRUARY 28, 2023

Titus, the Netflix container management platform, is now open source,” [link] Apr 2018 - [Cutress 19] Dr. [link] 2021 - [Trader 21] Tiffany Trader, “Cerebras Second-Gen 7nm Wafer Scale Engine Doubles AI Performance Over First-Gen Chip ,” [link] Apr 2021 - [Ghigoff 21] Yoann Ghigoff, et al., "BMC:

Performance 110

Performance Latency Cache Virtualization 110

All Things Distributed

MAY 19, 2017

That database engine is now known as Amazon Aurora and launched in 2014 for MySQL and in 2016 for PostgreSQL. We believe the central constraint in high throughput data processing has moved from compute and storage to the network.

Database 117

Database Design Cloud Storage 117

Brendan Gregg

OCTOBER 8, 2018

Some one-liners: # New processes with arguments bpftrace -e 'tracepoint:syscalls:sys_enter_execve { join(args->argv); }'. Files opened by process bpftrace -e 'tracepoint:syscalls:sys_enter_open { printf("%s %sn", comm, str(args->filename)); }'. Pages paged in by process bpftrace -e 'software:major-faults:1 { @[comm] = count(); }'.

C++ 110

C++ Virtualization Programming Latency 110

Brendan Gregg

FEBRUARY 28, 2023

Titus, the Netflix container management platform, is now open source,” [link] , Apr 2018 [Cutress 19] Dr. [link] , 2021 [Trader 21] Tiffany Trader, “Cerebras Second-Gen 7nm Wafer Scale Engine Doubles AI Performance Over First-Gen Chip ,” [link] , Apr 2021 [Ghigoff 21] Yoann Ghigoff, et al.,

Performance 52

Performance Latency Cache Virtualization 52

All Things Distributed

DECEMBER 8, 2016

To increase online readership, it worked with AWS Partner Network (APN) Partner ClearScale to develop a personal recommendation capability. We are excited to offer a robust portfolio of services from our foundational service stack for compute, storage, and networking to our more advanced solutions and applications.

AWS 155

AWS Cloud Lambda Innovation 155

Percona

APRIL 12, 2023

2016: The company adds service-loaded MongoDB Professional to its mix. and partners) and in the process, stifling the upstarts, the truly open source little folks dreaming big. Here are other key developments in company history: 2013: 10gen changes its name to MongoDB, Inc. looking out for MongoDB Inc.

Open Source 81

Open Source Programming Database Servers 81

Adrian Cockcroft

JANUARY 29, 2023

I presented this analysis of response time distributions talk in 2016 — at Microxchg in Berlin ( video ). For high traffic systems, processing the individual response times for each request may be too much work. I’ve been thinking about this for a long time. > system.time(wait1 <- normalmixEM(waiting, mu=c(50,80), lambda=.5,

Lambda 98

Lambda Latency Cache C++ 98

Percona

SEPTEMBER 19, 2023

Copyright (c) 2016, 2023, Oracle and/or its affiliates. The backup operation on InnoDB is a non-blocking operation, so the ibd files are allowed to change while the backup process is running. PXB runs a background process that watches the transaction log files and copies changes from them. Type 'help' or '?' X, but got no.

C++ 101

C++ Servers Storage Speed 101

Brendan Gregg

OCTOBER 7, 2018

Some one-liners: # New processes with arguments bpftrace -e 'tracepoint:syscalls:sys_enter_execve { join(args->argv); }'. Files opened by process bpftrace -e 'tracepoint:syscalls:sys_enter_open { printf("%s %sn", comm, str(args->filename)); }'. Pages paged in by process bpftrace -e 'software:major-faults:1 { @[comm] = count(); }'.

C++ 40

C++ Virtualization Programming Latency 40

CSS - Tricks

JANUARY 31, 2020

From 2016 (PDF warning) to 2019 , the average mobile download speed in the U.S. Despite every significant improvement to the underpinnings of the web and the networks we access it through, we continue to build for it in ways that suggest we’re content with the never-ending Jevons paradox in which we toil. has increased by 73%.

Innovation 50

Innovation Mobile Best Practices Speed 50

O'Reilly

APRIL 23, 2024

Perhaps a more direct way to say this in the context of economic value creation is that companies such as Amazon and Google and Facebook had developed a set of remarkable advances in networked and data-enabled market coordination. But over time, something went very wrong. Cory Doctorow calls this the “enshittification” of Big Tech platforms.

Social Media 111

Social Media Google Internet Internet 111

Smashing Magazine

JANUARY 15, 2019

This includes the work done by the server, the client and the intermediary communications networks that transmit data between the two. All of the major browsers provide developer tools that allow us to measure network activity. In 2016, O’Reilly published “ Designing For Sustainability ” by Tim Frick. Large preview ).

Website Performance 74

Website Performance Website Energy Performance 74

Dynatrace

JUNE 7, 2019

Although, the upgrade process involves reinstallation of the ActiveGates that make this location or adding new ones enabled for synthetic monitoring. LTS, Red Hat server 7 and CentOS server 7 for the execution of private browser monitors; Windows Server 2016 support is coming soon as well. Can I use it for browser monitors as well?

Network 152

Network Monitoring Servers Hardware 152

Rigor

SEPTEMBER 11, 2019

Reading time 16 min Whether you’re a web performance expert, an evangelist for the culture of performance, a web engineer incorporating performance into your process, or someone new to the web performance entirely, you probably identify as curious, excited about new ideas, and always learning. ” Paul Calvano. Paul Calvano.

Performance 89

Performance Education Google Website 89

SQL Server According to Bob

FEBRUARY 10, 2019

Network Trace. Looking at network traces helps build our understanding of the problem. Network Layer Overloaded. In contrast to the failing SQL Server login, this is what a network trace looks like when the network layer becomes overloaded and is not completing requests to the SQL Server. SQL Authentication Impact.

Servers 40

Servers Network Database Systems 40

Alex Russell

DECEMBER 18, 2022

TL;DR : To serve users at the 75 th percentile ( P75 ) of devices and networks, we can now afford ~150KiB of HTML/CSS/fonts and ~300-350KiB of JavaScript (gzipped). This is a slight improvement on last year's budgets , thanks to device and network improvements. That is, what was the average device in 2016?

Performance 113

Performance Network Mobile Latency 113

All Things Distributed

DECEMBER 13, 2016

In just this past month we've had HSBC , ARM , Missguided , and most recently at re:Invent 2016 , Trainline , talking with us about how they are using AWS to transform and scale their businesses. Take GoSquared , a UK startup that runs all its development and production processes on AWS, as an example.

AWS 166

AWS Cloud Artificial Intelligence IoT 166

Alex Russell

MARCH 6, 2021

Thanks to progress in networks and browsers (but not devices), a more generous global budget cap has emerged for sites constructed the "modern" way: ~100KiB of HTML/CSS/fonts and ~300-350KiB of JS (compressed) is the new rule-of-thumb limit for at least the next year or two. Modern network performance and availability.

Performance 124

Performance Network Cache Metrics 124

Testsigma

JANUARY 7, 2022

In 2016, Google made it clear that since mobile traffic is more than all else, mobile-friendly websites will be prioritised when a user searches on mobile. QAOps is a term derived by combining the two processes – the DevOps and QA into one. When we fuse QA into the DevOps process the newly integrated process is called QAOps.

Artificial Intelligence 89

Artificial Intelligence IoT Software Software 89

CSS Wizardry

MARCH 3, 2019

The best request is the one that never happens: in the fight for fast websites, avoiding the network is far better than hitting the network at all. Any asset that carries the no-store directive will always hit the network, no matter what. How is your knowledge of caching and Cache-Control headers? Now we’d use no-store.

Cache 264

Cache Latency Strategy Servers 264

CSS - Tricks

APRIL 1, 2021

Using a network request inspector, I’m going to see if there’s anything we can remove via the Network panel in DevTools. I like using Fiddler , and when I inspect the network requests I see that there are indeed some old URLs and redirects floating around. We can repoint all our redirect-to-slow-site URLs to the updated URL.

Cache 92

Cache Social Media Media Website 92

High Scalability

AUGUST 8, 2018

Niantic launched Pokémon GO in the summer of 2016. Prior to launch, they load-tested their software stack to process up to 5x their most optimistic traffic estimates. Pokémon GO’s success greatly exceeded the expectations of the Niantic engineering team.

Google 130

Google Games Entertainment Cloud 130

Brendan Gregg

NOVEMBER 29, 2017

In this configuration, the AMI and boot is paravirt (PV), the kernel is making hypercalls instead of privileged instructions, and the system is using paravirt network and storage drivers. AWS EC2 uses a different image type and boot process for PV and HVM, as described on the [Linux AMI Virtualization Types] page. The first was c3.

Virtualization 111

Virtualization AWS Hardware Storage 111

Brendan Gregg

SEPTEMBER 5, 2017

LTS (April 2016). There's also a ZFS send/recv code path that should try to use the TASK_INTERRUPTIBLE flag (as suggested by a coworker), to avoid a kernel hang (can't kill -9 the process). I wrote about it in a previous post, [DTrace for Linux 2016]. Both of those should be easy fixes. It's the official profiler.

Virtualization 76

Virtualization AWS Engineering Hardware 76

Alex Russell

MARCH 31, 2022

De-duplication can prevent multiple copies on disk, but we're out of luck on reducing network costs. Late in the drafting of this post, it was pointed out to me by Frederik Braun that Brad Hill had considered the downsides of a site-populated SRI addressable cache back in 2016. It turns out this idea has been circulating for a while.

Cache 82

Cache Government Traffic Network 82

O'Reilly

OCTOBER 29, 2019

Network propaganda works in much the same way as microphones and music. We saw those feedback loops in 2016, we’ve been seeing them ever since, and we will certainly see them used in 2020. While many of our systems are vulnerable to attack, social networks have proven the most vulnerable. Propaganda isn’t a new phenomenon.

Social Media 53

Social Media Media Network Energy 53

Smashing Magazine

JULY 12, 2021

There is also the matter of convincing the management or project stakeholders to invest the time and resources into the refactoring process. For example, the client is in process of rebranding and the website needs to be updated with a new design and revamped content. Incremental and flexible process. Large preview ).

Best Practices 145

Best Practices Website Architecture Media 145

Alex Russell

OCTOBER 22, 2017

We constrain ourselves to a real-world baseline device + network configuration to measure progress. Budgets are scaled to a benchmark network & device. Answering it requires an understanding of how browsers process resources (which differs by type) and the concept of the critical path. ” A good question!

Performance 68

Performance Benchmarking Network Mobile 68

Alex Russell

MAY 9, 2022

Effective managers understand that products require processes to manage, yet front-end discourse assumes that switching frameworks will introduce a golden age. Reduced network and device capacity correlate with other access challenges. Only teams that master their systems can make intentional trade-offs.

Performance 86

Performance Latency Metrics Engineering 86

O'Reilly Software

APRIL 4, 2018

We live in the era of the connected experience, where our daily interactions with the world can be digitized, collected, processed, and analyzed to generate valuable insights. Accumulating all this data to process overnight is not an option anymore. Do we need to process each record individually? Process tolerance.

Architecture 80

Architecture Scalability Google Operating System 80

Brendan Gregg

NOVEMBER 28, 2017

In this configuration, the AMI and boot is paravirt (PV), the kernel is making hypercalls instead of privileged instructions, and the system is using paravirt network and storage drivers. AWS EC2 uses a different image type and boot process for PV and HVM, as described on the [Linux AMI Virtualization Types] page. The first was c3.

Virtualization 40

Virtualization AWS Hardware Storage 40

Alex Russell

APRIL 29, 2021

At the time of the last Confluence run, the gap had stretched to nearly 1000 APIs, doubling since 2016. Real-time network protocols for enabling videoconferencing, desktop sharing, and game streaming applications. Modern, asynchronous network APIs that dramatically improve performance in some situations. Delayed five years.

Media 145

Media Games Education Engineering 145

Smashing Magazine

JULY 6, 2021

Moving to a headless CMS is not an easy decision and the selection process should not be taken lightly. With Luminary’s 21 years of experience in the digital space, my experience of 17 years in the CMS space as well as our focus on Headless since 2016, here are my two cents on what you should look out for. Microservices architecture.

Cache 143

Cache Website Mobile Google 143

Tim Kadlec

JANUARY 31, 2019

They prevent abuse by ad networks and many people are realizing the benefits of that with increased performance and better battery life. Never-Slow Mode isn’t that far removed from two ideas that we had back in 2016 when Yoav Weiss and I met with the AMP team to discuss some standards-based alternatives to AMP. Probably not.

Tuning 76

Tuning Cache Development Performance 76

Rigor

SEPTEMBER 3, 2019

This information can then be used to set up tests within a synthetic monitoring data platform such as Rigor to test how a user experience those flows on every type of device, network, connection, and location. This process is often quicker than making the origin request and further reduces the load on your infrastructure.

Website 74

Website Cache Ecommerce Traffic 74

SQL Performance

SEPTEMBER 2, 2021

SQL Server 2016 brought us STRING_SPLIT , a native function that eliminates the need for many of the custom solutions we’ve needed before. STRING_SPLIT() in SQL Server 2016 : Follow-Up #1. STRING_SPLIT() in SQL Server 2016 : Follow-Up #2. UPDATE: September 2nd, 2021 (Originally published July 26th, 2012.).

C++ 59

C++ Servers Testing Database 59