Alex Russell

JANUARY 30, 2024

It's time once again to update our priors regarding the global device and network situation. HTML, CSS, images, and fonts can all be parsed and run at near wire speeds on low-end hardware, but JavaScript is at least three times more expensive, byte-for-byte. What's changed since last year? and 75KiB of JavaScript.

Performance 90

Performance Network Mobile Speed 90

The Morning Paper

NOVEMBER 14, 2019

Taiji: managing global user traffic for large-scale internet services at the edge Xu et al., It’s another networking paper to close out the week (and our coverage of SOSP’19), but whereas Snap looked at traffic routing within the datacenter, Taiji is concerned with routing traffic from the edge to a datacenter. SOSP’19.

Traffic 42

Traffic Internet Internet Latency 42

The Morning Paper

JANUARY 30, 2020

Edge servers are the middle ground – more compute power than a mobile device, but with latency of just a few ms. Since we’re talking about mobile applications, we have to assume a changing environment over time, including the possibility of losing internet connectivity altogether. The Mobile Web Worker (MWW) System.

Mobile 104

Mobile Cloud Latency Games 104

Brendan Gregg

FEBRUARY 28, 2023

My personal opinion is that I don't see a widespread need for more capacity given horizontal scaling and servers that can already exceed 1 Tbyte of DRAM; bandwidth is also helpful, but I'd be concerned about the increased latency for adding a hop to more memory. Ford, et al., “TCP

Performance 110

Performance Latency Cache Virtualization 110

Brendan Gregg

JULY 4, 2021

AWS Graviton2); for memory with the arrival of DDR5 and High Bandwidth Memory (HBM) on-processor; for storage including new uses for 3D Xpoint as a 3D NAND accelerator; for networking with the rise of QUIC and eXpress Data Path (XDP); and so on. I also wrote about these topics in detail for my recent [Systems Performance 2nd Edition] book.

Performance 52

Performance Latency Hardware Storage 52

Brendan Gregg

FEBRUARY 28, 2023

My personal opinion is that I don't see a widespread need for more capacity given horizontal scaling and servers that can already exceed 1 Tbyte of DRAM; bandwidth is also helpful, but I'd be concerned about the increased latency for adding a hop to more memory. Ford, et al., “TCP

Performance 52

Performance Latency Cache Virtualization 52

VoltDB

JUNE 1, 2017

It was – like the hypothetical movie I describe above – more than a little bit odd, as you could leave a session discussing ever more abstract layers of virtualization and walk into one where they emphasized the critical importance of pinning a network interface to a specific VM for optimal performance.

Hardware 56

Hardware Latency Virtualization Cloud 56

Smashing Magazine

AUGUST 3, 2021

Lots can go wrong: a network request fails, a third-party library breaks, a JavaScript feature is unsupported (assuming JavaScript is even available), a CDN goes down, a user behaves unexpectedly (they double-click a submit button), the list goes on. The more enriched sentence (right) is an enhancement for when the network request succeeds.

Network 145

Network Engineering Availability Monitoring 145

Testsigma

JANUARY 30, 2021

With the rapidly increasing use of smartphones and ease of access to the internet across the globe, testing has spread across vast platforms. For example, if you are using internet banking via a Mobile Web application, it will not allow you to save cards or mark any transaction as favourite. What are Mobile Web Applications?

Mobile 52

Mobile Testing Website Internet 52

VoltDB

JUNE 1, 2017

It was – like the hypothetical movie I describe above – more than a little bit odd, as you could leave a session discussing ever more abstract layers of virtualization and walk into one where they emphasized the critical importance of pinning a network interface to a specific VM for optimal performance.

Hardware 40

Hardware Latency Virtualization Cloud 40

Highly Scalable

SEPTEMBER 18, 2012

Historically, NoSQL paid a lot of attention to tradeoffs between consistency, fault-tolerance and performance to serve geographically distributed systems, low-latency or highly available applications. A database should accommodate itself to different data distributions, cluster topologies and hardware configurations. Data Placement.

Database 213

Database Latency C++ Scalability 213

All Things Distributed

JULY 13, 2010

Customers with complex computational workloads such as tightly coupled, parallel processes, or with applications that are very sensitive to network performance, can now achieve the same high compute and networking performance provided by custom-built infrastructure while benefiting from the elasticity, flexibility and cost advantages of Amazon EC2.

Cloud 119

Cloud AWS Automotive Latency 119

KeyCDN

JULY 11, 2019

Using an image CDN, such as KeyCDN, can significantly reduce the latency of your image delivery. This is crucial due to mobile devices requiring additional optimizations because they typically have less powerful hardware and a slower network connection when compared to desktop devices. Mbps ( Opensignal ).

Speed 52

Speed Website Mobile Latency 52

Alex Russell

APRIL 29, 2021

In almost every area, Apple's low-quality implementation of features WebKit already supports requires workarounds not necessary for Firefox (Gecko) or Chrome/Edge/Brave/Samsung Internet (Blink). Real-time network protocols for enabling videoconferencing, desktop sharing, and game streaming applications. is access to hardware devices.

Media 145

Media Games Education Engineering 145

Alex Russell

JULY 14, 2021

Modern browsers like Chrome and Samsung Internet support a long list of features that make web apps more powerful and keep users safer. On Android today and early iOS versions, WebViews allow embedders to observe and modify all network traffic (regardless of encryption). Hardware access APIs, notably: Geolocation. Web Bluetooth.

Google 121

Google Mobile Engineering Internet 121

Smashing Magazine

JANUARY 6, 2020

If you don’t have a device at hand, emulate mobile experience on desktop by testing on a throttled 3G network (e.g. To make the performance impact more visible, you could even introduce 2G Tuesdays or set up a throttled 3G/4G network in your office for faster testing. 300ms RTT, 1.6 Mbps down, 0.8

Performance 145

Performance Cache Network Metrics 145

Smashing Magazine

JANUARY 11, 2021

Networking, HTTP/2, HTTP/3 OCSP stapling, EV/DV certificates, packaging, IPv6, QUIC, HTTP/3. If you don’t have a device at hand, emulate mobile experience on desktop by testing on a throttled 3G network (e.g. Estimated Input Latency tells us if we are hitting that threshold, and ideally, it should be below 50ms. 300ms RTT, 1.6

Performance 145

Performance Cache Media Metrics 145

Smashing Magazine

JANUARY 7, 2019

If you don’t have a device at hand, emulate mobile experience on desktop by testing on a throttled network (e.g. To make the performance impact more visible, you could even introduce 2G Tuesdays or set up a throttled 3G network in your office for faster testing. 150ms RTT, 1.5 Mbps down, 0.7 Mbps up) with a throttled CPU (5× slowdown).

Performance 103

Performance Cache Metrics Network 103

Tim Kadlec

SEPTEMBER 25, 2014

It’s a device actively being sold by a few prepaid networks. Another interesting note was how significant the impact of hardware has on the timing. Those markets are going to dominate internet growth over the next few years, and affordability is a more necessary feature than a souped up device.

Latency 41

Latency Google Hardware Servers 41

Brendan Gregg

SEPTEMBER 5, 2017

Here's some output from my zfsdist tool, in bcc/BPF, which measures ZFS latency as a histogram on Linux: # zfsdist. Tracing ZFS operation latency. At my employer we sometimes use SR-IOV for direct network interface access, and NVMe for direct disk access. Hit Ctrl-C to end. ^C x series.

Virtualization 76

Virtualization AWS Engineering Hardware 76

Highly Scalable

AUGUST 20, 2013

This system has been designed to supplement and succeed the existing Hadoop-based system that had too high latency of data processing and too high maintenance costs. In the previous section, we noted that many distributed query processing algorithms resemble message passing networks. Pipelining. However, things are not so simple.

Big Data 154

Big Data Processing Lambda Database 154

All Things Distributed

APRIL 4, 2017

As well as AWS Regions, we also have 24 AWS Edge Network Locations in Europe. This enables customers to serve content to their end users with low latency, giving them the best application experience. In 2011, AWS opened a Point of Presence (PoP) in Stockholm to enable customers to serve content to their end users with low latency.

AWS 117

AWS Airlines Latency Games 117

Smashing Magazine

AUGUST 22, 2021

As we will see, QUIC and HTTP/3 indeed have great web performance potential, but mainly for users on slow networks. If your average visitor is on a fast cabled or cellular network, they probably won’t benefit from the new protocols all that much. Two-way latency is often called round-trip time (RTT).

Performance 105

Performance Network Latency Servers 105

All Things Distributed

NOVEMBER 15, 2010

For example, the most fundamental abstraction trade-off has always been latency versus throughput. Modern CPUs strongly favor lower latency of operations with clock cycles in the nanoseconds and we have built general purpose software architectures that can exploit these low latencies very well. General Purpose GPU programming.

AWS 136

AWS Latency Programming Architecture 136