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

The Morning Paper

JUNE 13, 2019

As we saw with the SOAP paper last time out, even with a fixed model variant and hardware there are a lot of different ways to map a training workload over the available hardware. The following figure highlights how just one of these variables, batch size, impacts throughput and latency on ResNet50.

Hardware 58

Hardware Latency Serverless Energy 58

The Morning Paper

MAY 14, 2019

Last time around we looked at the DeathStarBench suite of microservices-based benchmark applications and learned that microservices systems can be especially latency sensitive, and that hotspots can propagate through a microservices architecture in interesting ways. on end-to-end latency) and less than 0.15% on throughput.

Big Data 45

Big Data Cloud Performance Hardware 45

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

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 40

Hardware Latency Virtualization Cloud 40

Abhishek Tiwari

JULY 12, 2018

Here are 8 fallacies of data pipeline The pipeline is reliable Topology is stateless Pipeline is infinitely scalable Processing latency is minimum Everything is observable There is no domino effect Pipeline is cost­-effective Data is homogeneous The pipeline is reliable The inconvenient truth is that pipeline is not reliable.

Latency 63

Latency Analytics Scalability Engineering 63

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 118

Cloud AWS Automotive Latency 118

CSS - Tricks

JULY 25, 2019

Lighthouse uses Chrome’s Remote Debugging Protocol to read network request information, measure JavaScript performance, observe accessibility standards and measure user-focused timing metrics like First Contentful Paint , Time to Interactive or Speed Index. Estimated Input Latency. Estimated Input Latency. Speed Index.

Google 71

Google Engineering Speed Mobile 71

Scalegrid

FEBRUARY 8, 2024

Key Takeaways Distributed storage systems benefit organizations by enhancing data availability, fault tolerance, and system scalability, leading to cost savings from reduced hardware needs, energy consumption, and personnel. This strategy reduces the volume needed during retrieval operations.

Storage 130

Storage Systems Big Data Azure 130

Scalegrid

DECEMBER 21, 2023

With its widespread use in modern application architectures, understanding the ins and outs of Redis® monitoring is essential for any tech professional. Identifying key Redis® metrics such as latency, CPU usage, and memory metrics is crucial for effective Redis monitoring. Redis®, a powerful in-memory data store, is no exception.

Strategy 130

Strategy Monitoring Latency DevOps 130

Abhishek Tiwari

DECEMBER 27, 2017

optimised container networking and security. For instance, scatter/gather pattern can be used to implement a MapReduce like batch processing architecture on top of Kubernetes. direct access to raw block storage [18] without the abstraction of a filesystem for workloads that require consistent I/O performance and low latency.

Big Data 63

Big Data Storage Benchmarking Hardware 63

The Morning Paper

NOVEMBER 10, 2019

Snap: a microkernel approach to host networking Marty et al., This paper describes the networking stack, Snap , that has been running in production at Google for the last three years+. The desire for CPU efficiency and lower latencies is easy to understand. SOSP’19. Emphasis mine). It reminds me of ZeroMQ.

Network 92

Network Transportation Latency Entertainment 92

The Morning Paper

MAY 12, 2019

An open-source benchmark suite for microservices and their hardware-software implications for cloud & edge systems Gan et al., A typical architecture diagram for one of these services looks like this: Suitably armed with a set of benchmark microservices applications, the investigation can begin! Hardware implications.

Open Source 61

Open Source Hardware Benchmarking Systems 61

Sutter's Mill

FEBRUARY 13, 2017

Tue-Thu Apr 25-27: High-Performance and Low-Latency C++ (Stockholm). On April 25-27, I’ll be in Stockholm (Kista) giving a three-day seminar on “High-Performance and Low-Latency C++.”

Latency 48

Latency C++ Hardware Performance 48

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. Marz, “Big Data Lambda Architecture”.

Big Data 154

Big Data Processing Lambda Database 154

Dotcom-Montior

NOVEMBER 24, 2021

There was a time when standing up a website or application was simple and straightforward and not the complex networks they are today. Gone are the days of monolithic architecture. These systems can include physical servers, containers, virtual machines, or even a device, or node, that connects and communicates with the network.

Systems 74

Systems Monitoring Hardware Network 74

All Things Distributed

MARCH 11, 2016

The expectation was that with each order or two of magnitude, we would need to revisit and revise the architecture to make sure we could address the issues of scale. We needed to build such an architecture that we could introduce new software components without taking the service down. The importance of the network.

AWS 144

AWS Hardware Retail Virtualization 144

The Morning Paper

MAY 19, 2019

Last week we learned about the [increased tail-latency sensitivity of microservices based applications with high RPC fan-outs. Seer uses estimates of queue depths to mitigate latency spikes on the order of 10-100ms, in conjunction with a cluster manager. So what we have here is a glimpse of the limits for low-latency RPCs under load.

Latency 80

Latency Hardware Network Architecture 80

All Things Distributed

JULY 20, 2015

Today, I want to explore the Amazon ECS architecture and what this architecture enables. The pool of resources, at this time, is the CPU, memory, and networking resources of Amazon EC2 instances as partitioned by containers. networks ports, memory, CPU, etc). Let’s talk about what Amazon ECS is actually doing.

Latency 166

Latency Architecture AWS Open Source 166

Percona

NOVEMBER 9, 2023

Kubernetes manages and orchestrates these containers, handling tasks such as deployment, scaling, load balancing, and networking. Your workloads, encapsulated in containers, can be deployed freely across different clouds or your own hardware. Just consider the sheer number of people who stream Netflix every night!

Efficiency 99

Efficiency Cloud Healthcare Open Source 99

The Netflix TechBlog

FEBRUARY 9, 2021

The homepage needs to load in a reasonable amount of time, even in poor network conditions. Server-generated assets, since client-side generation would require the retrieval of many individual images, which would increase latency and time-to-render. Here’s what the final architecture looked like.

Engineering 165

Engineering Storage Latency Entertainment 165

John McCalpin

AUGUST 1, 2018

According to Dr. Bandwidth, performance analysis has two recurring themes: How fast should this code (or “simple” variations on this code) run on this hardware? The user environment defines the mapping of MPI ranks to hardware resources (cores, sockets, nodes). The MPI runtime library. in ways that are seldom transparent.

Hardware 40

Hardware Transportation Performance Latency 40

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

SQL Performance

APRIL 27, 2020

Gen 5 is the primary hardware option now for most regions since Gen 4 is aging out. Hyperscale achieves high performance from each compute node having SSD-based caches which helps minimize the network round trips to fetch data. New Hardware Configuration for Provisioned Compute Tier. GB per vCore. Serverless Database.

Azure 55

Azure Database Serverless Hardware 55

The Morning Paper

APRIL 14, 2019

The paper sets out what we can do in software given today’s hardware, and along the way also highlights areas where cooperation from hardware will be needed in the future. There are also stateless interconnects including buses and on-chip networks. Requirement 4 : State flushing must be padded to its worst-case latency.

Hardware 44

Hardware Cache Latency Speed 44

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

Percona

APRIL 17, 2023

The swap issue is explained in the excellent article by Jeremy Cole at the Swap Insanity and NUMA Architecture. The CFQ works well for many general use cases but lacks latency guarantees. The deadline excels at latency-sensitive use cases ( like databases ), and noop is closer to no schedule at all.

Best Practices 52

Best Practices Design Tuning Database 52

The Morning Paper

MAY 28, 2019

The goal is to produce a low-energy hardware classifier for embedded applications doing local processing of sensor data. The resulting system can integrate seamlessly into a scikit-learn based development process, and dramatically reduces the total energy usage required for classification with very low latency. Introducing race logic.

Energy 52

Energy Hardware Efficiency Architecture 52

Abhishek Tiwari

DECEMBER 27, 2017

In 2018, we anticipate that ETL will either lose relevance or the ETL process will disintegrate and be consumed by new data architectures. Unified data management architecture. Apache Arrow's in-memory columnar layout is specifically optimized for data locality for better performance on modern hardware like CPUs and GPUs.

Big Data 49

Big Data Artificial Intelligence Storage Hardware 49

All Things Distributed

NOVEMBER 15, 2010

Building general purpose architectures has always been hard; there are often so many conflicting requirements that you cannot derive an architecture that will serve all, so we have often ended up focusing on one side of the requirements that allow you to serve that area really well. General Purpose GPU programming.

AWS 136

AWS Latency Programming Architecture 136

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. One distinct trend is a belief that a JavaScript framework and Single-Page Architecture (SPA) is a must for PWA development. The median user is on a slow network.

Performance 68

Performance Benchmarking Network Mobile 68

O'Reilly Software

APRIL 4, 2018

Considerations for setting the architectural foundations for a fast data platform. Google was among the pioneers that created “web scale” architectures to analyze the massive data sets that resulted from “crawling” the web that gave birth to Apache Hadoop, MapReduce, and NoSQL databases. Back in the days of Web 1.0,

Architecture 80

Architecture Scalability Google Operating System 80

SQL Server According to Bob

SEPTEMBER 26, 2016

When we released Always On Availability Groups in SQL Server 2012 as a new and powerful way to achieve high availability, hardware environments included NUMA machines with low-end multi-core processors and SATA and SAN drives for storage (some SSDs). As we moved towards SQL Server 2014, the pace of hardware accelerated.

Availability 40

Availability Servers Hardware Benchmarking 40

The Morning Paper

SEPTEMBER 12, 2019

SQL provides a declarative programming interface, below which the system itself can figure out the most effective execution plans based on data size and statistics, layout, compute hardware etc. For the feed-forward neural network, TensorFlow using GPUs is still significantly faster with smaller hidden layers, but cannot scale.

Network 44

Network Database Programming Hardware 44

SQL Server According to Bob

JANUARY 11, 2020

Many high-end disk subsystems provide high-speed cache facilities to reduce the latency of read and write operations. Example 1: ​​ Hardware failure (CPU board) Battery backup on the caching controller maintained the data. Important Always consult with your hardware manufacturer for proper stable media strategies.

Servers 40

Servers Cache Media Hardware 40

Alex Russell

JULY 14, 2021

Because they lean on the system-provided WebView component, they do not need to pay the expense of a heavier app download to support rendering HTML, running JavaScript, decoding images, or loading network resources. Hardware access APIs, notably: Geolocation. Basic navigation and window management features to (e.g. Web Bluetooth.

Google 121

Google Mobile Engineering Internet 121