Directory of RSS feeds
Statistics

RSS feeds in the directory: 2797

Added today: 0

Added yesterday: 0

Hi-Tech / Other

Habrahabr / Best publications per day

Best publications in the last 24 hours

Other RSS feeds of this site

Habrahabr / Interesting / topical publications

Found 3034 items

In the last post, I talked about Kubernetes, about how ThoughtSpot uses it for their own needs to support development. Today I would like to continue the conversation about the short but nevertheless interesting story of debugging, which occurred recently. The article is based on the fact that containerization != virtualization. Besides shows how containerization processes competing for resources even under optimal restrictions on cgroup and high performance cars. Read more →

( Illustration ) Senior web developer s Anton and Alex continue the story of the difficult fight against Nuxt. In the previous round of battle with this framework they showed how to run the project on Nuxt so everyone was happy. In a new article, let's talk about the real application of the framework. We started to rewrite a project with huge technical debt. Monthly audience amounted to 6-7 million unique visitors, but the existing platform has taken too many problems. It was therefore decided to retire her. Of course, the performance was our biggest fear, but I also don't want to sink on SEO. After a couple of rounds of discussion decided not to rely on the traditional approach with server-only rendering and not drive themselves into the trap client-side...

In the heart of information retrieval systems and Meltwater Fairhair.ai works a set of Elasticsearch clusters with billions of articles from the media and social media. the Index shards to the clusters differ in the structure of access, workload and size, which raises some really interesting issues. In this article we will explain how to apply linear programming (linear optimization) to maximize uniform distribution of the workload of search and indexing for all nodes in the cluster. This decision reduces the likelihood that one node will become a bottleneck in the system. As a result, we increased the search speed and save on infrastructure. Read more →

Elbrus Processors reached capacity, which allows their use in developing supercomputers, according to the developers of the MCST. Previously, domestic CPU was used only in desktop computers and servers, was built only a few small systems the size of a closet. However, the new project with a budget of more than 1 billion rubles — this is a supercomputer, which probably will be included in the list of Top 500 most powerful supercomputers in the world, writes "Kommersant". photo: IC CPU 1891VM12Ya — calculator server class with an advanced set of vector commands. Contains 8 cores of the architecture of "Elbrus" 5-th generation with clock rate of 1500 MHz. Allows you to build multiprocessor servers and workstations, as well as on-Board computers Read more →

I don't like to read book ratings for two reasons. First, they often represent a list of books, selected by unknown author by unknown criteria. Second, descriptions of books are more like advertising copy publishers that are hard to believe. because Of this, most of these materials are of little use, despite the fact that can contain explanatory books. I have long wanted to write a useful review, which will not impose certain material, and will allow the reader to choose the most appropriate. Read more →

Approx. transl.: This article continues the series of materials from technical writer from Google, working on the documentation for Kubernetes (Andrew Chen), and Director of software engineering SAP (Dominik Tornow). Their goal is available and clearly explain the basis for the organization of Kubernetes. Last time we translated an article about high availability and now we will talk about such a basic concept in Kubernetes as pod. Kubernetes — engine orchestration of containers created to run containerized applications on multiple nodes, usually called a cluster. In these publications, we use the approach of system modeling to improve understanding of Kubernetes and its underlying concepts. It is recommended that readers already have a basic understanding...

I'd like to tell you about the GPU of the console Nintendo DS and its differences from the modern GPU, as well as to Express their opinion about why the use Vulkan instead of OpenGL in the emulator will not bring any benefits. I don't really know Vulkan, but from reading I understand that Vulkan is different from OpenGL that operates at a lower level, allowing programmers to manage GPU memory and things like that. It may be useful to emulate more modern consoles that use proprietary graphics API that provides levels of control that are not available in OpenGL. Read more →

Yesterday the company "Yandex" has launched the project "Yandex.Radar" — a ranking of the most popular resources in Runet, which is built on mathematical models based on aggregated data "Yandex.Browser." If desired, sites can connect data from "Yandex.Metrics" to add in the statistics of the demographics of the audience etc. the New service will not please everyone. The main competitor of Yandex in the Russian segment Mail.Ru immediately declared mistrust to this metric, and was asked to remove from the ranking of their sites. Mail.Ru believes that "Yandex" he is a player of the advertising market, and therefore it is incorrect to maintain such a rating. In addition, he obviously did not accurately reflect the audience of the sites, because it relies on...

Some time ago (autumn 2016), the development of the next version of the technological platform 1C:Enterprise inside the development team raised the question of support for the new C++14 standard in our code. The transition to the new standard, as we have suggested, would allow us to write many things more elegant, simpler and more reliable, simplified support and maintenance of the code. And the translation seems to be nothing extraordinary, if not for the scale code base and specific features in our code. For those who don't know, 1C:Enterprise is an environment for rapid development of cross-platform business applications and runtime to run on different operating systems and DBMS. In General terms in the product are: Cluster of application servers running...

Engineers build a prototype of a robotic factory that will create water, oxygen and fuel on the surface of Mars the Martians: the representation of the artist about robots, excavators, which can earn one day on Mars, long before the first people set foot on the planet 2038. After 18 months living and working on the surface of Mars a team of six researchers is immersed in a space transport rocket and sent to Earth. People do not remain, but the work continues without them: Autonomous robots will continue mining and chemical synthesis, begun by them for many years before the first manned mission fell on the planet. The factory produces water, oxygen and gas fuel, using local resources, and it will methodically create all needed materials for the next Mars...

Is a network of approximately 90 major stores in Russia. Each store will very on the tape library (in the photo below — ZIP). Then they take the tapes and taking them by car to the archive. the mechanical Device: they break, fail, we go to fix it. Then they go with the extended warranty, and it's all angry. At some point they become obsolete. But the budget was exactly on the new version of the tape library. At this point, the customer appeared on the verge of a significant amount and asked if it is possible to think of something within it. We had thought about Central installation of one large glands, but the situation was complicated by the fact that the channels from the shops is limited to 5 Mbit/s (the most distant). Read more →

As you know, HTTP 1.1 is a text transmission Protocol. HTTP messages are encoded using ISO-8859-1 (which conditionally can be considered the extended version of ASCII that contains umlauts, diacritics and other characters used in Western European languages). In the body of the message, you can use a different encoding which should be indicated in the header "Content-Type". But what if we need to specify non-ASCII characters not in the message body and the headers? Probably the most common case is putting the file name in "Content-Disposition" header. This is seemingly a pretty common task, but its implementation is not so obvious. TL;DR: Use the encoding defined in RFC 6266 for the "Content-Disposition" header and convert the text into the Latin alphabet (a...

Translator: I am neither a professional programmer nor a professional translator, but the appearance of the described tool from the Creator of the popular libraries D3.js made a strong impression on me. was surprised to find that on habré, and generally in the Russian-speaking Internet, more than a year is unfair to ignore this tool. So I decided that is simply obliged to contribute to the development of arts programming, JavaScript in particular. Meet d3.express, integrated research environment. (31 January 2018 d3.express called Observable and lives on beta.observablehq.com) If you've ever had a rebate on my code or to understand someone else's, then you are not alone. This article is for you. the Last eight years I have developed tools for visualizing...

Participants Saint TeamLead Conf called the report of Alexander Sisi one of the best probably because of communication skills team-lead depends on many things, and they developed, as a rule, not very good. the Story will consist of four conceptual blocks: About communication. Will get to that, what is communication, what is the main problem with communication, why is it so many say and write. All scientists are philosophers, beginning with Aristotle's try to solve this problem, but the final decision to "take and make" is still there. high performance communications: 4 types of positional communication. This part is devoted to technical issues related to the development of highly effective communication. Roughly speaking what to do in a particular situation...

The Rating of calls is a key part of quality control for call centers. It allows organizations to fine-tune the workflow, so operators can do the job faster and more efficiently, and avoid meaningless routine. bearing in mind that the call center must be effective, we worked on the automation of the assessment calls. In the end, we came up with an algorithm that handles the calls and distributes them into two groups: the suspicious and neutral. All suspicious calls immediately went to the team of quality assessment. Read more →