The program of the Linux Piter 2016 conference, which will be held on November 11-12 this year in St. Petersburg, is 95% formed. This year, there are a lot more English speakers on Linux Piter than they did last year, and the topics are perhaps even more interesting.



Conference subsections are organized as follows:
')

• First of all, these are presentations that are somehow related to the Linux kernel - 5 pieces, including two reports on power management.
  
• Secondly, containers and clusters - 3 reports, including an overview presentation on light virtualization and a report on creating your own industrial PaaS on your knee from the improvised components.
  
• In the third - a couple of speeches about embedded systems with a detailed story about the device of a modern debugger.
  
• Next, 3 topics about networks, including an explanation on the fingers of the bottle what buffer bloat is and a story about the news in the TCP / IP stack.
  
• And at the end there is a little DevOps with tools for measuring system performance and admin panel.
  

Under the cut a little more ...

Linux Kernel

• “ So you want to write a Linux driver subsystem? ”- Michael Turquette (BayLibre, President and CEO, USA) from Los Angeles will talk about how to and how not to design and write the perfect framework for new driver subsystems in Linux. And generally how to write drivers correctly.
  
  
• “ Let's talk about NVMe over Fabric ” - Sergey Platonov (RAIDIX, Head of Department, St. Petersburg). A detailed presentation on the possibility of building SAN systems based on NVMe over RDMA, whose support appeared in the kernel 4.8. Comparison with traditional technologies, test results in numbers.
  
  
• “ The quest for low latency with block I / O ”. The report by Paolo Valente (Universita di Modena e Reggio Emilia, Assistant Professor, Emilia) from Italy is devoted to using the well-known BFQ I / O scheduler to overcome the I / O limitations of the blk-mq stack for some types of loads.
  
  
• “ Introduction to Power Managemen t”, again Michael Turquette. Mike has extensive experience with SoCs in Texas Instruments and in developing new smartphones, so we decided to ask him to read the report and on the state of affairs in modern power management, since this topic is familiar to him from top to bottom.
  
• “ PowerCI.org: the product development CI loop based on kernelci.org .” Paired to the previous report by Benoit Cousson (BayLibre, Co-Founder, President & CTO, Nice) Mike's colleagues from France, about how to build a system for measuring energy consumption and what to do with the resulting metrics.

Containers and Clusters

• “ These containers were given to you! ". Report by Pavel Emelyanov (Virtuozzo, Architect, Moscow) about the place of containerization in IT, about the "competitors" of this technology and what awaits it in the future.
  
  
• “ High Aavailability in Linux: Pacemaker problems and what we do with them .” Victoria Cherkalova's report (Dell EMC, Senior Software Engineer, St. Petersburg) will be about the problems associated with the automatic configuration and performance of the most common cluster manager, and how they can be circumvented.
  
  
• “ How to make PaaS using Docker, Consul and tying in Python ”. Presentation by Konstantin Nazarov (Parallels, Head of Infrastructure Team, Moscow) on how to make a small PaaS based on one person using Docker, Consul and Python as an orchestrator for one’s company.

Embedded

• The report of the engineer from ARM Pawel Moll (ARM, Principal Engineer, Cambridge UK) about Linux debuggers will be equally interesting for those who work with embedded systems and PCs. This is about how a modern debugger works and what mechanisms it uses.
  
  
• “ LEDE is the most widely available embedded Linux distribution .” Alexey Brodkin (Synopsys, Senior Software Engineer, St. Petersburg) will talk about the details of the configuration mechanisms and the initial load of one of the OpenWRT derivatives on various platforms.
  

Network

• “ How Linux beat Bufferbloat ” - Stephen Hemminger (Microsoft, Principal Software Architect, Portland USA). A fun presentation about what Bufferbloat is, literally - “buffer swelling” associated with excessive buffering in the network stack, and how it is struggled in the Linux kernel.
  
  
• “ Linux network report ” - Stephen Hemminger. Since Steven is the official Linux Linux bridging and iproute2 utilities and can easily tell about new features, drivers and new problems in the Linux TCP / IP stack, we asked to give his talk on this topic too.
  
  
• “ Tempesta FW: yet another Linux kernel Web accelerator ” - Alexander Krizhanovsky (Tempesta Technologies, CEO, Moscow). Presentation of a hybrid Web-accelerator and a firewall designed to filter large volumes of HTTP traffic. A story about why it is needed, how it was built and why other solutions are not suitable.
  

Storage

• “ Linux IO internals for database administrators ” - Ilya Kosmodemyansky (PostgreSQL-Consulting, CEO and consultant, Saarbrücken Germany). This time, Ilya will talk about how I / O works in Linux, how pages from databases travel from disk to shared memory and back, and what mechanisms exist to control this process.
  
  
• “ Experience of replacing XFS with BlueStore at Ceph ” - Maxim Vorontsov (Engineer, Moscow). Ceph BlueStore should already be stable, and in 2017 it should become a default backend, so Maxim has a story about the migration experience from FileSystem Backend to BlueStore.
  

Devops

• “ Application well supported in production ” - Nikolay Sivko (okmeter.io, Co-founder, Moscow). Talk about what to do to make the above application easy to support for people who have not written it. What is needed for the transparent work of the application in production? What is and is not enough in the known systems for collecting metrics.
  
  
• “ How devops exhaust themselves, and what will happen next ” - Cyril of the Evening (Jetware, CTO, St. Petersburg). A provocative title, and a report on what will happen next with DevOps, about the next generation of models for designing and operating server applications in public clouds and on classic servers.
  
  
• And finally, Alexander Chistyakov (DataArt, Chief Engineer, St. Petersburg) with traditional annealing!

There is still a month before the conference, and we still have a couple of free slots for super interesting reports. If your topic falls into the subject line (and especially if it is about networks and storage systems), then we will gladly accept your application for consideration.

Linux Piter is a conference about systems, platforms, and tools.

For students, a discount of 35% .