The Perdition and nginx IMAP Proxies (Jan-Piet Mens)
Abstract: The Perdition and nginx (Engine X) IMAP/POP-3 proxies enable clients to be directed to arbitrary back-end IMAP/POP-3 servers depending on the result of a database lookup. We discuss why you would want to implement such a proxy, what these tools can do for you, and show you how to configure the tools to suit a variety of situations.
Bio: Jan-Piet is the author of Alternative DNS Servers, a 700+ page book discussing choice and deployment, and optional SQL/LDAP Back-Ends in sundry Open Source DNS servers, and he has authored different technical publications.
Designing and deploying a manageable virtual desktop infrastructure (Mike Banahan)
Abstract: Partly by accident and mostly by design, Cutter Project has constructed a rapidly deployable and highly manageable desktop infrastructure.
Based around thin client technology and remote access facilities, centralised servers can deliver multiple different operating system and software environments to end users who are not tied to software installed on classical desktop PCs.
No great claims are made for any single step of earth-shattering innovation. Instead, common sense ideas and some ruthlessness are proven to work in producing very much lowered cost of deployment and maintenance, including the ability to perform near 100% remote maintenance and management of centralised services.
As an example, an outpost of the UK Foreign and Commonwealth office was remotely provisioned and deployed then remotely managed from some 5,000 miles distant without requiring any site visits by technical or commercial staff (though for other reasons, such a visit was indeed made).
What is valuable about the tactics and approaches covered in this talk is that they have been thought through and there are numerous reference sites in UK schools and universities where the ideas can be seen being put into practice. A saving of twenty five to one in maintenance engineering effort is argued with some confidence. Other substantial cost savings also accrue.
This talk covers the architecture and design decisions taken. It will also discuss areas where thin clients are very much not the correct solution.
Bio: Mike is an old lag from the early days of Unix, still having a 9-track tape on his shelf labelled 'Bell Systems Labs, Unix System, 7th Edition'. A former committee member of UKUUG he sold his soul to the dark side many years ago and has run consulting and training companies for a long time. Cutter Project is a hardware and support operation focusing on low-maintenance thin client deployment.
Simplifying software appliance creation with SUSE Studio (Matt Barringer)
Abstract: Today the Linux world is dominated by one-size-fits-all Linux distributions that include thousands of packages and are often several gigabytes once fully installed.
Matt will introduce the benefits of streamlined, single-purpose software appliances based on Linux. This can be in the form of virtual machine images, bootable and installable media, or live USB keys. He will also show how quick and easy it is to create and share such appliances with SUSE Studio (http://susestudio.com), all without ever leaving your web browser! SUSE Studio is available publicly and is completely free of charge.
Bio: Matt Barringer is a Senior Software Engineer with SUSE/Novell who has worked on SUSE Studio since it was a prototype, developing the back-end components dealing with packages, package repositories, and dependency resolution.
PostgreSQL High Availability & Scalability (Simon Riggs)
Abstract: PostgreSQL is the World's Most Advanced Open Source Database, with a variety of offerings to support High Availability, Scalability and enterprise-class performance. This talk will explain the various features in current Postgres (8.4), and explore the major new features coming in our next major release in mid-2010.
Bio: Simon Riggs is a PostgreSQL developer, consultant and advocate, having worked on major features for the last 6 releases.
Building an infrastructure for open source development (Simon Wilkinson)
Abstract: This talk will discuss the challenges of building a development infrastructure for a large Open Source software project. We'll look at issues of revision control systems, code review tools, and code verification and testing. Through the example of OpenAFS, we'll look at how one large project has moved it's entire repository to git, their use of gerrit for code review, and how continuous build verification is performed with buildbot.
We'll then example how the same tools can be adopted on a smaller scale, using as examples a number of projects at the School of Informatics, where small local development teams are benefiting from the same technologies.
Bio: Simon has presented at a number of UKUUG conferences, in particular on Kerberos, OpenAFS, and identity management issues. He designed, and implemented OpenAFS's new development infrastructure, including performing their conversion from CVS to git, and developed the integration tools that tie their environment together. Simon is a Systems Architecture Specialist for the School of Informatics at the University of Edinburgh, who kindly allow him to work on a variety of open source projects.
GNUBatch (John Collins)
Abstract: GNUbatch is a networked batch scheduling system enabling sequences of jobs to be synchronised and interlocked across multiple platforms under the control of dependency variables.
A wide variety of interfaces are available.
The product was adopted by GNU in February 2008 after development over two decades by the author.
The talk will present an overview of GNUbatch and discuss the features of the next release currently under development.
Bio: John Collins, after graduating from Cambridge in 1973 has worked in UNIX since 1980 and Linux since about 1998. He's run Xi Software since 1986 which specialises in UNIX Systems Administration Tools.
Security for the virtual datacentres (Sas Mihindu)
Abstract: In recent times, many technologists have focused on building datacentre environments based on the virtualisation technologies which allow and isolate multiple users while satisfying the security requirements of each user. Machine virtualisation alone provides immediate isolation of computing resources such as memory and CPU between guest domains. However, the network remains a shared resource as all traffic eventually passes through a shared network resource and ends up on a shared physical medium. As a result, we need mechanisms to (a) control the information flow between virtual machines, (b) configure virtual and physical network resources, and (c) separate network resources used by each networking domain. Having said that, securing the access to data on persistent media and transfers over the network are serious problems in distributed virtual datacentres and cloud computing scenarios.
This paper compares the security requirements of virtual networking and the current approaches in terms of their capability to enforce and support domain policies. It describes how each approach reinforces network separation and isolation within and across network domains. Newer solutions have been developed which rely on VLAN separation and address mapping techniques to control the information flow within and across networking domains. Similar solutions and also the implementation of distributed virtual gateways to (re)direct traffic within and between networking domains are discussed. Also, a framework based on Trusted Virtual Domains and Trusted Computing for secure network and storage virtualisation which include mechanisms for verifying the integrity of VMs and security policy enforcement points are detailed. This will facilitate the provision of so called Trusted Virtual Domains that provide a virtual environment for each user within the virtual datacentre.
Bio: The author has developed skills and experience by working for academic, research and business enterprises in Europe, and Australasia over two decades and is the founding member of UK’s first ‘BIM Research & Special Interest Group’. His current research interests are human-centric technology infrastructures for future workplaces (FWS), strategies of distributed knowledge and data management for community collaboration and virtual reality, virtual environments and virtual technology infrastructures. He has been leading and delivering for many European Commission projects with Information Society Technologies (IST) Programmes via University of Salford, Greater Manchester. He has recently been engaged in the design and deployment of web infrastructures for technological observatory of VR, VE and FWS, and Strategic Knowledge Management infrastructures for scientific and professional communities. In addition to this he has also written many research papers on Knowledge Engineering, Communication and Collaboration Tools and Distributed Virtual Networking infrastructures for collaboration. One of his future research goals is to establish a framework for Virtual Technology Infrastructures to support effective communication and collaboration for virtual communities.
21st Century Systems Perl (Matt S Trout)
Abstract: Over the past few years, the state of the art in Perl application development has advanced by leaps and bounds, notably as a result of the Catalyst MVC web framework and the Moose metaprotocol and object system.
However, systems scripting has a different set of priorities to applications development - ease of modification by a sysadmin not hugely familiar with the language is essential, as is fast startup time for usage in shell scripts, and deployment must be made as trivial as possible.
In this talk, I intend to discuss how the same principles that enable the new perl enlightenment in applications development best practices can be repurposed in a way that both makes perl more of a super-shell than ever before - but also to allow easy and rapid development of complex, reliable utilities in a manner still consistent with the requirements of systems administration tools.
Coherent and Integrated configuration of Virtual Infrastructures (Panagiotis Kritikakos)
Abstract: Although virtualisation provides both flexible, easy deployment of systems and also server consolidation, its adoption to provide infrastructure services does not completely eliminate problems in system configuration. On the contrary, it multiplies them, by increasing the complexity of configuration and management. Separating system and application configuration, for the physical machines and their virtual guests, raises its own issues – caused by the lack of a common framework that efficiently integrates these different configuration aspects.
Such systems require configuration of both their "outside" (hardware) and "inside" (software) aspects including the virtualisation aspects, that may differ in each system. Other major issues that arise are: How configuration of a virtual infrastructure can be integrated with the configuration and management of the existing physical infrastructure; and if the configuration framework should differ from an administrator's point of view.
In this paper we present the use of the lcfg-xen component of the LCFG framework to configure and manage Xen host systems and their virtual guests. We discuss automatic installation and centralised configuration of systems and applications - for both the physical and Virtual Machines - using a common descriptive configuration language. The flexibility of LCFG enables us to manage the "inside" and "outside" of a Virtual Machine in the same way as we would with a purely physical machine; and further, allows us to integrate these with an existing managed heterogenous infrastructure.
Bio: Panagiotis works as a Computing Officer at EPCC, at the University of Edinburgh. He is involved with configuring and administrating infrastructure services on managed Linux servers, as well as configuring and deploying managed Linux desktops. He has also been involved in European projects such as OMII-EU, DEISA2 and HPC-Europa2. He holds a recent Bachelor of Engineering degree on Internet Computing from Edinburgh Napier University. Morover, Panagiotis is a member of the Greek FreeBSD Documentation Project translation team and has been the moderator of the Greek Slackware Linux documentation translation team.
MySQL HA with pacemaker (Kris Buytaert)
Abstract: Database High Availability is often the core component to build a highly available infrastructure.
This presentation will guide you trough the different options available with their advantages and disadvantages when choosing a method to setup a Highly available MySQL setup.
We'll cover MySQL Cluster, MySQL DRBD, and MultiMaster based HA setups and we'll go in depth discussing the setup and integration with Pacemaker the successor of the Heartbeat framework within the Linux-HA project.
Bio: Kris Buytaert is a long time Linux and Open Source Consultant doing Linux and Open Source projects in Belgium , Europe and the rest of the universe. He is currently working for Inuits.
Kris is the Co-Author of Virtualization with Xen, used to be the maintainer of the openMosix HOWTO and author of different technical publications. He is a frequent speaker at different international conferences.
He spends most of his time working on Linux Clustering (both High Availability, Scalability and HPC), Virtualisation and Large Infrastructure Management projects hence trying to build infrastructures that can survive the 10th floor test, better known today as the cloud.
His blog titled "Everything is a Freaking DNS Problem" can be found at http://www.krisbuytaert.be/blog/
Hudson hit my puppet with a cucumber (Patrick Debois and Julian Simpson)
Abstract: What happens when Continuous Integration meets Puppet? In this session Patrick Debois and Julian Simpson will demonstrate Test Driven Development of Puppet modules using Cucumber, Hudson and VirtualBox. We'll start with a very simple example of a puppet manifest and a Cucumber scenario that will demonstrate that some of the functionality works.
We'll iterate several times to add functionality as we test drive our custom puppet module to install the Apache web server and help us automate configuration items like virtual hosts, modules, users and custom behaviour. As we add functionality we'll write tests and then make them pass.
We'll use Continuous Integration to demonstrate that we haven't regressed or introduced other issues along the way and at the end we'll show you how it can be used to do cross-platform testing.
You'll learn how to borrow tools from developers and put them to good use in systems administration.
Situation Normal, Everything Must Change (Simon Wardley)
Abstract: In today’s computing world, it can often feel like we are drowning in wave after wave of new trends. This sea of concepts are simply the evolution of our industry from a product to a service based economy.
This talk will examine the evolution of technology and why cloud is happening now. We will then examine the risks & benefits of cloud, the management challenges this brings and the common misconceptions.
We'll then examine the whys and hows of building your own cloud before using all these concepts to provide some predictions for the future.
Bio: As a geneticist with a love of mathematics and a fascination in economics, Simon has always found himself dealing with complex systems, whether it’s in behavioural patterns, environmental risks of chemical pollution, developing novel computer systems or managing companies. These days Simon works as the Software Services Manager for Canonical, helping define future cloud computing strategies for Ubuntu. He is a passionate advocate and researcher in the fields of open source, commoditisation, innovation and cybernetics . He is also fond of ducks. As he says “they’re fowl but not through choice”.
Libferris teaching filesystems new tricks (Ben Martin)
Abstract: Libferris allows you to see many diverse data sources like XML, PostgreSQL, ISAM files, and Web services through the same filesystem interface. Any libferris filesystem can be indexed and searched, and any filesystem can be seen from XQuery, SQLite, or as a virtual DOM.
If updating a cell in a Google spreadsheet from a PostgreSQL table, 'cp'ing a buffer from emacs into a value in a Berkely DB4 file, or drag and dropping a video from your mobile phone over bluetooth to youtube/vimeo are your idea of filesystem interaction this is the talk for you. On the other hand you might like to rsync an XML file into a DB4 file to improve query performance or mount a CSV or Apache log file and get at it from an SQLite prompt? This talk will cover examples of these things and more.
Libferris is pacakged for both Fedora and maemo.
Bio: Ben Martin has been hacking on libferris for around ten years. He has a BIT, MIT, and PhD in computer science. His PhD was applied research on creating search and browse spaces using finite lattice theory. When not tinkering on open source for fun and profitWW, Ben likes to write about Linux and Open source.
C Language Variants - An Overview (Donald Forbes)
Abstract: A brief history of the C language will be presented, with a fast forward to present tense. Ideas for better expressiveness will be discussed; the key appears to be good use of compiler frameworks. References will be made to the development of languages in general and then developed into a live discussion of concepts as to how best to go forward.
Detailed discussion will be developed from this to include the development of object oriented systems based on the C language and the constructs they use to acheve richness of expression.
There are several main streams of interest; the C++ and Java developments, with the Objective-C style as a variant. It is also possible to treat C as a portable assembly language and this leads straight into the business of compiler frameworks. It is proposed to discuss the relationship between simple expressions in a C language style (albeit in some other language) and the capability they deliver.
Detailed discussion can be developed from this to include the relationship between time dependencies and language constructs (most relevant to hardware design) and separately the expressiveness of languages and their complexity. This leads into the concept of 'semantic richness' with its corollary ambiguities (otherwise known as 'compiler headaches'.
A (brief) review of news for compiler frameworks yields the result that they have a few development areas, but relatively little activity as viewed from the perspective of their application to improved code in production environments.
The State of Linux Installs (Donald Forbes)
Abstract: Several linux installs will be available for live experiment; there will also be documentation as to the incompletenesses (or, by contrast, completenesses) of as many mainstream Linux systems as can be reviewed; the objective is to have useful comment on the 10 most popular, with at least one or two others referred to.
There will be two tracks of discussion 'what happens when I do the basic install' and 'what happens when I try to do something useful with it?.
Both these topics can give rise to pleasure or alarm; sometimes in equal measure. It is not the case that all systems are equally easy to install, nor are given installs necessarily very complete. Both of these require experience and pointers will be available for further development.
Ubuntu Enterprise Cloud: From bare metal to cloud in 30 minutes (Thierry Carrez)
Abstract: A technical overview of the Ubuntu Enterprise Cloud software stack, parallelized with a live installation that will show that your own private cloud is just a few minutes and a few questions away ! This talk will tackle themes like deployment topologies, new features coming up, technical challenges for a Linux distribution in delivering them, and future developments.
Bio: Thierry Carrez is the Technical Lead for the Ubuntu Server team. He works for Canonical and is home-based in a small village in France. In what seems like a previous life, he used to be an IT Manager for small and large companies.
Ubuntu Server 10.04 LTS: The making-of (Thierry Carrez)
Abstract: A sneak peak into the process that ultimately delivers the Ubuntu Server product. We'll discuss the Ubuntu development cycle and how the long-term-support perspective affects it. We'll follow up with an overview of the upcoming new features in 10.04. Finally we'll tackle the issue of community pre-release testing, a few days after the release of the Beta1 milestone.
Bio: Thierry Carrez is the Technical Lead for the Ubuntu Server team. He works for Canonical and is home-based in a small village in France. In what seems like a previous life, he used to be an IT Manager for small and large companies.
Herding a highly available MySQL cluster at Google scale (Andrew Stribblehill)
Abstract: Google has a large partitioned and globally replicated business-critical database. We discuss what procedures and technologies we use to keep it running stably and efficiently, including QoS-based load balancing, automated server replacement, semi-synchronous replication, monitoring, on-server tweaks and an awful lot of love and care.
Bio: Andrew Stribblehill is a Google Site Reliability Engineer responsible for production services, including a large business-critical MySQL database. He used to manage Durham University's central web, MySQL and HPC services. Before that, he studied physics at Durham and consequently knows way too much about ancient Babylonian astronomy. He is a coffee and whisky fan, though not in parallel.
The New VVorld (Bjoern Zeeb and Robert Watson)
Abstract: For over a decade, FreeBSD has supported lightweight OS virtualization through the Jail framework. The jail approach allows a single kernel to share multiple user space installations, each with their own root user and management environment. Jail has been widely adopted by Internet service providers to achieve virtual machine densities of hundreds or thousands of virtual instance, per server, but also for enterprise service virtualisation. However, a key limitation of the Jail technology has been the one-IP-address-per-Jail constraint, a property of how Jail was integrated with the FreeBSD network stack. This talk describes exciting on-going work to virtualise the FreeBSD network stack, giving each Jail to operate with its own virtual network interfaces, firewalls, traffic management, IPsec stack and forwarding tables, and much more.
What makes FreeBSD's virtualisation framework different from other approaches, such as fully virtualised hardware, and when is it appropriate? What are the design principles and how are things implemented? How does it scale? Why could this approach be interesting to people doing R&D, Internet Service Providers, and other businesses? What interesting things can be done using the framework? How could you extend the framework for your needs, and what are our future plans?
This presentation will attempt to answer these questions to help you evaluate how to use FreeBSD jail-based virtualisation to save time and resources for your business or research.
Video Streaming with Red5 (Tim Williams)
Abstract: Red5 is a java based flash video and communication server. I will discuss the capabilities of the Red5 server, move on to cover Application development and finish with a practical demonstration of a Red5 based video broadcast system designed to work with the Moodle eLearning platform.
Bio: Tim Williams works for EuroMotor-AutoTrain LLP, an eLearning technology development company started as a spin-off from the University of Birmingham in 2006. Tim has been working on eLeaning development projects for over 10 years and has been using and developing eLearning solutions using video streaming for much of that time.
More than just profit: GNU/Linux in Social Enterprise (Richard Smedley)
Abstract: Everybody uses Windows, and charities don't want to be second best with all that free stuff. The advantages of GNU/Linux for the voluntary sector may be obvious for Unix admins, but it's uphill work convincing decision makers to make that choice. Find out why it really matters, and just whay you can do to help.
Bio: Richard Smedley is a social entrepreneur working with and for third sector groups involved in environmental education and social change. A longtime advocate of software freedom, he is also writing a series of introductory GNU/Linux articles for Linux User magazine.