7 communities driving open source development

Not so long ago, the open source model was the rebellious kid on the block, viewed with suspicion by established industry players. Today, open initiatives and foundations are flourishing with long lists of vendor committers who see the model as a key to innovation.

Open Development of Tech Drives Innovation
Over the past two decades, open development of technology has come to be seen as a key to driving innovation. Even companies that once saw open source as a threat have come around — Microsoft, for example, is now active in a number of open source initiatives. To date, most open development has focused on software. But even that is changing as communities have begun to coalesce around open hardware initiatives. Here are seven organizations that are successfully promoting and developing open technologies, both hardware and software.

OpenPOWER Foundation
The OpenPOWER Foundation was founded by IBM, Google, Mellanox, Tyan and NVIDIA in 2013 to drive open collaboration hardware development in the same spirit as the open source software development which has found fertile ground in the past two decades.

IBM seeded the foundation by opening up its Power-based hardware and software technologies, offering licenses to use Power IP in independent hardware products. More than 70 members now work together to create custom open servers, components and software for Linux-based data centers.

In April, OpenPOWER unveiled a technology roadmap based on new POWER8 process-based servers capable of analyzing data 50 times faster than the latest x86-based systems. In July, IBM and Google released a firmware stack. October saw the availability of NVIDIA GPU accelerated POWER8 systems and the first OpenPOWER reference server from Tyan.

The Linux Foundation
Founded in 2000, The Linux Foundation is now the host for the largest open source, collaborative development effort in history, with more than 180 corporate members and many individual and student members. It sponsors the work of key Linux developers and promotes, protects and advances the Linux operating system and collaborative software development.

Some of its most successful collaborative projects include Code Aurora Forum (a consortium of companies with projects serving the mobile wireless industry), MeeGo (a project to build a Linux kernel-based operating system for mobile devices and IVI) and the Open Virtualization Alliance (which fosters the adoption of free and open source software virtualization solutions).

Open Virtualization Alliance
The Open Virtualization Alliance (OVA) exists to foster the adoption of free and open source software virtualization solutions like Kernel-based Virtual Machine (KVM) through use cases and support for the development of interoperable common interfaces and APIs. KVM turns the Linux kernel into a hypervisor.

Today, KVM is the most commonly used hypervisor with OpenStack.

The OpenStack Foundation
Originally launched as an Infrastructure-as-a-Service (IaaS) product by NASA and Rackspace hosting in 2010, the OpenStack Foundation has become the home for one of the biggest open source projects around. It boasts more than 200 member companies, including AT&T, AMD, Avaya, Canonical, Cisco, Dell and HP.

Organized around a six-month release cycle, the foundation's OpenStack projects are developed to control pools of processing, storage and networking resources through a data center — all managed or provisioned through a Web-based dashboard, command-line tools or a RESTful API. So far, the collaborative development supported by the foundation has resulted in the creation of OpenStack components including OpenStack Compute (a cloud computing fabric controller that is the main part of an IaaS system), OpenStack Networking (a system for managing networks and IP addresses) and OpenStack Object Storage (a scalable redundant storage system).

OpenDaylight
Another collaborative project to come out of the Linux Foundation, OpenDaylight is a joint initiative of industry vendors, like Dell, HP, Oracle and Avaya founded in April 2013. Its mandate is the creation of a community-led, open, industry-supported framework consisting of code and blueprints for Software-Defined Networking (SDN). The idea is to provide a fully functional SDN platform that can be deployed directly, without requiring other components, though vendors can offer add-ons and enhancements.

Apache Software Foundation
The Apache Software Foundation (ASF) is home to nearly 150 top level projects ranging from open source enterprise automation software to a whole ecosystem of distributed computing projects related to Apache Hadoop. These projects deliver enterprise-grade, freely available software products, while the Apache License is intended to make it easy for users, whether commercial or individual, to deploy Apache products.

ASF was incorporated in 1999 as a membership-based, not-for-profit corporation with meritocracy at its heart — to become a member you must first be actively contributing to one or more of the foundation's collaborative projects.

Open Compute Project
An outgrowth of Facebook's redesign of its Oregon data center, the Open Compute Project (OCP) aims to develop open hardware solutions for data centers. The OCP is an initiative made up of cheap, vanity-free servers, modular I/O storage for Open Rack (a rack standard designed for data centers to integrate the rack into the data center infrastructure) and a relatively "green" data center design.

OCP board members include representatives from Facebook, Intel, Goldman Sachs, Rackspace and Microsoft.

OCP recently announced two options for licensing: an Apache 2.0-like license that allows for derivative works and a more prescriptive license that encourages changes to be rolled back into the original software.

---

Best Microsoft MCTS Certification, Microsoft MCITP Training at certkingdom.com

7 killer open source monitoring tools

Network and system monitoring is a broad category. There are solutions that monitor for the proper operation of servers, network gear, and applications, and there are solutions that track the performance of those systems and devices, providing trending and analysis. Some tools will sound alarms and notifications when problems are detected, while others will even trigger actions to run when alarms sound. Here is a collection of open source solutions that aim to provide some or all of these capabilities.

Cacti
Cacti is a very extensive performance graphing and trending tool that can be used to track just about any monitored metric that can be plotted on a graph. From disk utilization to fan speeds in a power supply, if it can be monitored, Cacti can track it -- and make that data quickly available.

Nagios
Nagios is the old guard of system and network monitoring. It is fast, reliable, and extremely customizable. Nagios can be a challenge for newcomers, but the rather complex configuration is also its strength, as it can be adapted to just about any monitoring task. What it may lack in looks it makes up for in power and reliability.

Icinga
Icinga is an offshoot of Nagios that is currently being rebuilt anew. It offers a thorough monitoring and alerting framework that’s designed to be as open and extensible as Nagios is, but with several different Web UI options. Icinga 1 is closely related to Nagios, while Icinga 2 is the rewrite. Both versions are currently supported, and Nagios users can migrate to Icinga 1 very easily.

NeDi
NeDi may not be as well known as some of the others, but it’s a great solution for tracking devices across a network. It continuously walks through a network infrastructure and catalogs devices, keeping track of everything it discovers. It can provide the current location of any device, as well as a history.

NeDi can be used to locate stolen or lost devices by alerting you if they reappear on the network. It can even display all known and discovered connections on a map, showing how every network interconnect is laid out, down to the physical port level.

Observium
Observium combines system and network monitoring with performance trending. It uses both static and auto discovery to identify servers and network devices, leverages a variety of monitoring methods, and can be configured to track just about any available metric. The Web UI is very clean, well thought out, and easy to navigate.

As shown, Observium can also display the physical location of monitored devices on a geographical map. Note too the heads-up panels showing active alarms and device counts.

Zabbix
Zabbix monitors servers and networks with an extensive array of tools. There are Zabbix agents for most operating systems, or you can use passive or external checks, including SNMP to monitor hosts and network devices. You'll also find extensive alerting and notification facilities, and a highly customizable Web UI that can be adapted to a variety of heads-up displays. In addition, Zabbix has specific tools that monitor Web application stacks and virtualization hypervisors.

Zabbix can also produce logical interconnection diagrams detailing how certain monitored objects are interconnected. These maps are customizable, and maps can be created for groups of monitored devices and hosts.

Ntop
Ntop is a packet sniffing tool with a slick Web UI that displays live data on network traffic passing by a monitoring interface. Instant data on network flows is available through an advanced live graphing function. Host data flows and host communication pair information is also available in real-time.

---

Best Microsoft MCTS Certification, Microsoft MCITP Training at certkingdom.com

Ubuntu 13.10 takes aim at Microsoft, Apple, Google

Canonical is making a serious run at the cloud, client, server and smartphone markets

Ubuntu is moving into the rarified class of operating systems that cover x86/x64 clients and servers, ARM-based tablets/smartphones, and commodity cloud instances. Meaning that it's taking on everybody from Microsoft to Red Hat to Apple and Google.

On the cloud front, Canonical’s announced compatibility with OpenStack APIs for both internal, Ubuntu-hosted or external clouds speaks to Canonical's attempts to make Ubuntu instances a default choice. (Watch a slideshow version of this story.)

On the client side, Ubuntu’s Unity GUI starts to go places where Ubuntu-derivative/associated distros like Linux Mint cannot: onto smartphones and tablets. It took pains for us to obtain a smartphone that would run Ubuntu 13.10, but the Nexus 4 we used is essentially identical in functionality to installations we put on notebooks and virtual machines.

The server version is targeting enterprise scale-out as well as scale-up implementations. Ubuntu server editions are still Debian with Canonical clothing, but thanks to attention paid to hypervisor compatibility and leadership in both the OpenStack compatibility and Amazon Web Services EC2 worlds, 13.10 Server wants to rock rapid rollout and play everywhere, with every hypervisor. That includes stacking application instances inside of itself for what Canonical calls “compactness”.

+ ALSO ON NETWORK WORLD Ubuntu 13.10: The good, the bad and the ugly +

It works, and there are aids to rolling out 13.10. But the stability of OpenStack will be a target for criticism as various vendors tilt and mangle OpenStack towards their own pursuits and goals.

Ubuntu 13.10 client

Dubbed Saucy Salamander, the 13.10 client has changed very little from the previous version, although the GUI plumbing underneath is in transition. This may be the last version of Unity that uses X.org underpinnings, as Canonical transitions to the windowing/GUI system Mir and its X.org-Mir translator.

The support life for this OS is only nine months and the next version, 14.04, will be a five-year supported version that might contain the debut of XMir and Mir -- which would be somewhat radical for Canonical in an Long Term Support release.

The Ubuntu Dash dashboard has been upgraded to allow opt-out searches through the Ubuntu One cloud storage service. Queries sent when opted-in are sent to Ubuntu One, and Canonical serves as a proxy information broker among the current 50-plus third-party providers, listed here, sending the results back to the user. Call it: Answers-As-A-Service, or AaaS.

On one hand, this allows searches across numerous providers, and hopefully, the Canonical parsing engines are smart enough to pick suitable answer-providers from among the third-party supporters. This also means that as Canonical ostensibly anonymizes user information and privacy policy prevents Big Data ad engines from coming into play. This will please privacy advocates.

The other hand is that Canonical does store the query and account information associated with the search transactions by Ubuntu One accounts. This information is stored and can be the subject of a subpoena or other order, hence possibly revealing user/query information -- to the chagrin of privacy policies and mandates.

The Dash update, however, is deceptively handy. Yes, Amazon is featured prominently in searches for products. A sampling of various queries was enlightening, and OCD query searchers will become rapidly addicted to the amount and variety of results, when using the uprated Dash. It's perhaps the key secret sauce that makes this edition useful.

Inside this version are updates of the Linux kernel (3.11), Firefox (24), and when we tried to update SAMBA 4.04, it cratered the instance, even though a prior version was installed. The second time was a charm.

Ubuntu Server 13.10

Through installation, we saw no real server changes on the surface. We could choose to install Open SSH, DNS, LAMP, Mail, Print, SAMBA, Tomcat, and/or virtual machine hosts or add a Postgre database. Or our own manual package selection from the installation menu. We tried various combinations in several instances.

Oddly, Canonical still packages MySQL as its default LAMP database, where many in the industry have gone to MariaSQL, a fork of MySQL. Perhaps it's been retained for compatibility sake -- and the fact that Oracle can be hired to make it work.

Part of the reason we like Ubuntu Server 13.10 is the ability to use Juju to spawn app instances into Linux Containers/LXC. Our problem with this is that LXC isn't due to go to production (a 1.0 version) until February of 2014. If you're adventurous, and we were, it's possible to use a Juju charm to deploy applications into containers that are similar to Oracle/Sun's original idea of a container -- with a few additions, such as a bridge with network address translation and unique IP addresses (only IPv4 for now), and other walled-instances that remind us of how SELinux and other neo-virtualization systems can place significant walls around app instances within servers.

The LXC components start with a shared bridge, and several other adaptations that define kernel name space, and kernel addressing -- but not those found in full bare-metal hypervisors, rather, the kind championed by Parallels in their Virtuozzo kernel-sharing scheme. Instances aren't necessarily unique, and they're not as simple as chroot, which redefines user space for partitioning an application.

Prior to 13.10, Juju's behavior was to dutifully launch applications into their own instance, which was fast, but wasn't very efficient, and made a lot of revenue for cloud providers, as they often charge by the instance. It's possible, using the pre-production LXC method, to launch apps into their own containers, and provision the containers on the fly with comparatively simple commands. But this is alpha code, and is still experimental in nature, rather than more mature products whose rules and qualities are comparatively well known.
Ubuntu

The LXC setup contains templates that can be used, although some are experimental. We chose to boot a simple application (a WordStar-like text editor) into an AppArmor shell. SELinux barriers and constructs are also supported. We set up an LXC bridge, provisioned it, then stuffed the application inside the container. It takes about 10 lines of code from initial start until the app is deployed, all of which can be built into a script with arguments to provision many containers simultaneously, much as the popular Puppet instance controller product provisions OS instances through a communications bus. Everything can be setup to boot at once, by dependency choice, and/or at restart time. File system mounting is easily supported, and one needs to be careful that file locks don't thwart multiple app instance access.

---

Best Microsoft MCTS Certification, Microsoft MCITP Training at certkingdom.com

Top open source network management tools

Zenoss Core edges Nagios Core, NetXMS and OpenNMS in four-product roundup

As networking continues to expand and diversify, encompassing a growing number of wired and wireless devices, the demand for network monitoring tools remains high. While feature-packed commercial products abound, the growing market for monitoring tools has also fueled robust offerings from the open source community.

We reviewed four popular open source products - Nagios Core 3.5, NetXMS 1.2.7, OpenNMS 1.10.9 and Zenoss Core 4.2. All four products are mature, have extensive monitoring capabilities similar to their enterprise-grade counterparts, and are currently updated with good community support.

The products offer event management, performance monitoring, alerting and reporting on network devices such as servers, routers, switches, printers, UPSs, websites, business applications and mobile devices. Monitoring can be as simple as a PING request to a single device to full-fledged management using SNMP (Simple Network Management Protocol) or a native agent.

---

Best Microsoft MCTS Certification, Microsoft MCITP Training at examkingdom.com