Linux A Future

t never fails to amaze me just how the same arguments keep coming up over and over again, like bad pennies. Most of them ignore facts so blatantly it gives me a headache.

The I’m referring to today is the one about Linux Viruses.

First of all, Linux Viruses exist. Yep, it’s true, they exist. There’s actually a couple of them. Wikipedia lists an even 30 of them. There’s even Virus protection for Linux.

True, compared to MacOS and Windows, that’s a drop in the bucket. Barely worth noticing. And that’s where the same old argument comes up time and again.

“The reason that Linux doesn’t have viruses like MacOS and Windows is because it’s not as popular!” The argument is that if more people used Linux, there actually would be viruses for Linux. The completely ignores the fact that the desktop is the exception, not the rule. Linux dominates elsewhere. Servers, smart phones, super computers, etc. Linux is kicking ass and taking names. A virus that could infect all those systems would be hugely valuable both monetarily and for a reputation. Yet, it doesn’t happen.

Even  the ones that already exist pretty much suck compared to their Microsoft and Apple counterparts.

So is there a reason to even consider running anti-virus on your Linux computer? Shockingly, the answer is still yes, but not for the same reasons some Mac users and all Windows users should.

The past couple of years have seen a fundamental shift in the way enterprise IT is done. This trend started with the consumerisation of corporate devices, and the gradual introduction of BYOD policies. This has paved the way to the “user voice” being heard more than ever. We’re seeing this trend continuing into the backend of corporate IT, from the infrastructure services to the enterprise applications. Business users are slowly demanding tools and environments that will enable them to do their jobs better, they are no longer satisfied with using whatever IT mandates. This applies just as much to enterprise applications as it does to mobile devices.

There is clearly a need for easier access to new enterprise applications, enabling users and companies to try them, and make an educated choice based on experience rather than a sales pitch. Companies can no longer afford the long wait times that normally come with procurement cycles, nor do they want the narrow options they provide.

On the other end of the spectrum, IT service providers are becoming more sophisticated about the way they offer their services. Many telcos and service providers have recognised the value of offering public cloud services, and most of them already have clouds up and running. As they become more experienced cloud providers, they realise the importance of diversifying their vendor base to avoid lock-in, as well as looking for lower-cost options that will still give them access to the latest features and enterprise-grade support. Telcos and service providers have strict SLAs that they are looking to meet. They don’t take the decision to offer cloud services lightly, and are looking for the right partners to help them achieve their standards for security, reliability, and uptime. The sophistication also extends to looking at the building blocks of their cloud infrastructure. So, instead of looking for “cloud infrastructure”, they are now looking at optimising the different aspects of that infrastructure in a service-driven manner.

We are seeing three main areas in which this shift is taking place: compute, storage, and networking. OpenStack continues to gain traction as the best platform on which to build cloud infrastructure, but customers aren’t always satisfied with the default options that come with OpenStack. They want more choices that suit their businesses, and they want to be able to switch with minimum hassle.

At Canonical, we believe we stand at the intersection of those two trend lines, with Ubuntu 13.04 and paving the way to 14.04 LTS. The move towards more user-driven technology choices fosters the message at the core of open-source software, namely to drive adoption among end-users and slowly expand that to the rest of the enterprise. We’ve seen this happen with technologies like Gmail, Gdocs, Skype, and others, which have moved from mainstream use to being legitimate enterprise applications. We believe that open-source will be the next step of that trend. The increasing ease with which tools like Juju can be used to deploy and manage cloud-based services provides a strong foundation on which to build on-demand, fast-paced enterprise offerings. It transforms the way enterprise applications are created, distributed, sold, used, and billed, and opens up the way not only for smarter IT choices for companies, but to a whole host of application providers that couldn’t access the enterprise market before.

On the cloud infrastructure side, the expansion of our ecosystem plays well into the diversification we see in technology choices along the cloud stack. Ubuntu now works seamlessly with various products such as Ceph for storage, Floodlight for SDN, and the Canonical-VMWare collaboration links OpenStack compute (Nova) to the ESX hypervisor. Offering a ‘High Availability’ deployment configuration for core OpenStack components such as RabbitMQ and MySQL helps meet service provider requirements for maximum uptime and reliability. We expect this to be extended to other Juju charms in the near future, offering a complete HA environment for cloud infrastructure.

Well, wipe that smile off your face, fleshy one. Killer robots are a real threat to our future and must be outlawed now, according to a campaign launched in London on Tuesday by five international NGOs, led by Human Rights Watch.

The Campaign to Stop Killer Robots — yes, that is its real name, and you can find its website here — calls for a comprehensive ban on the development, production, and use of fully autonomous weapons. The launch event came a month in advance of a UN report on the subject, set to be delivered to the Human Rights Council in Geneva on May 27.

"Killer robots would cross moral and legal boundaries, and should be rejected as repugnant to the public conscience," said Human Rights Watch's Arms Division Director Steve Goose. "Lethal armed robots that could target and kill without any human intervention should never be built."

The campaign brought out its big guns for a press event at the Houses of Parliament in Westminster: Jody Williams, a Nobel Peace Prize laureate from the International Campaign to Ban Landmines, and Dr. Noel Sharkey, an artificial intelligence expert from Sheffield University.

Undermining the message slightly, they also brought the cute-looking 1950s-style sci-fi robot pictured above. What, no T-1000?

Still, the campaign is in earnest, and Sharkey ticked off many reasons why he fears killer robots in our future: the fact that the Pentagon is currently hiring more drone operators than actual pilots, for example, and the building of its X47B unmanned plane, which works like a drone on steroids.

But Sharkey didn't single out U.S. research efforts; he fears this is a global problem. "There are a lot of people very excited about this technology in China, Russia, Israel," he told the Guardian, "very excited at what is set to become a multibillion-dollar industry. This is going to be big, big money ... We won't hear about it until China has sold theirs to Iran."

"I think we are already there," Sharkey told reporters at the event. "If you asked me to go and make an autonomous killer robot today, I could do it. I could have you one here in a few days."

A slightly bemused UK government spokesman told Reuters that "there are no plans to replace skilled military personnel with fully autonomous systems."

For a little light reading, try the Human Rights Watch 50-page report on the subject, Losing Humanity: the Case Against Killer Robots.

Photo by Oli Scarff/Getty Images

Topics: killer robots, Robot, Tech, World

t appears that LinkedIn accesses something other than just the contacts in the mailbox when a person provides them with corporate login credentials.

I ran some tests with two brand new mailboxes, and it seems that LinkedIn accesses both the Contacts and the Sent Items. For a test mailbox with no Contacts or Sent Items at all the LinkedIn page returned an error that it wasn’t able to recognise the webmail URL I had provided. But as soon as I sent just one email, the next attempt returned that email address in the list to invite to connect on LinkedIn.

Is this a bad thing? Absolutely. especially when LinkedIn has trouble telling the difference between a mailbox and a mailing list.

So what does LinkedIn store when you give them your login details?

In the first screenshot of the article you can see I had to provide the email address, username, password, and webmail URL.

On a second test run, I only had to provide username and password.

Enigmail is a security extension to Mozilla Thunderbird and Seamonkey. It enables you to write and receive email messages signed and/or encrypted with the OpenPGP standard.

Sending and receiving encrypted and digitally signed email is simple using Enigmail.

When starting it for the first time, you are guided through the basic setup. We also prepared a new users' guide that explains how to use OpenPGP.

Boise University PhD candidate Joshua Kiepert has built a 32-way Beowulf cluster from Raspberry Pis.

Kiepert says his research focuses on “developing a novel data sharing system for wireless sensor networks to facilitate in-network collaborative processing of sensor data.” To study that field Kipert figured he would need a decent simulator, preferably a cluster so he could simulate lots of distributed sensors. The University possesses just such a cluster, comprised of 32 nodes each packing a quad-core Intel Xeon E3-1225 CPU humming away at 3.1GHz.

That's a lovely facility and is therefore much in-demand, which meant Kiepert could not guarantee access for lengthy experiments. That got Kiepert thinking that if he had a cluster of his own he could tweak as required, and that cluster was a bit closer to the low-level hardware used in sensors, that would be a fine thing.

Kiepert's mathematical skills then did some multiplication: at $45 per Pi, including an 8GB SD card, he could acquire the raw materials for a 32-way cluster for $1500, or the same price as one Xeon-powered PC. As he contemplated the design for such a cluster, Kipert settled on Arch Linux for its tiny size. He eschewed the Pi's micro-USB port as a power source, as he felt it would complicate cabling, instead using a 5V pin on the machines' I/O headers.

USB power would also have complicated the housing for the cluster by adding weight, a problem Kiepert said was not insignificant because 32 ethernet cables were a drag on the slim and light computers.

“In order to keep the cluster size to a minimum while maintaining ease of access, the RPis were stacked in groups of eight using PCB-to-PCB standoffs with enough room in between them for a reasonable amount of air flow and component clearance,” he writes. “This configuration suited our needs for power distribution very well since it allowed for a power line to be passed vertically along each stack.”

“Using this orientation, four RPi stacks were assembled and mounted between two pieces of acrylic. This created a solid structure in which the cluster could be housed and maintain physical stability under the combined weight of 32 Ethernet cables.”

Joshua Kiepert 's 32-way Raspberry Pi cluster

The Pis were also overclocked, using Turbo Mode, to give Kiepert the grunt needed to run his simulations.

Kiepert's now doing all his research on the cluster, writing in a lengthy (PDF) account of the build that “I have found performance perfectly acceptable for my simulation needs, and have had the luxury of customizing the cluster software to fit my requirements exactly.”

Custom software doesn't, however, mean faster performance: Kiepert admits performance of the cluster isn't stellar, even after he re-wrote simulation software for his cluster.

But the price was right: the PDF above includes a bill of materials that includes cabling, lighting and even screws needed to assemble his acrylic racks. The total came to $1967.21. ®

Linux operating system is expected to improve security, stability

Microsoft Corp.'s (MSFT) Windows XP is among the company's most beloved products. Generally lauded for its stability, compatibility, and security since Service Pack 3 rolled out, the aging operating system has still been hanging around in many enterprise deployments -- including the International Space Station (ISS).  But the ISS and others have found out the hard way of late that slower patching and a phase-out of support have left the once unbeatable OS a major security risk.

I. Phasing Out XP, Moving to the Penguin

Keith Chuvala of United Space Alliance, a contractor who handles much of the ISS operations, decided enough is enough after a 2008 security breach; he's switching the "dozens of laptops" aboard the ISS to Debian 6, a Linux operating system.

Linux has been used on the ISS systems since its 1998 launch (see this 2001 article [PDF] in Linux Magazine, for example) and in NASA ground system since its inception. However, most laptops used aboard the station for day-to-day activities like viewing stock inventory, controlling scientific experiments, or checking the current station location -- a cluster dubbed "OpsLAN" -- ran Windows.

But NASA received a wakeup call in 2008 when a Russian cosmonaut inadvertently brought a W32.Gammima.AG worm aboard his Windows laptop, which infected the other laptops onboard, requiring painstaking cleanup.  While Linux machines are not immune to worms and other malware, their small market share and quick patching (thanks to the open source community) have made Linux distributions more secure than their Windows counterparts from a practical perspective.

Phone, Text, Eyes, Mail

Subscribe to Silent Circle for one, low price and get full access to all four encrypted communication tools included in the Silent Suite. Silent Phone, Silent Text, Silent Mail and Silent Eyes - everything you need to communicate more securely