Novell releases new version of appliance toolkit

Novell is releasing Version 1.1 of its SUSE Appliance toolkit.

“It s delivering the broadest cloud support of any imaging solution,” enthused Michael Applebaum, director of Linux and appliances marketing for Novell.

“It has The most enterprise use cases. It’s the fastest way to bring applications to the cloud today.”

The new version has a number of important new features, he added.

“We’re adding support for Amazon EC2. You’ll be able to run your SUSE application in EC2. We’re adding support for building those applications within SUSE Studio, the solution that’s part of the SUSE Appliance toolkit. We’re also adding KVM and ODS support.”

Applebaum gave an example of how this will work in practice.

“If you’re a retailer and you have a large network of stores, you want to keep your system flexible by having storage separate from the devices, and having them reboot over the network when you change software.

“Instead of having a complicated store you can then manage it separately and remotely, assuring security, assuring patching, and expanding the use cases.” The company talks about going “from zero to cloud in a month.”

SUSE Studio began as a stand-alone offering a year ago. The idea at the time was that independent software vendors would use it to make their applications cloud-ready.

Applebaum said that over the last year the company has heard from many enterprises who are looking for their own cloud-prep solution “We’re now calling on in house IT shops, because we can package their applications and ready them for any cloud development.”

These remain early days in the cloud era, Applebaum concluded, with people thinking about clouds in many different ways.

“I’m not sure if they’re confusing or equating clouds and boxes per se, but they’re increasingly understanding the key tenets of the cloud approach, and they’re looking for practical solutions that work in an incremental way, without having to have a completely siloed management approach.”

 

Open source runs a house of cards

All software is a house of cards.

(This picture, from CBS, shows the world record-breaking card house, from Bryan Berg, of the Venetian Casino. The man who controls the Venetian empire, Sheldon Adelson, formerly ran the Comdex computer shows.)

It’s inherently fragile. For any software program to work it must be perfect, or it collapses. The larger the system, the harder this is to maintain.

Then there is the fact that you have all sorts of people trying to blow holes in it all the time — the number and tenacity of such people increases as the software becomes more critical.

The difference between “open” and “closed” software is mainly a legal one. You can see and edit Microsoft code. Criminals do it all the time. Both open and closed source are equally vulnerable.

But most people are not criminals. Most people are good. Most people even follow EULAs, that’s how good they are. This is the social reason why open source works. Both types of software have people trying to blow holes through them. But open source can bring more resources to bear in fixing them.

Anyone who has tried to build a large, proprietary system from scratch knows how hard it is. It gets exponentially harder as the system scales. You might call it Moore’s Law of Software Degradation. Bigger stacks mean more things can go wrong.

Thus we have Steve Jobs’ secret source. Every once in a while he drops what he built before and starts again from scratch. The Mac OSX is large, but based on an open source LinuxUnix, so it can maintain. The iOS is new, and relatively small — everything old becomes new again.

Innovation can happen within smaller systems. The houses are smaller. They can be built by small teams. So throughout my 30 year career I’ve watched constant reiterations of the same old things. Web games of the ’90s copied PC games of the ’80s. Phone games of the ’00s were the same. Apps, again, mostly, the same. Only the displays get better.

As a system gets bigger the need for an open source process grows, the need to cooperate grows if this skyscraper of cards is to keep standing and growing.

That’s why Mozilla joined the Open Invention Network — to protect its growing code base. It’s also why the Bilski case matters little in the end — the incentive for people to cooperate despite patents is too great.

Once you understand that all software is a house of cards, subject to Moore’s Law of Software Degradation, the need for open source becomes clear.

If it didn’t exist we’d have to invent it.

 

Fedora 14 with SPICE, MeeGo support due in November

The Fedora project has announced the release of the Fedora 14 beta, code named “Laughlin.”

Wander Mule

General availability is expected in November, the open source project also announced.

The Fedora 14 beta incorporates support for the Simple Protocol for Independent Computing Environments (SPICE) framework, which was developed by Qumranet for the virtual desktop infrastructure. SPICE is an open source solution for integrating with virtual desktops.

Red Hat, sponsor of Fedora, acquired Qumranet two years ago.

The Fedora 14 beta also supports a number of other emerging technologies, including the Intelligent Platform Management (IPMI) interface for remote or powered-off server management, as well as D programming language support and Rakudo Star, an implementation of Perl 6, the project announced on Tuesday,

It also offers, the press release states:

Improved debugging and integration with existing tools to enhance Fedora as a RAD platformAlternative environments like Sugar, and software from the MeeGo projectThe next-generation systemd management for faster startup and on-demand loading and unloading of services is expected to be offered as a technology preview in Fedora 14, and is likely to become the default initialization system in Fedora 15

Paula Rooney is a Boston-based writer who has followed the tech industry for almost two decades.

Defaults set policy on Google Code and CodePlex forges

My Italian friend Roberto Galoppini (right) has a way of answering questions most analysts are not even asking.

He sometimes does this by examining defaults.

We know, for instance, that carriers wrecked Android by making their own offerings into defaults on phones they controlled. The idea was that while the savvy might complain, and a few might jailbreak their gear, most users would not know the difference and empty their pockets into carriers’ vaults.

Thus, you might say something is open, but through your actions you have it treated as closed.

Recently, Galoppini took this insight to Microsoft CodePlex and Google Code, both of which now claim to be very, very open.

Guess what he found?

The license options Google offers Code users by default do not include the AGPL, which it now loudly claims to support.

To offer your code with this license you have to select “Other Open Source,” then write it in. It is, in other words, a Murkowski license. You have to know it and write it in to get it.

Codeplex does the same thing, he found. Its box lists the Microsoft licenses that are no longer popular, according to Black Duck Software. And it doesn’t list more popular licenses like the Artistic License or GPLv3.

That’s not all. Google Code search does not, by default, let you search by license or by download popularity, a very popular metric. You can get licenses by turning on the “label” function, but the numbers are considered private. Microsoft Codeplex search, on the other hand, does allow these sorts.

In business you can say whatever you want, and while it may be technically true, you can still tweak things to make it untrue in practice. This is what carriers have done with Android. It’s what these corporate forges also seem to be doing.

Enterprises must understand a mainframe is not a cloud

For proprietary companies clouds are a tough deal.

Clouds don’t care about operating systems. They don’t care about where data is. They don’t care about constraints like CPU speed, memory, or formats. They expand as needed, in the background.

Mainframes are not clouds. Neither IBM nor Oracle want you to know that. It would kill them for you to understand that.

In time it will.

So an effort is underway by proprietary companies to embrace and extend their technologies into the cloud metaphor, inducing (or forcing) customers to take them into the new age.

It’s a process that will take this whole decade to shake out. Clouds are still so new enterprises can’t yet get their heads around them. So they are easy to market against.

You mean I don’t care what hardware is on my desk?You mean I don’t need a server room?You mean I can’t see my stuff? My applications, my data, are they still mine?

In Amazon’s EC2 cloud, which remains the flexible enterprise cloud of choice, the business process is seamless. Users don’t need to know their applications and data are somewhere else. They keep working as before. All that complexity is abstracted, outsourced.

To make this happen, of course, your old stuff must plug into the cloud. This gives proprietary companies their opportunity, because they can (at first) dictate the terms of this move.

My point today is, this won’t continue, if enterprise customers understand that the cloud should be their goal, that abstraction is the means to that end, and the way forward is to take only what they need into the cloud, only the essential tools needed to keep the transition seamless.

Once your data and tools are all in the cloud, all those proprietary advantages become simply formats, choices that can be transformed and remade once tools are available which virtualize enterprise software as enterprise hardware is now virtualized.

You go in with all your IT baggage, and over time work to reduce that baggage, repeatedly making choices based on minimizing your costs.

Which means minimizing your commitments to any particular vendor. Even Amazon.

 

CodePlex Foundation becomes OuterCurve

The CodePlex Foundation, originally created to support Microsoft’s CodePlex code site, has been renamed the OuterCurve Foundation to help make its mission clearer.

The mission is to unlock the value in corporate code vaults, building a museum of projects with complete development and documentation support.

Since its launch a year ago, the foundation has developed a governance model, hired key staff, and elected an independent board. The board also created a process for accepting projects, and has brought in nine to date.

Paula Hunter is executive director of the foundation, which is based in Wakefield, Mass. but incorporated in Washington state. She made the announcement in conjunction with the Open World Forum meetings in Paris, which will be covered extensively by ZDNet Open Source.

Before we both headed east, Hunter discussed the effort with ZDNet.

“People thought we owned Codeplex and we never had a legal relationship. It’s Microsoft’s forge. The foundation was a separate entity whose mission was enabling the exchange of code and creating open source communities. We’re not a forge, our projects can use any forge.”

So is this more like Eclipse or Apache? “What distinguishes us from Eclipse and Apache is that we’re license agnostic and platform agnostic. We are also not imposing a strict development methodology. We’re providing guidelines, we’re showing what the process should be, but we’re not dictating.” To date six projects have been approved under the new guidelines.

This does not mean OuterCurve will just be a collection of random projects. “We created a museum model. We have galleries, which are sponsored. Galleries are collections of projects centered around some theme – like the .ASP .NET gallery or the systems infrastructure and integration gallery.

“The gallery manager is the person to connect with initially,” if you’re interested in launching a project under the OuterCurve structure. Over time the hope is that gallery managers will be seen as leaders within their fields. (Programming frogs will become development princes?)

The foundation will hold all IP rights, and provide infrastructure back-up, “like getting co-sign certificates, or helping get money together for documentation,” Hunter explained.

By providing structure and support for corporate-backed code, OuterCurve hopes to coax more of this code out of corporate vaults and transform it into open source projects that will have a wide following, Hunter said.

It’s a point Sam Ramji, the former Microsoft employee who still serves as the President of Outercurve’s board, said in a variety of ways when I interviewed him over the last two years. Open source development is becoming a mainstream IT strategy, but IT departments are looking for a model to turn that strategy into reality.

OpenCurve’s mission is to provide that strategy and create that reality.

 

Why the Super WiFi skeptics are wrong

One of the most interesting points in the current “Super WiFi” debate, which could climax tomorrow with new rules for unlicensed use of TV white spaces, is the apparent lack of opposition.

It’s the dog that hasn’t barked. (Shown is a Hawking WiFi dish adapter, from Amazon.com.)

Ardent Bell proponents like Scott Cleland have been silent on the issue, preferring to focus on their fight against Google and net neutrality. The so-called “Progress and Freedom Foundation,” which called WiFi itself “communism” a few years ago, has likewise kept quiet about the issue.

This has led some to claim the allocation is no big deal. No national allocation is being made. There will be a lot more spectrum in small cities than in big ones.

The critics and skeptics are missing an important point, namely whom tomorrow’s order will be addressed to. While most spectrum orders are destined for carriers or their lobbyists, this order will be addressed to equipment manufacturers.

As with WiFi or any other unlicensed use, the Super WiFi order should be giving chip-makers like Broadcom and Intel an all-clear to begin making chip sets to certain specifications. In this case, this will include a function which tells a radio to use its GPS and check an Internet database before broadcasting.

That’s a big deal. This kind of functionality has not been specified in hardware before. The fact that advocates say adding this is no big deal tells us a lot about where radio technology stands in 2010.

There is a second technical point to be made, one which is just as important.

One reason carriers are so crying over a looming spectrum shortage lies in how they design their networks. They are designed as wireless networks supporting mobile devices. Sounds simple and logical, but WiFi isn’t designed that way. It’s fixed.

WiFi is designed as an on-ramp to a wired network. Your WiFi router connects to your wired broadband. The coffee shop’s signals are vacuumed into the larger Internet.

The purpose of Super WiFi is not to change this structure, but to enhance it and create competition for the wired component. It’s just a longer on-ramp.

Instead of going into the nearest wire, Super WiFi makes it possible to find another wire, perhaps miles away, and funnel traffic to it. That wire might be competitive fiber, priced at a competitive rate, rather than copper offered at a monopoly rate. The resulting competition could drive down broadband prices for everyone.

It’s true that in this case the Super WiFi signal could become a bottleneck. If 6 restaurants a half-mile from a phone switch decide to share backhaul using Super WiFi, their traffic could fill the air. And if someone else puts up a second or third or fourth Super WiFi set in the area, frequency contention could ensue.

Fortunately there are ways to engineer around that problem as well. A directional antenna. Aim the signal to a specific point and interference is minimized. Both sides of the frequency transaction get maximum use of the resource.

All the little tips and tricks that have evolved to maximize the capacity of WiFi over the last 10 years — Pringle Cans, new encoding, maximum use of the wired infrastructure — can be applied to Super WiFi by equipment manufacturers, and will be.

WiFi was an economic boon to wireless equipment companies, and Super WiFi can take the technology even further.