Open-source software has won the argument. Now a new threat to openness looms | Hacker News Reader

Open-source software has won the argument. Now a new threat to openness looms

economist.com

30 pointszzkt16 comments

This is the usual, but valid, argument that as apps move to the cloud, we get a new kind of vendor lock-in.

The first step you can take is to use the Affero GPL instead of the old GPL:

"... version differs from the version of the GNU GPL on which it is based in having an additional provision addressing use of software over a computer network. The additional provision requires that the complete source code be made available to any network user of the AGPL-licensed work, typically a web application."

This doesn't solve all problems, as it doesn't cover software developed from-scratch by the big players.

I think an "Affero LGPL" would be good to get companies to contribute back library enhancements without "infecting" and thus having to give up their whole IP.

Regarding data portability, I've become a fan of json - it's easy to understand and process, and it forces interfaces to be simple (compared to xml, e.g.).

The AGPL won't help. As someone said earlier, the problem is data portability, not source code availability.

The thing that a lot of people seem to forget in this debate is that there is room for all 3. On a personal level, I make extensive use of commercial software (especially MS SQL Server) for some things, wide use of open source software (FireFox and Python are first amoungst many) and some cloud software (gmail, google calandar come to mind).

On the business side though, I would be hesitant to ever advise an employer or client to make use of cloud resources for a few reasons. One of them is the lock in as mentioned in the article. Another is the risk of loss of data that is outside the businesses control. Another is the fact that the cloud service could upgrade the software in a way we didn't like with very little recourse on our part.

After all, there are times when you want to keep the older version. For instance, numerous businesses have declined to switch to Vista and I will not be leaving python 2.54 until some of the libraries I use finish their updating process.

Of course, all of the concerns I mentioned can be addressed, but it would require substantial due diligence on the IT side to verify that the cloud computing company has adequate protections in place and then on legal's side to ensure that they are contractually enforceable.

For small, technically-oriented companies, this article probably sounds kind of lame. But think about 80legs, or SmugMug, or someone with a 200-machine farm and 20TB of data squirreled away in S3. Moving to a different cloud would be an expensive, tedious, big-bang proposition.

It doesn't matter how open the data formats are. If it's difficult to simultaneously use two or more vendors at a time, you are locked in.

What about the non-cloud version? Is it easy to move 20TB from one storage system to another. What about to a new data center? What about to a cloud provider?

I think if you have 20TB of data that you need to have online all the time, things are going to be hard regardless.

Yes, but if you have your own datacenter and some spiffy new software or device comes on the market, you can buy it and set it up. You can take advantage of innovation from anywhere.

If you are sitting in Cloud A and Cloud B introduces some new thing... you're out of luck until Cloud A (who knows you're not going to move) decides to offer something similar.

It's not a threat to openness, it's a threat to portability. It's not quite the same thing. You can be open without being portable and vice versa.

Yeah, the article includes SaaS as cloud computing. The SaaS portability problem has been around for years.

Cloud computing as I think of it-- AWS, Mosso, etc-- is generally built on familiar and open technologies. If anything it is easier than ever to redeploy an application with a different cloud provider...and vendors like CloudKick and RightScale make switching trivial.

I think the economist has done more harm than good with this article.

For sure. Talk about "clouding" the subject. The cloud this article is talking about is not the "cloud" of today, it is the "cloud" of yesterday, i.e., this has been a problem since the first web app went online for commercial use well over a decade ago. Today's cloud, to me, is about software and service providers using cloud resources, and so far those are the same resources that were used in-house (Linux, MySQL, etc., etc.). There is little danger in getting data out of one MySQL DB and importing it into another at another cloud provider.

I think Amazon has a chance of achieving serious lock-in with AWS, not by doing anything pernicious, but simply by adding so many valuable differentiating features that at some point (perhaps already) it will be too much of a pain to switch.

rightscale.com is going to support multiple cloud vendors, and I would guess they'll offer consulting services to help people move.

Come on, it's just another vendor. Being tied to a vendor is a choice we've always made, the beauty of Unix is the abstraction. Present an attractive abstraction and developers will come, not only because of the strength of your provision but because we also have to recognize how hard it would be to change environment providers.

The Unix model presents files and folders, it is a fantastic model with a few operations. You present me with a file system and I'll manipulate it. We've spent 30-40 years building tools to manipulate files grep, sed, awk, join, sort etc.

I don't see any threats. Just don't use services that don't let you port your data. And encourage the use of those who abide by standards and fair play.

Demand portability!

Ask Richard Stallman about "Easier said than done".

Reply on news.ycombinator.com