Thomas Delaet

Uncloud

2011-02-05T20:18:00.000+01:00

We (= team from DistriNet) are participating in an online competition, named Battle of Talents. In short: engineers form virtual ventures and MBA students can invest in these virtual ventures. The most successful venture wins. Our venture is called Uncloud and solves this problem in a more generic way. Our product is a software service that provides a secure backup of data that is stored in cloud applications and enables easy conversion between competing cloud applications. It is positioned as a 'data insurance'. Check out the website@http://www.uncloud.eu and enjoy our promotional video's!

I don't believe in the cloud

2011-01-10T21:05:00.001+01:00

I do not believe in the cloud. At least not as an outsourced solution for all my data/applications. The reason is simple: cost.

Suppose I have 2TB of data. How much does it cost to store this in the cloud? I did some simple calculations for different providers and this is the result:

The 'NAS' option is when you buy yourself a NAS device (actually 2 NAS devices since you need a backup) and add electricity costs.

With the exception of Flickr for storing your photos, the NAS option is by far the cheapest option. I used two years as a timeframe since that is a realistic timeframe to buy a new NAS device (actually most people keep NAS devices longer so the NAS option becomes even cheaper).

Other advantages of using NAS devices for storing your data:

privacy friendly/piracy safe
content is close to where it is consumed most frequently (at home)

When you look at different types of content from a cost perspective, you can distinguish the following categories:

music/video's: not cost-effective to store at a cloud provider. You also need a lot of bandwidth.
photo's: cost-effective since there exists fixed-fee services like Flickr.
all other types of content: you can find free (or very cheap) services. For example for mail, tasks, contacts, documents, notes, ...

Regarding the pricing of cloud storage, I am wondering which of the two following options is true:

Flickr is too cheap for storing unlimited number of photo's/HD home video's.
General purpose storage services like Amazon S3 make huge profits.

Torrenting in the cloud

2011-01-03T16:49:00.000+01:00

I want to have my torrents as fast as possible. Since my main computing device is a laptop, I'm not online all of the time: I commute to work, I sometimes sleep (when my children aren't sick), ... So I want an always-on device to take care of my torrenting needs. Luckily, I have a server/router device that is responsible for my Internet access. I installed transmission on this server (actually 2 servers, it's a failover setup. Yes, I'm a geek). In addition, I installed this Google Chrome extension that allows me to manage my torrent application running on my server from within my laptop browser, wherever I go! Whenever I click on a torrent link, it is automatically added to the application on my server and starts downloading. My browser gets notifications about finished downloads, I can see the list of downloads, pause downloads, set speed limits, ... all from within my browser! What is left is a mechanism to import finished download in a (non-existent?) web application that manages my videos and music ...

Here's a screenshot of the torrent extension in Chrome:

Google Web Store

2010-12-24T14:19:00.000+01:00

The Google Chrome Web Store is Google's web-answer to the iPhone app store and the many other app stores that emerged. Web applications are nothing more than normal 'websites' with javascript functionality. Gmail is a web application, Facebook is a web application. The added value of the Web Store is that you can easily discover, rate and review web applications. For developers, the web store offers a framework to monetize your apps: you can launch a free trial, use in-app payments, use google's infrastructure to charge for your app, .... The goal is to get as many applications in the web store use advanced HTML5 functionality like offline support and advanced graphics. When this is realized, there will be no functional difference between web and desktop applications.

This video is from the event when the Google Chrome Web store was announced.

Micropayments and Internet utilities

2010-12-20T09:53:00.000+01:00

In an interview with a Belgian newspaper, one of the founders of the World Wide Web makes the following statements:

He is opposed to monopolies like Google and Facebook because it is against the philosophy of the open and collaborative web.
He argues for the idea of micro payments for search results and other web services (like consuming media) as an alternative to advertisement based services (like Google does with its search)

I like the idea of micro payments. I would have no problem with paying for search results, listening to music, watching video's, ... as long as the unit price is low enough (less than one euro cent for a clicked search result).

In an ideal world, a world-wide government would control 'utility' services on the Internet like search, much like national governments operate the electricity,rail and other infrastructure. Micro payments for search results would go directly to the authors of web pages in such a model. Unfortunately, this is utopia.

Dutch version of interview
Google Translate version of interview

Chrome OS

2010-12-14T15:35:00.000+01:00

Last week, Google launched a beta program for the first real cloud computer: the CR-48. The CR-48 is the first computer running Chrome OS, google's cloud operating system. The concept is simple: A minimal operating system that only starts a browser (Google Chrome) with some added functionality like configuration of network settings. Will these type of devices accelerate the shift from desktop apps to web applications? I don't know, but I do see some compelling advantages when people use a Chrome OS device compared to a traditional PC/Mac:

It has a non-existent learning curve. If you can work with a browser, you can work with a Chrome OS device
The device requires no maintenance. No security updates, no application installs, no operating system patches, media codecs, ...
It is fast. Really fast. Booting is under ten seconds. Resume from sleep modus is instantly.
It is interchangeable. All your data lives in the cloud. You have everything available whether you are in your browser on your 'classic' Desktop OS or on your Chrome OS device.
It is cross-platform. You are not locked in to one 'app framework' like iPhone apps or Android apps. Every website is a potential application for Chrome OS.

The first real notebooks running Chrome OS are expected in mid 2011. I hope someone will also commercialize a nettop/desktop device running Chrome OS. I know at least two computers in my home that could be immediately replaced by a Chrome OS device.

Video that demonstrates some of Chrome OS's capabilities.

Greplin - the search bar for your life

2010-12-14T15:15:00.000+01:00

Greplin indexes all your web applications (facebook, linkedin, google calendar, gmail, ...) and provides a quick as-you-type search box for them. Still in beta, but looks promising.

Greplin Demo from greplin on Vimeo.

Resident web applications

2010-11-03T12:24:00.000+01:00

I'm a GTD freak, so my todo app is open all the time. The same goes for my Gmail. My browser offers the option to 'pin' tabs. Pinned tabs are smaller (only the icon is shown) and are resident: they open automatically when you start your browser.

Meyer on cloud computing (google docs)

2010-10-25T12:42:00.000+02:00

In his blog post, the cloud and its risks, Bertrand Meyer describes a bug he found in Google Docs and uses this case to highlight the strenghts and weaknesses of a cloud-based solution like Google Docs.

Bottom-line: you depend on the cloud provider. If you want to avoid all potential data loss, you have to make an extra backup of your data on local storage or use an online storage service like Amazon S3 as an extra backup. A precondition is that your cloud-based solution (like Google Docs) offers an API so that you can get your data out of the service.

Thinking about this: wouldn't this be a compelling business case: an application that regularly gets all your data out of your Facebook account, LinkedIn, Google Docs, Gmail, Hotmail, whatever and makes an extra (secure) backup of that data.

Use Google Docs as download directory

2010-10-12T21:22:00.000+02:00

In my last post, I migrated all my documents to Google Docs. If I click a document on a website, my browser downloads it to my local hard drive. Then, I need to upload the same document to Google Docs. Enter this extension for my browser.

The extension scans each webpage for office documents (word, powerpoint and pdf) and rewrites the web page so that the document opens with Google Docs viewer. The latter has the option to save the document in Google Docs.

Right-click menu when extension is enabled. Default behavior is to open in Google Docs Viewer

Going Google Docs

2010-10-03T19:33:00.001+02:00

For as long as I remember, I have been managing all my personal documents with a version control system: first CVS, then subversion, and during the last few years mercurial.

With CVS and subversion, I configured my home router as the version control server. I manually synced changes with the version control server using the command line tools that CVS and subversion provide.

Mercurial is a distributed version control system and fits better with my workflow: I do not need a central server: all computers can sync with each other.

As part of my experiment to migrate all my data and applications to web applications, I uploaded (almost) all my documents to Google Docs.

In Google Docs, I defined everything I work on as a project. Every project gets a folder and the folder contains all spreadsheets, notes, presentations, documents, PDFs and other types of files that are associated with that project. This project-oriented approach is inspired by the productivity methodology I follow.

Google Docs gives me instant access to all my documents from the two laptops I use and our kitchen computer. All my documents are always the latest version and I no longer need to manually sync changes between my computers. Furthermore, synchronization conflicts that occasionally happened when I used a version control system are eliminated.

As of today, I use Google Docs for all my documents with two exceptions: accounting data and LaTeX documents.

Accounting data: I have yet to find a web application that supports the flexibility of my current accounting application. So for the near future, I keep my financial data in a local repository.
LaTeX documents: Apparently, there exists an extension to write and compile LaTeX documents in Google Docs. Using this extension is on my to-do list.

Moving to the cloud ...

Intro - living in the cloud

2010-09-20T11:23:00.001+02:00

I believe that computing can become much easier. One of the mechanisms to make computing easier is to use web applications instead of classic desktop applications. This is the first post in a series of posts that will document my experiences in migrating all my applications and data to web applications.

Why do web applications make computing easier? Well, they require no management: you don't have to deal with software updates, patches, new versions, etc. In general: they require no management from the end user's perspective. This is a good thing. As a driver, you also don't care about the management of your car.

Another advantage of web applications is device interchangeability: I can access my applications and my data, independent of the device I am using.

The biggest downside is vendor lock-in. If Google (or another company) stores all your mails, how do you migrate to another provider? More and more web applications have public API's that can be used to extract your data if you choose to migrate to another provider.

Other disadvantages of web applications include:

Poor support for offline operation: the new HTML5 standards do have some support for caching data and application code in your browser to support disconnected operation. However, this only works for applications that operate on small amounts of data and nowadays very few applications have support for HTML5's offline capabilities.Standards such as the File API that provide better support for offline operations are still in development and not yet implemented in today's browsers.
Emerging technology: HTML5 and associated standards are an emerging technology. They are promising but not yet fully implemented nor fully standardized.
Performance: running applications on a remote server introduces latency. Last mile connections need to get much better and service providers need to provision enough servers to create a better user experience.