XdbE Twitter Updates for 2009-03-15

  • @wefollow #socialgraph #globalsocialgraphing #globalsocialgraph #openweb #web3.0 #socialgraphaggregator #openproject #XdbE #GSG #
  • RT @XdbE Connecting DOTs! and following all who shares #linkedDATA idea! http://tinyurl.com/cahhd5 #XdbEproject #
  • All who share #LinkedData idea (Tim Berners-Lee TED talk) welcome on board! http://tinyurl.com/cahhd5 #
  • Sharing is Caring – XdbE Model ://xdbe.com/blog/articles/sharing-is-caring-xdbe-model/ #linkeddata #
  • Sharing is Caring – XdbE Model http://tinyurl.com/6mwtgq #
  • Let Public Data Remain Public! Join XdbE network and connect DOTs! http://xdbe.com/faq/ #
  • Let’s aggregate all Social Graphs in one place, owned by no one but shared by everyone! http://xdbe.com/ #
Posted in Twits | Leave a comment

XdbE Twitter Updates for 2009-03-14

Posted in Twits | Leave a comment

FAQ Section

Just updated our web site! Now it has FAQ Section. The page is as always a work in progress, so keep an eye on it. To avoid a situation when there are plenty of Qs but they are never yours, please, post them in our Google Group and the most interesting of those will be re-published on our site. Remember that “questions are never indiscreet: answers sometimes are” ;)

Posted in Updates | Tagged , | Leave a comment

XdbE Twitter Updates for 2008-11-12

Posted in Twits | Leave a comment

XdbE.com launch

Today I’m happy to announce that we have launched our site, blog and wiki. After almost one year of working on project and blogging off line, we feel relieved to see the result and share with you our thoughts. But at the same time, to my horror, I realised that most of our posts need to be edited because most of them were written for internal use only. I hope it will be done ASAP and you could read more about XdbE project history and subject GSG related articles.
Enjoy! :)

Posted in Updates | Tagged , , , | Leave a comment

XdbE Operating Principles

At this stage, when public release of first XdbE website and after this XdbE software already seems to be not so far away it is about the time to formulate some basic governing principles under which the system is supposed to operate. Sooner or later a certain legal base should be established for XdbE to function and it is quite timely to discuss the ruling principles now.

Participation Open for All

It is the most important principle derived from the fact that data constituting User Circle AKA Social Graph are in Public Domain and not only the very body of data but also access to this data should remain open and public. XdbE provides this via means of XdbE DS Network which is the backbone of the whole system. The importance of the data itself calls for measures to ensure data quality and consistency. Hence, when participation in XdbE DS Network remaining Open in principle it should nevertheless carry some obligations aimed at providing quality service and ensuring fairness of participation. First of all it concerns service quality provided by hardware requirements. The condition of becoming a DS Operator should include an obligation to provide sufficient hardware and bandwidth resources to the network to ensure its normal functioning. Failure to comply with system requirements may lead to temporary or permanent exclusion of such a server from XdbE Network.

Sharing

To ensure fairness in sharing not only of system resources but the content as well the participation of Online Communities should be qualified too. It is not unimaginable that some Online Communities might wish to access the data and run their own XdbE DS without actually releasing their own data into the system. In theory this is possible by setting up a local XdbE DS that provides data to such a community by the way of passing it for an independently run DS. This obviously contradicts the principle of sharing and fairness too. In order to prevent this from happening this situation shall be clarified in advance in the following way: each DS found to be a part of a non-participating network shall be permanently excluded from the XdbE Network.

Open Data Access

Although the default way of getting structured access to the data set is XdbE DS there is also an option of requesting data by clients which are not members of XdbE Network. The system is designed specifically to provide load-balancing and reliability required for servicing such requests. However, one should not rule out the possibility that at some stages of XdbE development the demand may exceed available system resources. This inevitably leads to one sort of limitation of client requests or another. It could be done either by limiting the number of client calls served for one single IP or via issuing developer keys. Needless to mention, that such limitations should be lifted as soon as system resources allow doing so.
This basic outline provides a good starting point to begin developing the future XdbE Terms Of Service which shall be in place before the service becomes operational.

Posted in Articles | Tagged , , , | Leave a comment

Sharing is Caring – XdbE Model

XdbE is about sharing and this concept is so important that it makes a perfect sense to make it more and more explicit.
First of all who shares what.

  • Online Communities share their data sending it to XdbE.
  • XdbE DS and RS operators share their bandwidth and CPU resources
  • Developers share their knowledge and experise creating applications based on XdbE dataset

Now why would anyone want to share? What XdbE benefits are?

  • Online Communities just by the virtue of participating and sharing their data first of all get exponential exposure of their data but much more importantly they have immediate local access to the data sent by all other participating communities
  • XdbE DS operators are getting local access to the whole XdbE database thus having an excellent opportunity to develop their own products and services based on XdbE data set
  • XdbE RS operators are getting enormous amount of feel-good factor combined with good karma for just knowing that their participation makes the Internet a better place
  • Developers are getting a unique opportunity to connect their applications to reliable, fast and constantly updated load-balancing source of data

But what are the costs?
Let us consider the cost of the single heaviest element of XdbE Network, namely DS. The cost of setting up and running one XdbE DS should not exceed an average cost of a single dedicated server with perhaps software setup costs added. Altogether setting up a DS is within limits of several thousands US$ and estimated monthly running costs are just about a couple of hundreds US$ and for those who just want to do something good for the internet community there is always an option of running a simple XdbE RS which is even cheaper. For an Online Community one shall add a setup cost of DSC and DSS software but by all means total costs related to XdbE functionality are insignificant compare to total running costs of any online community. For developers apart from setting up pre-programmed XdbE DSC or alternatively writing their own one thee are no other costs at all.

With so many parties benefiting from XdbE it is bound to provide a substantial change in Web 2.0 infrastructure ultimately benefiting end-user as well.

Posted in Articles | Tagged , , , | Leave a comment

User Data – Privacy, Ownership and AGL

Developing such an innovative project as XdbE undoubtedly is one has to do quite a lot of rethinking of apparently obvious and common concepts to ensure that the project ideology not only is up-to-date, but preferably ahead of events. Among the most burning issues of the current phase of Web 2.0 development are issues of user privacy and ownership of user data. Such questions as “Who owns user generated content?” and “How far can you go without infringing user privacy?” need to be answered clearly and unambiguously at the formation phase of our project.

Very much in line with the approach advocated by TheSocialWeb TV among others we firmly believe that all user-generated content belongs to the user who generated this content. Whatever ToS may be put forward by some Online Community Hosts we believe that common law overrides those and the ultimate owner of user-generated content is that user him or herself. No Online Community Service (Social Network) could own this content or otherwise claim any intellectual property rights to it simply because such services only act as means of publishing for such content. At the same time one should respect decision of a user to publish his or her content at a particular service as although this content is not owned by the service but it in some way “belongs” to that given community. Any sort of user content aggregation should exist either as being set up by the user him or herself or with explicit permission given by the user.

XdbE does not deal with user-generated content at all but all that reasoning presented above is nevertheless quite important as it first of all explains why it is so and secondly defines our attitude towards user-related information circulating within the XdbE network. The intended purpose of XdbE is to provide real-time representation of global User Circle (Social Graph) or GSG (Global Social Graph) and hence all that is not within the scope should be left out not only because taking it on board would make the whole system heavier and less manageable but also because we can’t take what really belongs to somebody else. So, the principle number one: XdbE does not contain any user generated content.

But trying to build a universal User Circle (GSG) database we inevitably have to explicate our policy for dealing with user personal data. In our approach to this question we are leaning heavily on thoughts brought forward by LZZR in his Internet States of Presence article and elsewhere. First and foremost we do not pretend, like FOAF Project does for example, that information we are dealing with here is about individuals. Instead we firmly believe that user information available to us only reflects actions of an individual (or sometimes groups of individuals and possibly even some internet spam bots). Hence, we consider the basic block to build our User Circle database with to be not personal data but rather user account data. With this approach the boundary between personal data and user generated content might become sort of blurred. Without complicating things any further it is suffice to state here that we simply accept that we have no means to know how the kind of representation we operate relates to real-life individuals. More importantly, we don’t need this information to become personal information for obvious reasons of protecting user privacy. There are quite many different approaches to this subject currently in circulation. To illustrate let me give you a quote form ToS along which Yahoo! is currently operating:

Personal information is information about you that is personally identifiable like your name, address, email address, or phone number, and that is not otherwise publicly available.

If you read exactly what is written it means that all information ever published about an individual falls outside the scope of this definition. In contrast we prefer rather more restrained approach. Instead of turning XdbE into a database of personal information we decided to limit this sort of information to the necessary minimum. Here we had to balance the need to protect user privacy and the need to maintain usability and functionality of XdbE that would allow building basic applications without the need to retrieve information from any third-party sources. Although for building a global User Circle database it is enough to identify each node by its user name and community such limitation would make this database almost useless in terms user experience it could provide. Ultimately, attempting to minimize the amount of user information circulating within the system we came up with a concept of AGL. AGL or Age, Gender, Location is in our opinion the lowest amount of user data that both make the system functional and at the same time keeps user information below personally identifiable threshold. We obviously can not restrict, and more importantly we have no intention of prohibiting building third-party databases where user data driven from XdbE are combined with other data including those of more detailed and personal nature. However, all technical and legal responsibility lies with those third parties and not with XdbE.
Another tricky bit we had to decide upon was relationship status. Sometimes a user might want to keep some of his or her friends visible only for other friends or even only to oneself. Excluding such data completely from XdbE would make the system lame in terms of end-user usability substantially limiting data portability. At the same time we should obey the decision of a user to restrict the circulation of his relationship status. The way out is to assign three possible states of friendship relationship: public, private and friends only and treat them accordingly. All unauthorised and non-participating clients will receive only data on relationships marked as ‘public’ while participating and authorised clients will receive all. To ensure that the data does not end in wrong hands XdbE Terms of Service shall include a clause making it legal responsibility of all participants be it Online Community Services or XdbE DS operators, or application developers or anybody else for that matter to obey relationship statuses assigned by a user.
Needless to mention that data encryption algorithms built into XdbE data exchange protocol serve the same purpose of protecting user privacy as well.

Posted in Articles | Tagged , , , , , , , | Leave a comment

Introduction to Project Terminology – OCS, DSC, DSS, DC

There are several terms being for some time used internally for the project I think is timely to introduce publicly.
OCS is Online Community Service. This is any web service which has a community component no matter connected it to XdbE or not. OCS can be a social network, photo site, blog host, wiki, etc. Interchangable with more common but less precise term Social Network.
DSC is Data Source Clinet. This is a special application working inside the OCS intranet, which connects that OCS to XdbE DS Network. DSC knows how to send data, how to encrypt/decrypt data, how to interpret a response from XdbE DS Network. As an input DSC takes clean users data from OCS core software.
DSS is Data Source Server. This application has an inverse functionality compare to DSC and works in conjunction with it. It accepts data clarification requests from XDBE DS Network, decrypts them and provides clean requests to OCS core software. OCS core provides responses through DSC in the manner described above.
DC is Data Client. This is an external entity which requests user data from XdbE DS Network. This can be of course installed at OCS or at any other client interested in recieveing the data. It provides authorisation and identification as well as follows built-in load balancing algorythm.
Further reading can be found in XdbE Wiki.

Posted in Development, Updates | Tagged , , , , , | Leave a comment

On XdbE Network Data Security

One of the most important aspects of XdbE Architecture is system security. As XdbE core stores a lot of user data we should design the system in such a way as not to allow unauthorised data insertion (data poisoning) into the system or mailcious data change. Also we should not allow anyone from outside the XdbE Network to use service channels or to manipulate the network in any other way (DoS attacks etc.).

My approach here is to apply symmetric encryption to all the traffic inside the network and to all the traffic from an external network with no exception. The unique encryption key (more than 128 bit) will be used for every pair of servers within the XdbE Network itself. Maybe you want to ask why we won’t use asymmetric (public-private key) encryption. The asymmetric encryption is much slower and we can’t use it to encrypt all the messages. Actually we will use that kind of encryption but only for signature service impacting messages (like configuration change request).

Posted in Development, Updates | Tagged , | Leave a comment