While Apple presented iAd, advocating for an “application-linked” publicity model, Google bought AdMob, betting on for a “web-link” (outside the application itself) advertisement model. Google claimed that ecosystems such as that proposed by iAd reduce creativity and shift benefits away from developers.

Also, Apple relies on its nice market penetration with the iPhone. Google has not started to get the best for Android’s yet, but this strategy will surely be continued.

We are not sure which model will be the winner (if any, since they both may co-exist). The

Cloud computing is transforming the way we use the web but there’s still a long way to go before we make full use of the promise it offers. Projects Magazine – the leading research and development magazine in the areas of science and technology – interviews Professor Ignacio M. Llorente, the head of the DSA-Research group.

The NoSQL movement claims the SQL query language as the source of many of the scalability issues that we face today with traditional database approach. The misconception here lied in linking SQL to the RDBMS so tightly. Several alternative query languages, trying to solve well-known SQL problems, such as a document model, or simpler approaches, such as Key/Value query, are very common today.

If you design your data model to fit into a distributed model, you may find that SQL can be a useful format to manage data. A good example is Hive and Google JPA /Bigtable: the underlying data store is based on a scalable Key/Value store, but the front-end query language happens to be SQL-based. The key is the decoupling of the query semantics from the underlying data-store .

This middle path is tremendously useful for big companies (e.g. carriers) which data and applications have traditionally been managed in a relational manner (i.e. crossing searches by relying on the power of SQL). These big companies face huge scalability challenges and this decoupled key/value store offered with SQL-APIs seems a very appealing prospect to be exploited.

4CaaSt is a 15-million-Euro EU-funded initiative (EU grant agreement 258862) funded by the 7th FWP (Seventh Framework Programme) under the Internet of Services, Software & virtualisation (ICT-2009.1.2) area, aimed at creating an advanced PaaS Cloud platform which supports the optimized and elastic hosting of Internet-scale multi-tier applications. 4CaaSt embeds all the necessary features, easing programming of rich applications and enabling the creation of a true business ecosystem where applications coming from different providers can be tailored to different users, mashed up and traded together.

DSA-Research will join a consortium of Europe’s leading experts in cloud computing, including UPM, 2nd Quadrant Limited, BonitaSoft, Bull SAS, Telefónica Investigación y Desarrollo, Ericsson GMBH, FlexiScale, France Telecom, Universitat St Gallen, ICCS/NTUA, Nokia Siemens Networks, SAP AG, Telecom Italia, UCM, Universitaet Stuutgart, UvT-EISS, and ZIB. OpenNebula, developed by DSA-Research, will provide the EU FP7 project with a powerful technoloy to build IaaS clouds supporting automatic scaling of resources to run the business use-case scenarios in real world conditions. This news consolidates DSA-Research’s position at the cutting edge of cloud computing research worldwide, following an announcement two months ago of its participation in the EU’s StratusLab project, aimed at bringing cloud and virtualization to grid computing.

Ignacio M. Llorente

A somewhat old (April) article by Wolfgang Gentzsch discussed the monetary benefits of cloud computing.

The article makes an interesting approach to understand when and how using cloud computing is really worth it. Although some assumptions are arguable, the article reveals an important point: if your company is already achieving a high utilization of the available hardware resources (or high server consolidation if you are using virtualization technologies), then the economical benefit of the cloud is not interesting for you.

Dr. Gentzsch proposes two interesting use cases from companies which R&D depts. make high use of the available resources to perform computationally intensive simulations. Is not the cloud for those as well?

Here, we would argue that it depends on the available infrastructure they already have and the average time they want their simulations done. If with the resources at hand they are capable of reaching acceptable productivity, then the cloud is not definitely for them. However, if on some occasions they need to speed up some priority simulations without stopping all others (which may have been running for weeks and even months), then the cloud is a reasonable source for additional computational resources. Some on-going efforts already point in this way such as Nimbus, StratusLab or Amazon’s CCI. The cost of maintaining a non used infrastructure for these peaks is what makes the cloud interesting: again a problem of resource utilization.

A missed point is that of the promised eased management (e.g. automatic scaling, no need for hardware configuration, etc.) of the cloud. These operational costs certainly play an important role for cloud user institutions and deserve further consideration.

A recent post by Jim Metzler in NetworkWord revealed three key points that justify building software layers on top of typical virtualization solutions.

1. Security is still a major concern and virtualization introduces several new issues than deserve careful attention.

2. VM management is cumbersome and there is an increasing need for automating network provisioning too (having VMs ready is nothing if we have to wait for ages to get them properly secured AND connected).

3. Some services and applications “dwelling” in the cloud are highly dynamic and require tons of management work. Automation tools by the most relevant vendors do not fill this gap, being “driven by manually constructed scripts and templates that are as inflexible and cumbersome as the traditional command-line interfaces.”

The topic of Network as a Service was dealt with before in this blog and several articles by Jeffrey Chase’s group also support this idea. Regarding how to provide abstract operations close to our customers’ mindset and automate them, while allowing for the required dynamism, Claudia’s service lifecycle manager is already filling this gap. A new article in a jorunal and a subsequent post here will reveal its inner guts very soon. Keep tuned.

In a recent post we aimed at hitting Google where it hurts to preserve net neutrality or, at least, to let users decide and negotiate with ISPs on what should and would not be prioritized.

Google’s power source settles on their information about… US!!!! We are literally bombed with recommended and directed publicity. Statistical anonymity has been presented as the “do not worry, dude” panacea…

First hit: If they want our data, let’s make them pay for it!! Yes, they offer us a very nice email, search engine, etc. but they are also making a lot out of that without our permission on what they can or cannot see (how much is your privacy worth? If you share all with them, shouldn’t you get more advanced services in return? Why can’t you sell your privacy in exchange for money?). Let us imagine how easy it would be for Microsoft (or to any mid-low skilled developer) to develop a web browser plugin to obfuscate searches in a random but mutable manner (e.g. including random searchers per every single user actual search on Google an filtering results back to the navigator so that users do not notice any difference but google trends get terribly affected).

Also, do you want them to peep at you emails while you are writing? well, similar obfuscation mechanisms are easily implementable for using gmail as usual, but making Google’s information worthless or, at least, inaccurate.

Nurtured with imprecise information their recommendation systems will prove useless and their advertising revenues will consequently suffer a tough hit.

Second hit: Commoditizing their advertisements. Although undeniably effective, Google advertising system is very primitive. Gaining data from every individual’s social networks is a clear step ahead to deliver users with the ads they want but in the form they want. In other words, you may have never noticed that whenever you write a gmail email several targeted ads are there for your eyes only. More elaborate and seductive content creation and distribution mechanisms can simply be built on top of Google’s infrastructure by combining social knowledge with their web indexing service. Will they be willing to pay for content elaboration obtained from their own sources? I guess it will be betting on a winning horse.

Third hit: Collaborative content generation. Google is exposing contents that are already catalogued and kept on very specific locations (see, for instance, The Prado at Google Earth). Indeed, Google is benefiting from collaborative efforts to re-sell what we built. Collaborative efforts towards content generation alternative to Google are a third hit.

From this blog, we will keep you posted on several initiatives we and others are starting to hit Google right where it hurts.