Archive for July, 2009

Virtualized Infrastructures: What is the next step?

Monday, July 27th, 2009

Posted by Theodora A. Varvarigou

Various attempts are conducted nowadays to shape the vision of the Future Internet, the Internet of Things and in general future flexible and scalable infrastructures, which will support and enable the provision of services that utilize different kinds of resources such as devices and digital equipment (e.g. sensors). Virtualization of resources (including computational nodes, supercomputers, workstation-clusters, network elements, data-storages, internet networks, etc) has been identified as a way that allows for service provision with Quality of Service attributes as requested by the end-users, maximizing at the same time the resource utilization on the service providers’ side. But what has been achieved nowadays in the aforementioned context? Current environments not only aggregate but also virtualize a large number of independent and geographically distributed computational and information resources. Furthermore, the advent of Service Oriented Infrastructures that take advantage of virtualization technologies made feasible the provision of services by addressing at the same time a set of challenges such as live migration, fault tolerance, quality of service etc. Nevertheless, dynamic virtualized infrastructures also include a number of non-virtualized resources (mainly referring to digital devices). And these are the resources that have not yet been virtualized, even though they consist as main elements in such environments: devices and digital equipment.

Taking into consideration that emerging applications are tightly coupled with the aforementioned digital devices, service offering is limited (a simple example refers to the guaranteed network links through virtual environments that cannot be provided to digital devices that are not part of them). Future applications, such as social networking environments, will not only use but will require device-enriched environments, which results to an increasing demand of services and as a domino effect to the complexity of their composition and management since the number of components disposed to failure will increase as the size of applications will increase in terms of factors such as devices, hardware components, software components and geographical scale.

Is virtualization of devices a trivial task? The answer is certainly “No”! In order to make the next step on virtualization, a number of aspects have to be addressed, ranging from data management – since high quality input will be transmitted from sensors of all kinds – to network management (due to the exponential traffic growth), privacy, data integrity issues as well as the economic and societal impact of new networked resources (sensors, actuators, communication devices, etc). Applications in the near future will connect any kind of devices and taking into account that computing became a commodity, research efforts need to focus on how digital equipment and devices with specific characteristics can be virtualized. Sensors, displays, actuators and – generally – any type of electronic, interactive mechanism will be soon regarded as an integral part of dynamic virtualized infrastructures.

Bookmark and Share

Grids, clouds & interoperability – Some key questions

Monday, July 20th, 2009

by Craig Lee, OGF President

“Grids” and “clouds” are buzzwords that have come to denote broad concepts and user bases in science and industry. While these concepts have distinct characteristics based on “where they came from”, they are nonetheless strongly related when general distributed computing requirements are considered. At just the very top-level, we can ask the following questions:

Are there sufficient market forces today to make cloud interoperability or portability a reality? In fact, when we say “cloud interoperability” or “portability”, what scope do we actually mean? This could mean interoperability at the virtual machine level, the platform level, the service level, or even the data domain level. When will vendors decide that interoperability will make the market bigger for everyone (rather than maintaining “walled gardens” for their customers)? How long will users have to consider building their own “interoperability layers” to get things to work together?

What user experience do end-users really need and want? Can programming frameworks, problem solving environments, or portals shield users from dynamic provisioning and compute/data affinity/locality performance issues?

Where are the economically viable market segments that will support the development of the necessary tooling and infrastructure? Beyond e-commerce and web site hosting, what are the viable market models for scientific and engineering HPC, civil protection/services, urgent computing, massive data mining, etc.?

How can we manage the dichotomy between concept and implementation? Technical “movements” usually start with a grand vision and concept that are labeled with a buzzword. Over time, however, buzzword labels become inextricably linked and identified with the dominant implementations that are in use. Will tools such as XMPP be viable interoperability vehicles? Will it fair better than the basic web services stack? Will it support the range of economically viable market segments? Are there other options?

Bookmark and Share

Engineering on Cloud Computing

Tuesday, July 14th, 2009

Posted by Andrea Manieri

ICT services in the next few years will be even more based on the concept of service. Current hosting services will be reshaped as infrastructure as a service, software applications will be re-designed following the Software-as-a-Service concepts, Data and content Repositories will be accessible as a service. Though not yet well-defined, cloud computing Services are the most interesting perspective for future business, the right step to pave the way toward the knowledge society based on the so-called Future Internet.

An interesting debate is underway on what a cloud is or is not, and the relation between cloud infrastructures and grids, which in recent years have become the buzzword for IT in the future. Analysts, enterprises, and academics, all are attempting to define the concept of cloud Computing and cloud Computing Services. The Engineering Group sees the cloud as a way to deploy a Service-oriented Architectures. Exploiting virtualisation technologies, cloud Infrastructures should provide basic services to developers that can be put together to provide the requested Service-oriented architecture (SOA) at the right level of abstraction. Virtualisation technologies allow to abstract from hardware of different vendors, from different operating systems, from different development platforms and from different service development environments.

Each of (All) these level(s) of abstraction could be provided by a cloud Infrastructure, be it private, public or hybrid. Security and trust of a cloud, dependability of a cloud-based system, ease of use/deployment and discovery of software services in a cloud, single-sign-on mechanisms adapted to public or hybrid clouds, accounting and billing technologies specifically targeting the cloud model, quality assurance methods for applications running over a cloud, are examples of research challenges still facing ICT research players. On the positive (or negative) side of these issues, the popularity of cloud computing will become real business.

The Engineering Group is part of this evolutionary process, bringing the deep knowledge of the Italian and European IT markets, strong expertise in grid and service technologies and the capacity of its data centers. Engineering is heavily involved in NESSI, the European Technology Platform on Software and Services, it is contributing to the evolution of the European e-Infrastructures and is investigating in new techniques to ensure high-quality secure and dependable ICT services delivered by its own data centers.

Bookmark and Share

My Life in the Cloud

Monday, July 6th, 2009

by Avner Algom, Israeli Grid Technology (IGT), CEO

Prefix
Grid computing was created to provide a better cost/performance solution for distributed data processing by sharing network resources. Previously both academia and industry had to develop their own Grid software, and to manage their own network resources. Today we have an alternative for our network resources of CPUs and Storage, it’s called Cloud Computing, and we can use it on-demand and pay for it per-use. I hope that the following story will illustrate the new upcoming reality of Cloud Computing benefits and risks.

My Story
It was 9:00 a.m. 2011 when I awoke in the Dead Sea resort near Masada. It is a spiritual location in which, scholars found refuge thousands of years ago to write their ancient books. I went there to find the time and distance from the mundane to develop my new idea for a service that can improve our lives. I was sitting in front of my $100 laptop, writing a sacred code that would change the world. Via the wireless connection, I entered the requirements for my application development. Out of the 30 clouds that were on the Coodle clouds broker list, I had to choose the most qualified one. The US based Alayon Cloud had the best cost/performance I could find. Thanks to my credit card, a development platform (PaaS) was at my service. To reduce the cloud costs, I had to make my application code efficient enough to use a minimum of CPU cycles and storage space. It took me seven days, but then I was ready for the scalability and performance tests. I ordered 1000 CPUs and 10-Tera storage to make sure that my business application would not collapse once it took off and gained popularity. I requested a transfer of 30 Gigabytes of test data in 1 minute. I got the confirmation to start the transfer and in one minute and 5 sec (SLA violation!) I had my 30 Gigabytes on the other side of the ocean, on the Alayon Cloud storage. Anyway, I was pretty impressed with the bandwidth on-demand. Of course, in order to have a secured upload of my application to the cloud, I requested to activate the on-demand VPN in advance.

After fixing the application errors, it ran smoothly on 1000 CPUs, but to improve the performance I had to upgrade the CPUs by adding 10 cents per hour per CPU. Since using 1000 CPUs for one hour costs the same as using one CPU for 1000 hours, I decided to run a long stability test on a smaller quantity of CPUs. Meanwhile I got a call from my colleague; he could not find a reasonable cloud price for his 1 Terabyte of medical images that needed to be processed by the next day for an important remote brain surgery. Via our open Cloud community, I was able to gather 11 physical locations that in total had enough storage and CPU resources for my colleague. We activated the Cloud of Clouds virtualization system, with a 40Gbit network bandwidth. On top of the Cloud of Clouds infrastructure; we activated the Grid Computing service, so we could process the images as fast as they come. My colleague also reminded me that next week we have an important board meeting with the national universities cloud counsel. (NUCC). After 10 years we were able to consolidate the universities ERP systems into one unified ERP-SaaS cloud, saving the country a huge amount of money. The ERP-SaaS customization features enabled each university to customize its user interface and its unique features. Meanwhile, my lawyer forwarded me an email from Alayon Cloud that the government asked to review the data I saved within their territory. I was shocked, but since the data was only test data, I really didn’t care; I asked Alayon Cloud to transfer the data to another country that does not grant the government the legal right to look at my data. After a month of checks and preparations the new cloud service was ready for public announcement. I registered the service in the worldwide cloud directory with the right keywords and the pricing model. The potential customers that subscribed got an introductory offer to use the new service for a free one week trial. After a month, my new service got so popular that at peak time Alayon Cloud needed to move Virtual Machines (VMs) to a partner cloud facility with an average transfer of 10 VMs per second! Thanks to the unique IP6 VM addresses, we were able to follow the VMs on every cloud.

The publicity about the new service captured the attention of some bizarre people. The hackers were looking for my application data! Luckily cloud intrusion prevention monitoring system blocked the attempt. It sent instant messages about unauthorized access attempts to the service provider and to me. The International Cloud police were able to locate the hackers – students of a well known college. Based on my great cloud service success, it was time to make a tough financial decision. Via the Alayon Cloud financial monitoring system, I checked the list of banks that are Alayon Cloud partners, following a short competitive analysis; so the huge revenues went to my bank.

Bookmark and Share