In this second of the three installments of the topic, the author raises some situations in which new models of applications in network information management will be known. By: Osvaldo Callegari*
The so-called virtualization and abstraction of a computer's resources is considered the cornerstone of architectures. The possibility of virtualizing servers (with operating systems incorporated in an abstract way), desktop applications, in addition to a wide range of IT resources is used through the form (on-demand) on demand.
Along with these benefits, it is necessary to grow the bandwidth capacity of the Internet since all information passes through there. That large bandwidth space makes virtualization possible and in turn lowers the costs of technologies. Today they have migrated applications to run from conventional browsers.
We can see virtualized online logistics applications for example. Tools of extensive storage capacities emerge, practically unlimited, practical handling of volumes, snapshots and clones (copy and overwrite). Images can be deployed immediately, radically simplifying the remote installation of resources.
Composition of the layers of architectural services of the CC: The Internet made it possible to manage layers through the 3-layer model or n layers, it emerged as a general architecture. Cloud virtualization has created a new set of layers namely.
These new layers not only encapsulate resources but also define a new application development model. Within each layer in turn it is possible to define new business opportunities.
They can be offered as pay-as-you-go. There the acronym SaaS takes preponderance, which indicates the use of software as a service. SaaS is located at the highest layer, offered as a complete application and with the modality of a single instance of it on the provider's server. This service serves multiple companies, an example that we can cite of SaaS is Salesforce.com. Google apps offer basic business services such as shared email.
In the middle layer or Paas (Platform as a Service) an abstraction of a development environment is encapsulated, within it there is a web server, a distribution of an operating system such as linux and a programming environment such as Ruby or Pearl. Paas can function as content managers or they can specialize in any particular area. Google app Engine serves applications from the Google infrastructure. These services can provide great flexibility but are limited by the capabilities of each provider.
IaaS (Infrastructure as a Service) is located at the bottom layer and provides basic storage media, server systems, connections and routers.
A commercial example of this type of service is Amazon Web Services, which offer compute and elementary storage services. Joyent also provides infrastructure services with the possibility of scaling in the management of web services, including in turn the administration of rich applications written in Ruby, Php, Python and Java.
The primary advantage of CC is that it hides the complexity of the infrastructure from developers and end users. They don't need to know or know what's behind the cloud, they just want it to deliver the services they need. Some choose to build clouds of their own, in this particular business they must possess knowledge of the management and abstraction of fundamental resources and take the appropriate steps in choosing critical technology.
Within CC virtualization, a single server can behave like multiple computers with the same resources. Hardware and software are used on demand, meaning they are required when requested, later. There are important benefits with virtualization, for example the improvement of the environment, a single computer on can work as five computers or more, this important energy saving is a benefit in the future.
By encapsulating physical resources, virtualization solves the following challenges:
- Higher utilization rates.
- Consolidation of resources
Use/cost, energy and space savings
The electricity required to run enterprise-class data centers is no longer available in unlimited supplies, and the cost is in an upward spiral. For every dollar spent on a hardware server, an additional dollar is spent on energy (including the cost of function servers and coolers). Using virtualization to consolidate makes it possible to cut total energy consumption and save money in a significant way.
Space savings: Server extension remains a serious problem in most enterprise data centers, but data center expansion isn't always an option, with construction costs averaging thousands of dollars per square foot. Virtualization can relieve tension by consolidating many virtual systems into fewer physical systems.
Reduced operating costs: the average company spends $8 of
Maintenance, for every dollar invested in new infrastructure. Virtualization tends to change the service/management radius and reduce the total administrative workload and total operating costs.
Factors with collateral benefit: Cloud architectures can solve central security, privacy, and regulatory issues.
Platform Virtualization
Platform virtualization allows arbitrary operating and ambient systems to run on a given system. There are two basic models for this virtualization system:
Total – virtualization, is a complete simulation of the base hardware.
Para - virtualization, which offers a "mostly similar" model of the base hardware.
The two models are implemented as Type 1 hypervisors, which run directly on the hardware, and Type 2 hypervisors, which run at the higher level of a traditional operating system.
Each of the top virtualization vendors offers variations of both. It is important to realize that there are design and performance exchanges in any model of the virtualization system. Generally, the more abstract the fundamental hardware OS is made, the fewer specific hardware features can be accessed.
Augmented OS abstraction can also increase the potential for reduction and performance limitations.
Network Virtualization
Load balancing techniques have been a hot topic in CC because, as physical and virtual systems within the cloud scale, so does the complexity of handling the workload that is executed to deliver the service.
Load balancers group multiple servers and services behind virtual IP addresses. They provide resource-based guidance on service requirements and automatic failure recovery when the node fails.
While hardware balancers can run software-based balancers, their flexibility is always limited. Engineers also end up writing software that interacts with the hardware via a user interface or using a large number of computers to solve the problem.
A significant challenge in the Cloud Computing network is not only the provision of individual virtual network interfaces to a given virtual environment, but also the growing need for cloud infrastructures to offer more complicated virtual and private data centers, which provide a set of different system roles and the logical interconnection between those roles.
Application Virtualization
The great possibility offered by web container technology makes it possible to expand the productivity and flexibility of the developer. Such a container is the part of the application server that handles Java Server Page JSP files and other web-tier (layered) components. Apache Tomcat has an open source technology but has some limitations when you want to go beyond web-tier. This is when clustering, messaging, etc. is needed.
Glassfish Project provides an integrated collection of JAVA EE containers with full capabilities. Currently the developer chooses the operating system and the development platform. The growing public clouds will increase the offer to the programmer at the best of levels, the high level.
Deploying applications
With CC the abstraction of fundamental hardware is increasing, systems are deployed through a set of previous decisions. Monolithic applications can be integrated with ERP and CRM.
Machine Images
Increasingly, an image-based deployment model is becoming the primary mechanism for deploy application development loads on virtual resource reserves. Machine images contain user-specific applications, libraries, data, and associated configuration environments and are hosted within the cloud.
Perhaps the best-known examples are the Xen images. This deployment model is the foundation of Amazon Machine Images (AMIs), which are built around a variety of cores. They are easy to modify and share, the tools are provided by Amazon. They are stored in the free storage service as in Open Solaris.
Software deployment
With cloud computing offering increasing abstraction of fundamental hardware, a related but separate set of decisions must be made regarding how software and applications are deployed on cloud infrastructure. The cloud computing model is flexible enough to accommodate applications of all types and sizes, in all phases of development and deployment. Architectures can be the delivery platform for proprietary, monolithic applications such as ERP and CRM; the development and deployment platform for a new production of lightweight, dynamically classified applications built on open source software; or a source of IDEs and testing resources.
The packaging of software components based on software, data, server and storage reserves and other cloud resources makes possible the efficient allocation, re-use, and management of resources. The packaging system is essentially a software delivery mechanism that simplifies and accelerates the installation of everything from operating systems to applications to end-user data.
The Image Packaging System (IPS) for the OpenSolaris™ OS, for example, makes it possible to create images and install, search, update, and handle packages in the image. THE IPS can also be used to create regular packages and deposits and publish and manage packages for deposits. Increasingly, cloud operators and data centers are moving from installation system software on each server, choosing to deploy golden images on server farms. In any case, basic software configurations must be supplied in the system resource reserves.
In the last part we will give an opinion about this new technological model.
*If you want more information on this topic you can write to the author of this article at the email [email protected]
*The trademarks and names expressed in these articles are trademarks and registered names of their own companies.
