The adoption of cloud computing has started to accelerate worldwide. But the first experience indicate that a complete replacement of the traditional model for cloud computing services will not happen soon. On the other hand the initial reasons that encourage the adoption of cloud, such as cost reduction, begins to be replaced by the agility with which the business is now available when using this template to create new and innovative processes and applications.

The model of cloud computing is just beginning to impact business and IT industry itself. But it still arouses suspicion and many companies are waiting cautiously that “early adopters” show the path. It is normal for this scenario. After all, much has been said about the distributed model, we currently use, and not everything promised has actually happened.

Besides, making predictions is always risky. And now again, we are facing stuck forecasts, such as “When the Paris Exhibition was over, no one else heard about electricity”. (Erasmus Wilson, Oxford University, 1879) and “The television will not work. People will have to stare at screen, and the average American family has no time for that. “(The New York Times, April 18, 1939, presentation of the prototype of a TV set).

Why were these things said? Were these people ignorant? No, scientists and professionals are well prepared. The point is, they do make wrong assumptions.

I recall here an interesting little story. In 1886, Gottlieb Daimler had just unleash the horses of a chariot and installed an engine behind her. He created the first car (or horseless carriage). His company teamed up with Karl Benz in early 1900s, tried to predict the size of the world market for these so smoky and noisy vehicles. After a careful analysis he predicted that, in the next century there would be around one million cars in use worldwide. But this prediction was proved entirely wrong. In 2000, there were over 600 million cars in the world! It was a long-term forecast, subject to weather, but still missed by a factor of a thousand. Why? The assumption used was wrong. They predicted that,  in hundred years the world population of professional drivers would be about one million and this would be limiting the growth in the use of horseless carriages. The assumption was that, every car need a professional driver, as at that time. That’s not what happened. Anyone can drive a car.

The same happens when we look at cloud computing from the perspective of the current model of IT and we cling to view this model as a simple modernization of outsourcing. But the possibility of a company and even creating new business processes without waiting for the traditional cycle of IT, and even without major capital investments, but only in operating costs (opex) opens up new and innovative spaces to be explored. Cloud can help you to transform your own business. Therefore, the cloud computing model should not be viewed solely from the perspective of technology, but as a strategic means to leverage new business.

However, the change will not occur by a “big bang”, but gradually. There are still barriers in the way and legal compliance still create risks for the business. Many providers do not have solutions and practical experiences. On the other hand, if you sit back and wait for things to happen, it can pass off good opportunities for competitive advantage. Also, frankly speaking, the data centers of some global cloud providers are much more advanced and secure than most enterprise data centers. So what to do?

In my opinion, companies should look at cloud from the perspective of a business strategy and begin to experiment with this model. If the organization already has familiarity with outsourcing, cloud can become a natural extension of your IT. The hybrid model, where applications on-premise clouds coexist with private and public clouds, will be the natural way for many companies. Expanding the current systems to operate in the cloud, while still maintaining the most sensitive data inside the house is a good way. This way you gain experience and cloud gradually detaches itself from the traditional model and it will become part of the company’s IT DNA.

The speed of adoption of cloud computing will depend on the culture and industry sector of each company. There are more regulated sectors and companies who are most aggressive in adopting innovations. There is no single recipe that suits all organizations. For example, a company can start by putting the majority of non-ERP systems and cloud with it, while reducing operational costs, you can immediately achieve greater agility to new demands by IT users. But the results do not come only with the tactical adoption of cloud. It represents a combination of the reorganization of IT to think more agile (Agile development) and not adhere to basic operational activities, standardizing and automating their operational environment (cloud computing).

Well, a few suggestions:

• Understand the nature of the cloud computing model and how to exploit this new model with new applications. Do not think about using the clouds just to do the same thing you do today.

• Adopt cloud computing if you design template applications for new and innovative systems. But keep mission-critical data at home, at least for now.

• Create a model of IT governance (policies, procedures and standards) covering cloud and do not forget the risk analysis in decisions to use the cloud applications that require certifications such as SAS70, HIPAA, etc., and service levels extremely hard.

• Start experimenting on specific and less critical workloads. Although cloud is still immature and primitive compared to five or ten years back, it already allows doing many interesting things. An example? How about BI in the cloud? It is perfectly possible to build a cloud platform for the demands of public business analytics, probably a fraction of what it would cost to build the environment in the home.

• And finally, have an attitude of being “open-minded” about cloud computing.

Today, we will discuss about the criteria that managers of a company should consider to make sourcing decisions effective for disaster recovery.

Basic Considerations

• Continuous pressures of senior management to improve the promptness of the recovery in times of tight budgets are forcing many companies to evaluate all the alternatives of data center services to reduce cost and risk of recurring recoveries.
• Fierce competition between providers of disaster recovery (DR) is causing a reduction in the price of services for customers who choose to evaluate the alternatives of service before the expiration of their current contracts.
• The more diverse the platform (e.g., combinations of Z and P series IBM, along with the heritage of midrange systems, UNIX, Linux and Windows), the more likely the continued use of rehabilitation services providers represents a more realistic approach to maintain a policy of mitigation consistent.

Recommendations

• Determine the changes in service and infrastructure necessary to meet Recovery Time Objectives (RTOs) and Recovery Point Objectives (RPOs) in 24 hours or less, is most important.
• Before the expiry of current DR contract, analyze alternative service provider to find out what is the best option to manage the service levels required for recovery.
• Manage the expense of disaster recovery to ensure that the infrastructure, processes and human resources are necessary to achieve and maintain the goals of RTO and RPO of applications that are available.

WHAT YOU NEED TO KNOW

In the last three years, we have seen several clients doing analysis of DR services long before their contracts expire. It is increasingly common because vendors want to know (up to six months in advance) if the service contract will be maintained or end. If a company chooses the second option, it is important to ensure that the transition from old to new provider should be performed promptly.

Due to the inherent complexity of web applications and the continuing effects of changing data centers, the need for a longer period of testing is also included in the essential criteria for an analysis of alternatives, especially for the evaluation of alternative demand and exclusive service.

ANALYSIS

In the last five years, a lot of people have started sourcing DR and emerging data center technologies like server virtualization and private cloud computing. These and other alternatives help to reduce the barriers of investment and operation for the use of disk-to-disk-replication, and have easy choices for specific services.

More important is that, many companies believe, a longer period of testing is necessary for IT to provide the best guarantee of recovery for complex services and applications online. This is because complex online applications require a higher level of dependence on software testing and data, mainly due to the accelerated pace of change in their datacenter. However, it is advisable to schedule a test period months in advance.

As these requirements keeps on changing, an increasing number of companies are starting evaluations of alternative DR before their current service contracts of DR expire. If a company decides to discontinue service, it is crucial to ensure that the process of transition from an old to a new provider  is defined and coherent.

However, simply evaluating alternative internal services, hosted or cloud-based does not imply a reduction in the cost of management. It is also important to ensure that the correct criteria is guiding the evaluation.

This include:

• Availability of data center for continues recovery.
• Availability of equipment and configuration of infrastructure required (for example., Server, storage and network).
• Support for failover (automated access to servers only when required) network services quickly, if necessary.
• Availability of production data in a period, consistent with the goals of RTO and RPO.
• Results of recovery tests consistent with RTO and RPO goals.

Choosing a disaster recovery services provider after doing proper research is clearly a more appropriate choice.

However, it is more likely that you may need more than one provider for effective management of recovery. For example, an external provider host may be the logical choice for the case of continuous availability of data center, a direct supplier of equipment is the most appropriate alternative for IT infrastructure equipment, the simultaneous use of multiple telecommunications providers is the best option for failover of network services, as well as test management, as it has professionals with practical knowledge of business applications and the priority of data recovery, and better understand what levels of recovery test will be effective and consistent.

At the same time, we found that the mix of internal computing platforms can often play a key role in determining the most viable approach. In conversations with customers, we found that the more diverse the mix of platforms, the easier it will define which vendor best suited to the case. This is because the task of providing hardware systems for medium and large companies, as well as managing a data center for recovery is not feasible for many companies, which makes the traditional approach of separate services for disaster recovery much more practical.

Moreover, the growing use of server virtualization may also mean that a smaller number of physical servers can expressly be needed in the recovery of a site to keep the workload of a specific application balanced. Regardless of how diversified is the corporate computing environment, certain issues should always guide sourcing decisions.

They are:

• The choice of sourcing basically need to provide the means by which levels of RPO and RTO services are effective and predictable critical applications are defined and maintained.
• We recommend that companies should use the 80/20 rule pragmatically to ensure that current budgets are allocated to support DR infrastructure, processes and human resources needed to achieve the goals of RTO and RPO, and at least the most important applications.

Ideally, the choice of sourcing should be based on a combination of sensible economic criteria of effectiveness and efficiency, as well as technology and vendor-specific advantages and disadvantages of management operation.

A research done by experts from different areas in the data center has published a compilation of the 10 countries that are most conducive in building a data center.

Where is the best location to build a data center?

So, according to experts, the best regions in the world to build a data center are as follows:

Where is it worse to build a data center?

Unfavorable regions for building a data center are based on the following criteria:

• Access to reliable energy sources, taking the growth prospects of the power consumption (data center scalability) into account.
• The use of existing and emerging cleaner, renewable energy
• The possibility of free cooling
• The possibility of submarine cables,
• The presence of stable, scalable network resources
• Favorable conditions in the country as a whole (stability, security)
• The risk of natural disasters with their consequences,
• The project cost (capital and operating costs)
• Availability of human resources,
• Availability of vendors (The official representative of manufacturers of equipment).

The Result For 10 Worst Countries:

Nigeria
Iraq
South Ossetia
Georgia
Yemen
Somalia
Papua, New Guinea
Haiti
Myanmar
Vatican

10 Worst Locations:

Mount Etna
British Antarctic Survey Station, Halley
Pyongyang
Galapagos Islands,
Los Angeles
Harare,
Venice
Kabul
Machu Picchu
Atlantis.

An interesting collection of unfavorable data center locations. Does any one have any desire to build a data center in Myanmar or somewhere on the slopes of Mount Etna?

Knowing the profile of customers has always been a competitive edge, even in past times. However, the dynamism of the modern market demands much more than that old custom of being personally greeted by the owner of the company at the door of the establishment.

In times of major mergers and acquisitions, organizations are faced today with dozens of islands of information that contain data for business strategy, often scattered and disconnected in transactional systems from many different flavors.

It would be virtually impossible to qualify the business data in order to make them useful, if information technology was not also on the rise sharply. Fortunately, technologies such as Data Centers, Data Warehousing and Business Intelligence are already mature, with over 30 years of existence, but still have gaps in functionality with respect to one of the most strategic goals: Know your customer.

In fact, the park now has a computer alphabet soup that usually causes much confusion among users or managers. Terms such as BI, DW and CRM are more popular, but initiatives such as MDM, IDC or PIM are still virtually an enigma to some.

Of all these acronyms, which in our view is most useful to study at the moment is just the IDC – Customer Data Integration – which can be understood simply, as a set of technologies applied to a single goal: to better recognize your client at the corporate level.

To explain clearly the meaning of IDC, we must first know the term Master Data. Master data is the key information about entities that are critical to your business. Databases containing customer information are the classic example. For large utilities the problem may not be as related to customers, but with difficulty in managing extensive lists of products for strategic or compliance objectives. In short, it’s technology to manage critical data and sensitive to each company at the corporate level.

We can define IDC Conceptually as technological infrastructure that aims at the integration, harmonization and management of master data. For management means from the application of business rules validation, enrichment and unification of corporate data to the management and strategic use of quality indicators of such information. Thus providing a rich source of master data due criticism, unified, and treated to meet the business processes. The exposure of this information typically involves the ability of the platform of choice in engaging these resources data buses and corporate services via SOA.

It is, therefore, a comprehensive set of computing resources combined to meet the common goal. There are hardware and software experts who are involved with integration capabilities and quality of data to create an identity resolution engine, and especially the involvement of people and processes to define the requirements of treatment can be determined as a corporate coverage.

Good joint depends, of course, experienced companies in projects with complex data so that the internalization of practice does not become another complication to the already troubled daily life of IT professionals of companies. The fact is that the abbreviation IDC, which is in focus at the moment, the world of business intelligence and its driving force could well be associated with that old practice which reads: Know your customer.

Aligned with other initiatives – some already well known as CRM, ERP – and some not so popular as the PIM (Product Information Management), the IDC technology is able to strengthen support for the complex process of decision making with respect to actions directly aimed at people who interact with the organization, be they customers or employees or related entities such as providers, for example.

When we purchase a product or service that requires a significant investment, it is essential to know in depth the characteristics or properties, if you have a product, and the advantages and security provided if a service.

When it comes to data center, there are multiple factors to consider:

First large data storage capacity. Not only that, but good exchange rate for the same. If we shut down  the server, but far from databases, the set speed will suffer.

The physical data center runs on electricity. It should have assured electricity even in disaster conditions (cuts supply network, fires, floods, etc.). It should not just have a efficient UPS but also the authentic electric generators to ensure supply.

Backups of course. The data is stored redundantly and must be accessible in a transparent manner if a database or server goes down.

The redundancy is applicable not only to the above, but for all the physical devices involved in the management and transmission of data.

Very important: We not only have to think about security and communications software , we must also think about good physical security for all equipment and wiring.

A data center that does not have a professional team, or have 24/7 support, the proper operation of the service is not very reliable. The provider should maintain and update the software (this task must be fully supported by this team). The provider must be available to customers to solve any problems that arise.

Increasingly, data centers are concerned to comply with the recommendations within the scope of  “Green IT” and this will benefit customers to lower energy costs through efficiency.

Finally, ensure the financial solvency of the company that provides service and … before you sign the contract rather read the Terms of Service and the fine print.

How to get all this information is easy to imagine. Almost always, gather this data from the service provider.

In a business world that does not forgive inefficiencies, the concept of value for the business becomes increasingly important and decisive in making executive decisions. Analyze the adoption of a technology or computer concept like cloud computing services, involves assessing their real value to the organization, whatever the valuation models, such as ROI (return on investment), ROA (Return on Assets), value the opportunity and so on.

Of course the value of technology depends directly on the importance of technology to business results. How quickly an application should be implemented to allow the creation of a new service or support the launch of a new product? The rate imposed by globalization and the Internet does not accept delays as found in the time of batch applications or client-server.

New technologies, speed and choices

Unfortunately most companies are not prepared to meet the demands at that speed. Allocating computing resources is not in the traditional models we have adopted to manage the technology infrastructure, a simple task.

Often it is necessary to enter into a process of selection and purchase of new dedicated servers. Other times, even when servers are available, you need an exhausting work of preparing the environment.

The infrastructure management models adopted by companies like Google, allow you to manage massive data centers automatically and efficiently. Why not use this model in enterprise data centers?

Models and solutions from the cloud providers

Consider the viewpoint of financial executives, the CFO (Chief Financial Officer). CFOs are keen to solutions such as cloud computing model and its pay-as-you-go “because this model exchange capital investment (Capex and Capital Expenditure) for opex (operating expense). The result is a cash flow, which is much better than the traditional model.

The CFO does not need to sign any checks before being able to have the computing power. Instead, he signs the checks as they consume computing resources.

The financial risk is also much lower, because, with the traditional model he spent the advance money on technology without knowing whether the result is even expected. In the cloud model, financial risk is monthly (use and pay) and he can more closely monitor how the money is being spent. Moreover, there is no depreciation of the asset. Finally, the CFO’s point of view, cloud is the model of your dreams.

Eliminating financial risks and benefits knowing

Although the economic appeal of cloud, converting capital expenditure (capex) in operational expenditure (opex) is very strong, the model of pay per use “pay as you go” quite adequately capture the economic benefit of the proposal.

The hours of computing acquired by a cloud can be distributed unevenly, i.e., we can use 80 hours of server today and tomorrow only 5, and pay only 85 hours.

In summary, the elasticity super provision eliminates the risk of excess or scaling (that generates under-utilization of resources) and sub provision (scaling down the requirements, creating bottlenecks) are differentiating factors and drivers of Cloud services.

Demand and possibilities

The ability to allocate resources and remove dynamically, in real time, to reconcile the demand with their resources adequately.

A simple example shows the benefit of elasticity. Suppose we estimate a demand in the peak period in 500 servers, but on the other hand, we also have a period of little use when we need only 100 servers.

The average use will be situated around a capacity of 300 servers per day, or 300 x 24, 7200 server times / day. But the cost of the servers, due to scaling the peak, is 500 x 24 or 12,000 servers per day. A factor 1.7 is higher than necessary.

In the cloud model, you pay only what was actually consumed. Gains can be substantial.

Another example. Suppose that 10% of users who receive poor service are online shoppers who drop out of the transaction and leave the basket, leaving the store. Business opportunities are lost.

The shop estimated a peak of 400,000 users (1000 users per server x 400 servers), but promotion has pushed demand to 500,000 users trying to access the site. In excess of 100,000 users who received a bad service, by our estimate, 10% or 10,000 have abandoned the transaction.

If this access profile is maintained, the store will be losing 10,000 transactions per hour, for a long time. In addition, negative ads, common in the Internet world, you can remove any new buyers.

In the model of cloud, the capacity fluctuates according to demand. There will be 500 servers when the store has 500,000 users, 400 servers when 400,000 users and 100 servers when only 100,000 users are accessing the store.

Do not lose customers and you pay only for what the shop demanded resources. That is, IT spending is directly related to the revenue of the business.

Economy in investments and according to the customer profile

Furthermore, as there is payment for the use, applications with different profiles of performance are collected in a much more appropriate. An example: an application that requires intensive CPU activity is low and I / O will be charged primarily by the cost of CPU time. Have another application, intensive I / O, will be billed by the volume of MBs transferred to and from the cloud.

There is another benefit of the Cloud hosting model through which it passes unnoticed: technological developments mean that the life of a machine as a server is relatively short. In a few years or months, a new server, with lower operating costs (such as energy and space) hits the market. In the traditional model, the cost of replacing the machines can be relatively high, preventing many companies from obtaining these benefits due to high capex required.

In the cloud , the provider (they have massive data centers) may replace equipment more easily and pass on these gains to customers due to competition in the market.

Let’s look at another positive point: the model of cloud company removes all the hassle and expense of administering all the technological stuff that usually is not their core business.

Operate and maintain a plethora of servers takes time and money. Offsetting this service to the cloud provider, the company concentrates its activities in IT in the business.

Cost reduction is a magic word when it comes to business and IT managers are not left out of this matter. Perhaps the ultimate goal really is to achieve better results, but reducing costs and enhancing business success. And that, thankfully, is no longer a mythical subject – may already be a reality, regardless of business type and company size.

How?

To get the answer we must first address some concepts such as convergence. In IT, more specifically when we talk about networks and servers, convergence can be defined as the willingness to travel with safety, efficiency and security, voice applications, video and data over a single network, wired or wireless. The concept is to mount a network end to all elements virtualized management with automated, optimized resources and guarantee availability.

Generally, corporate data networks are separated. A company has a telephone network with its traditional PBX for communication between employees and the outside world, the other for data network, which runs its data processing systems and shares the internet access.

A third network may exist, for information on video security systems and, finally, there is still a fourth network, via telecom operators to videoconferencing systems, among other specific systems that require connectivity solutions individuals.

Thus, companies reserve twice the energy required for installations and machines, as consumption is unpredictable. But what can be done to significantly reduce operating costs of these networks?

You can monitor how energy is consumed and therefore reduce consumption.

As discussed earlier, yes. This is no longer “something of the future”, but the present. Currently, the market offers several possibilities to obtain a converged infrastructure, ie have a model that virtualizes and converges IT environments of servers, storage and networks with data center facilities, creating a single instance of shared services, allowing greater flexibility and agility to meet new business demands.

A good example is the technology and Converged Infrastructure at HP, which relies on Networking, Storage, Dedicated Servers and Services. Thus, it is possible to serve companies of all sizes, with the most diverse portfolios, offering the best solutions, always respecting the client’s legacy and evolving according to their needs.

On one hand, proprietary infrastructure networks keeps costs high. On the other, many open standards must be followed exactly to make sure equipment from different manufacturers are compatible.

It takes trust in the solutions, they should have greater wealth of functionalities for users, lower cost of maintenance and especially safety. Modern architecture, open standards-based HP, evolving and innovating while consistently decreases the cost curve in enterprise networks.

Thus, reducing complexity, cost and exceeding their expectations in IT Management are the rules, with simplicity and efficiency, with open architecture, with maximum network security, with the converged infrastructure.

As a result, it is possible to guarantee the long-awaited reduction of costs, so simple that you will invest more in your own business instead of worrying about maintenance.

More for less with benefits. So it should be….

Or do you prefer the opposite?

Service-Oriented Architecture (SOA) – is a term that explains two very different things. The first two words explains a methodology for software development. The third word is a picture of all the software assets of a company, like an architectural plan is a representation of all parts that together form a building. Therefore, service-oriented architecture is a strategy that proclaims the creation of all software assets of a company using methodology-oriented programming services.

What is a service?

Services are portions – or components – software, developed so that they can be easily linked with other software components. The idea behind these services is simple: Technology should be expressed so that business people can understand, and not as an application enigmatic.

At the heart of the concept of services, it is the idea that makes it possible to define parts of the software code portions which is significant enough to be shared and reused in different business areas. Thus, some tasks become automated – for example, send a query to a credit reporting website to find out if a customer qualifies for a loan. If the programmers at a bank can abstract all that code to a higher level (ie, take all the code that was written to perform the verification of credit rating and package it into a single unit called “get credit rating” ), they can reuse that portion for the next time when the bank decides to launch a new loan product that requires the same information, instead of having to write code from scratch.

To achieve this, the developers created a complex wrapper around the bundled code. This wrapper is an interface that describes what the chunk does and how to connect to it. It is an old concept, dating from the ’80s, when the object-oriented programming has emerged. The only difference is that, today the software objects are much larger and more sophisticated.

In the U.S., for example, the service “get CSR” (get or obtain customer service record) is a complex clutter of software actions and data extractions that uses the infrastructure of enterprise integration to access more than 25 systems in four data centers around the country. Before creating the service “get CSR”, developers who needed that critical piece of data had to create links to all 25 systems – adding their own links on the complex web of links already hanging off the popular systems.

There are a lot of different ways to connect to the services, as like custom programming links or integration software suppliers, but from a couple of last years, a set of communication mechanisms to software known as web services, built upon the ubiquitous World Wide Web, has become a increasingly popular method for integrating software components.

What is the difference between SOA and Web services?

SOA is a comprehensive architecture for building applications within a company – think of an architectural project – but in this case, the architecture requires all that  programs which are created with a methodology for developing specific software, known as service-oriented programming. Web services are a set of standard mechanisms of communication that are created on the World Wide Web ie, Web services are a method to connect and communicate. While SOA is an IT strategy.

How do I know whether I should adopt an SOA strategy?

Being an architectural strategy, SOA involves more than mere software development. The creation of an architecture based on a portfolio of services demand that CIOs make a compelling case for enterprise architecture, a centralized development methodology and a centralized team of project managers, architects and developers. It also requires a CEO and an executive team ready, prepare the ground for the IT staff can dip into core business processes. Understanding these processes and getting buy-enterprise sharing are the cornerstone of a business transformation based on SOA.

Governance is critical. For services to be reused across the enterprise, there must be a methodology for developing single, centralized software so that different areas do not create the same service in different ways or use incompatible connectors. There must be a centralized repository so that developers know where to look for services – and IT will know who they are being used. The services must be well documented so that developers know what they are, as it integrates them and rules for using them. Some companies, for example, charge fees for use of services and create performance agreements to ensure that the services work well and do not burden the corporate network.

Most companies that have progressed on the path to SOA has created a centralized architecture group to choose processes that will be trained to service and see different areas of the company to create specific services. The centralized group also creates a convenient mechanism for governance. If all service requests have to go through the architecture group, the methods of service development, designs and performance agreements can be more easily managed.

Companies that have had more success with SOA so far are those that always have success with technology: big companies with big IT budgets whose business is based on technology (telecommunications and core banking services). They also tend to have business leaders involved in the IT field and willing to support their projects. For companies without these advantages, SOA may not be everything it promises.

For smaller companies, companies that bet the farm on integrated packaged applications and companies that adopt sound strategies for application integration, SOA has nothing to do with “when”, but “if.” CIOs must be careful because in service-oriented architecture, the elements of “service development” and “architecture planning” are distinct but not independent – must be considered and executed in parallel. Services that are reared in isolation, without taking the goals and architecture of the business into account, may have little potential for reuse (one of the most important benefits of SOA) or fail completely.

What are the benefits of SOA?

First of all, the benefits of service-oriented architecture must be contextualized. If your company is not large or complex, that is, if not more than two primary systems that require some level of integration, it is unlikely that the model provides great benefits. Amid all the hype around SOA today, can be forgotten that the development methodology in itself is not useful – is the effect it has on a redundant infrastructure and complex that it does. The architects say that creating a good service-oriented application involves more work than the usual application integration. (Research shows that SOA is being used to integrate traditional application in most companies.)

Thus, the development of SOA generates an extra initial cost. For this work produces benefits, SOA has to work on eliminating any other point, since the methodology itself does not generate benefits for business. So the first step is to find out if there are redundant and poorly integrated applications that could be consolidated or eliminated as a result of adoption. If this is the case, then there are potential benefits.

To understand the overview of the benefits touted by SOA, you need to examine it at two levels: first, the tactical advantages of service-oriented development and, second, the benefits of SOA as a strategy of global architecture.

Advantages of service-oriented development:

Software reuse: If the code package is a service that has the size and scope right (a big if, say veterans SOA), then it can be reused for the next time when the development team need a specific function for a new application that wants to develop. Say a telecommunications company has four different divisions, each with its own system for processing requests. All these systems perform some similar functions, such as credit checks and searches of client records. But, considering that each system is highly integrated, none of these redundant functions can be shared. The service-oriented development is gathering the necessary code to create a version of “credit check” that can be shared by all four systems. The service can be a totally new piece of software or a composite application consisting of code or some system of them all.

Anyway, the ‘composite’ is wrapped in an interface that hides its complexity. The next time when that developers need to create an application that needs credit check, will create a simple link to the new application. They don’t need to worry about connecting to individual systems – in fact, need not even know how the code has been included or where it comes from. Only need to create a connection to it.
In a company that constantly develops new systems that rely on similar functionality – an insurance company with many different divisions, each with slightly different products, for example, or a company that is getting ever more – the time saved in the tasks of developing, test and integrate this same functionality of software is an advantage.

But reuse is not guaranteed. If developers in other parts of the company did not know that services exist or do not trust that they are well built, or development methodologies vary within the company, services may decline and not be repeated. Companies adept at re-developed mechanisms of governance – development teams, centralized, single methodology for developing repositories and services – to increase your chances of reuse.

Sometimes, however, the service is simply not well designed. It does not perform operations enough to be widely applicable in business or attempts to perform other operations. Either the developers did not take into account that others may want to use the service in different ways. For professionals in the business, proper sizing of services – also known as granularity – is as much an art as a science, and bad granularity can dramatically reduce the possibilities for reuse. A research estimates that only somewhere between 10% and 40% of services are reused.

Increased agility: Even if services are not reused, they can add value to facilitate the modification of IT systems. There are no redundant applications or multiple business units clamoring for services. But with the division of the application process in discrete services, each component can be isolated and modified as necessary to cope with peak demand. In the system, a server responds to spikes in activity during each phase of the application process, transferring capacity to the specific service that needs it most.

Benefits of an SOA strategy:

1. Better alignment with the business. The service-oriented architecture is an overview of all processes and flows of a company’s business. It means that business people can visualize, for the first time, as the company is built in terms of technology. When IT projects are presented in terms of activities and business processes and not in the form of complex applications, the business people can better appreciate and support IT projects.

The grand vision of SOA is that, when fully enable IT services to the important processes of a business, business people can take control of change and merge the different services in new combinations of the process itself. But this view is still many years away.

2. A better way to sell architecture to the business (and IT). There are times that enterprise architecture has been the concept that dare not speak its name. Some CIOs get to the point of not using the term with colleagues for fear of scaring them, lose them or simply bore them. EA has always been a big undertaking, costly and complex. Your ROI, it is often unclear to the business. Standardize, map and track IT assets clearly does not make the business more flexible, capable and profitable. As a result, IT architecture efforts often fail or become completely IT-centric. The service-oriented architecture provides value to the business, enterprise architecture in the old, rarely was just a vague promise. Reuse, increased productivity and agility in IT infrastructure and a set of software for specific business processes are the bait to sell an enterprise architecture initiative for the business. But remember that architecture is not for everyone. Small or extremely decentralized companies may be unable to justify a centralized team of project managers, architects and developers.

How to balance the need for planning in SOA architecture with the need to prove the business value quickly?

Architectural planning is time consuming. The service-oriented development based on principles of programming known and widely available technology standards (SOAP, HTTP and so on), can be much faster. But the two have to happen in parallel, the experts teach.

How will SOA affect IT group?

If you have a decentralized company, be prepared to fight. SOA leads to centralization. In fact, ask centralization. You need someone to lead, you must have an individual or a small team managing the architecture. You need a group, a series of research and someone to ensure that development groups should adhere to the methodology for service development.

As the portfolio of services grows, the development process can begin to resemble an assembly line. “It turns into a factory”.

According to a survey of CIOs, only 25% of respondents have the teams they need to adopt service-oriented architecture – 49% plan to have or already have training programs for current staff to work at full steam.

While most of the major network providers have yet to predict the transition of data centers in the same type, unstructured network in the future, we will need to work on the practical application architecture, which will replace the hierarchical networks.

According to me, the future single-level network, as opposed to the standard three-tier network can best meet the challenges of management and productivity encountered by operators of large IT networks.

In ten years all the new data center will look like a single architecture. Technology pooling, also known as cloud hosting, are the main driver of this restructuring and modern hierarchical architectures cannot cope with their tasks.

If there is an increase in network traffic, the number of servers increases linearly and  the number of switching equipment grows exponentially.

Every time you double the number of storage systems and servers in data center, the number of patch panels increases four times. This means that the complexity of the network will grow exponentially. Today, operators of data centers have to raise the performance and reliability and to improve the quality of services provided. At the same time they need to strive to continuously expand their data centers, while controlling and minimizing costs. All these problems can be solved by radically simplifying the IT network.

This will be one big switch, comprising a plurality of physical devices, each port will be connected to only one port. It will unite all the devices into a single coherent network, that will be managed by the operating system, and a single management interface. The proposed solution is called the Stratus, and will have a modular type expansion.

Structured cabling system DPC (Data Center) should support the work of various applications, including those that are only planned to use in the future. Speed in data centers are increasing due to the use of server virtualization technology and the use of this technology leads to an increase in requirements for the cable system. Using an optical cable in data centers is primarily due to the fact that a large number of reliable and working technologies and solutions that use as a medium for data transmission singlemode and multimode fiber , the transmission of large information flows and the distance is much more than it can reach copper cables.

Optical cables have no problems with EMC and crosstalk occurring between the twisted pairs and copper cables, which makes use of copper cables in data centers .

The diameter of the optical cable is less than copper and it allows the use of cable channels for the optical cable of smaller sections, which allows better use of space in data center, reduce the costs of cable channels. Also reduces the load on the mount cable channel.

Using an optical cable increases the efficiency of the cooling system: optical cable run cool, creates fewer obstacles when placing cable channels under the raised floor, electric power consumption of the active ports of equipment on the optics, copper below.

A well-designed cabling in data center allows the owner of the data center to solve the problem of scaling data center, with no problems in connecting with a new network and equipment storage. And to solve the problem of scaling the data center without using an optical cable is impossible.

Multimode Or Singlemode Fiber In The Data Center

In the data center in accordance with the requirements and recommendations of all standards are allowed to use two types of optical fibers: single-mode fiber and multimode fiber.

Single-mode fiber is cheaper than multimode fiber, but due to the fact that the ports are active equipment to fiber mode are much more expensive, the single-mode fiber is not widespread in the data center . Most single-mode fiber is used for connecting services from the operator, who delivers it to the room of the cable entry point in the network interface. In contrast to SCS for commercial buildings in the cable system, data center allowed the use of single-mode fiber in the horizontal subsystem. Using a single-mode optical fiber may be required when filing service from service provider directly to the port of the active equipment installed in the machine room in the hardware distribution area.

Most of the data center equipment is used with ports that support the work of multimode optical fiber as the price of the port of active equipment for multimode fiber is cheaper than the ports of the active equipment for single-mode optical fiber.

The type of multimode fiber depends on the size of the data center and the applications that will be used at the time of design and in the future.

TIA EIA 942 Standard allowed the use of multimode optical fibers with core and shell 62.5/125um and 50/125 categories of OM1, OM2 and OM3. However, the TIA/EIA-942 standard recommend using multimode fiber optimized for use with a laser with a ratio of broadband multimode fiber is not less than 2000 MHz • km at 850 nm, that is recommended for multimode fiber 50/125 categories of OM-3.

European standard EN 50173-5 for SCS in the data center do not permit the use of 62.5/125 micron multimode fiber and even 50/125 OM1 category, and also as the standard TIA/EIA-942 recommend using OM3 multimode fiber category.

The international standard ISO / IEC 24764, which is not yet approved and which is designed with the latest requirements for SCS in the data center , generally spelled use in a data center of a multimode fiber does not lower OM3 category, ie the data center , which is designed should meet international standard requirements.

The use of OM3 multimode fiber category ensures data transfer rate 10Gbit/sec at a distance of 300 meters, however, if the insertion loss in the optical channel at a distance will not exceed 1.6 dB.

Multimode fiber category OM4

American and international working committees on standardization shall consider the adoption of a new category of multimode optical fibers. TIA / EIA working document TIA-492AAAD, which will include the specification of the characteristics of the fiber category OM4. And the international working group is working on a standard IEC IEC 60793-2-10 with a new type of multimode fiber. While the new category OM4 multimode fiber has not been accepted, but working groups have agreed on the minimum value of the coefficient of bandwidth for multimode optical fiber of a new category for VCSEL laser with 850 nm radiation. Coefficient of bandwidth for new category OM4 must be at least 4700 MHz * km.

OM4 multimode fiber category will allow data centers to move into the future without any problems on the use of 40 and 100 Gigabit Ethernet technology, allowing you to work 10Gbit/sec system at distances up to 550 m and reduce the requirements for insertion loss in the channel. A requirement for the insertion loss at the limit for the standard distance is very difficult to reach when using an optical cable with multimode fiber category OM3 and OM2 particular category.

The Design Of Optical Cable In The Data Center

The choice of design of optical cable depends on the number of optical fibers, method of installation, place gaskets and other technical requirements and conditions.
In order to reduce the diameter of the optical cable, it should not be used with fiber optic cables having a secondary buffer coating (external diameter of the fiber in the secondary buffer coating 900 micron).

If the number of fiber optic cable is more than 24 ex-tape, it is better to use fiber-optic cables, as they have a smaller diameter with the same number of optical fibers. This reduces the size of the cable channels or use them more effectively.

It is recommended to use the distribution cables in the data center, rather than the usual cables, which are used for the manufacture of optical cables and jumpers. The concept of “distribution cable,” defined by the standard ICEA S-83-596 and related to the level of mechanical strength of optical cable. Electrical cables can be routed on the tray and ladders.

If the optical cable is laid on top of the telecommunications closet 42U, it is usually the additional protection of the optical, however, if the optical cable will be run in places where access will take the technical staff (for example, under the raised floor), then you may need an optical cable with extra protection. The design of the optical cable may include additional reinforcing the strength of materials – Kevlar thread, fiberglass rods, etc.

If the data center must meet the requirements of the fourth level of reliability (Tier 4), you may need to use an armored optical fiber cable to ensure a certain level of protection for cabling infrastructure.