The modern enterprise IT architecture, be it cloud-based or local, can be presented as a combination of several different aspects.
The first level describes the main business processes of the enterprise in a general sense and defines the key points of the corporate IT strategy. The second level comprises definitions of specific tasks and use cases for business processes in the IT infrastructure. This level also includes the application stack designed to solve the tasks in the most optimal way and algorithms of interaction between applications. The third level systemizes data management processes and defines their lifecycle parameters.
The basic, or fourth, level describes the technology architecture: a set of clouds, servers, storage systems, and telecommunication hardware that provide the computing capacities necessary for business process automation.
The IT industry has a set of best practices that allow you to build and design the corporate IT architecture smartly in each such aspect. These practices are widely used by enterprise IT architects; they describe technologies, applications, and data processing approaches, and contain a formal description of the business processes themselves, which are automated by the basic bricks of the pyramid.
Trust: so necessary yet so fragile
The specifics of IT in our country is that many companies need to optimize their infrastructure further to minimize operating costs and develop an adequate IT strategy. Still, they often suffer from insufficient in-house expertise in comprehensive project implementation.
Russian businesses are have not yet started to make substantial investments in the development of large-scale IT architectures—they prefer partner services. Heads of corporate IT departments are reluctant to entrust the development of comprehensive architecture to integrators—companies try to determine the required system components and procedures on their own. Thus, it is crucial to focus on increasing the trust of enterprise customers to integrators—the customer must be 100% sure that the product (IT architecture, in this case) is worth the investment and the partner will offer reliable support for their business.
Western countries are a bit more advanced in terms of IT, and they managed to overcome this barrier. Fortunately, the trust of the Russian business community to system integrators is also growing.
For example, system integrators successfully solve specific tasks within the scope of expensive and responsible ERP system implementation projects—from delivery of required hardware to building the target computing infrastructure and turnkey development of the corporate technology architecture. Some of these projects demand to reinvent the enterprise IT architecture at almost all levels. Successful projects improve the market reputation of integrators, in turn, allows to maintain the necessary degree of trust, improve it, and build a long-term mutually advantageous partnership.
Secrets of a well-designed IT architecture
First of all, it is important to note that enterprise IT architecture does not exist for its own sake: it always serves business purposes and has a direct impact on corporate efficiency. The essential feature of a good IT architecture is that all its companies must be densely and optimally integrated without any compromises in terms of flexibility and ability to adapt to the ever-changing business situation and business development.
Most companies prefer medium-term over long-term planning, which reduces the volume of permitted capital investments and limits investments in IT infrastructure, requiring convincing justifications of its medium-term feasibility. Corporate IT budgets are becoming smaller, so CIOs look for ways to optimize costs even more rigorously. Thus, IT architecture must offer processes and technical tools that would solve all tasks with minimal short-term costs.
Like a living being, IT infrastructure is constantly changing. Technology keeps evolving, equipment becomes obsolete, internal business processes are undergoing various metamorphoses, new tasks emerge, and old ones are phased out. An adequately designed architecture enables you to adapt to business dynamics, update, scale, or decommission specific infrastructure components relatively quickly and without reinventing the infrastructure concept.
The relations between business and IT are not always that harmonious in our country. In Europe and the U.S., the market is stable, and large enterprises are well aware of the importance of IT architecture optimization. Having fallen into all possible traps, they eventually came to understand the need for a consistent and qualified approach to IT architecture design based on the market expertise.
What about Russia?
Successful startups try to meet the current trends and strive to prepare an efficient IT infrastructure development strategy from the very beginning. At the same time, the core of corporate IT infrastructure—a classic data center—often faces many significant problems. These problems are particularly common in large companies with a rich history in the market and mature and well-established information ecosystems.
The causes of problems can be traced back to the past. Then the IT infrastructure was expanded very chaotically in an attempt to respond quickly to business challenges, but these decisions were not aligned with a long-term IT strategy. The reason for this is that in the time of quick change, the IT department had no proven methods, no developed competencies, and no relevant IT services in the market.
The first and foremost problem of traditional data centers is the diversity of the hardware fleet within data centers. In large companies with a long history, the range of applied technologies can be really great. Due to the forced coexistence of utterly different hardware platforms (e.g., SPARC and x86), products from different vendors (HP, IBM, etc.), and various network stacks (e.g., Ethernet and FibreChannel) in one data center, the manageability, and scalability of such a "Behemoth" is far from perfect, and the operational expenses tend to be exorbitant. This is especially the case if the infrastructure has business applications tightly linked to legacy technologies that must also be supported.
The second problem—inefficient use of existing hardware—is directly related to the first one. Existing hardware resources are underutilized, obsolete equipment is not decommissioned and not replaced with cheaper and more efficient analogs. With the increasing shortage of available data center power capacities and non-linear increase in their cost over time, the situation may become critical soon. Businesses will have to spend a significant part of IT budgets to solve it without getting any benefits in return.
The third problem is the cost of qualified personnel. According to Gartner, the cost of skilled staff may reach 40% of the total data center costs. The higher the share of such fees is, the more complex and sophisticated the existing data center infrastructure becomes, which is not always justified by business objectives.
Many of our customers came to Softline with such problems. Within a reasonable time, the experts from our company offered them solutions and services that efficiently solve the current tasks using a trust-based approach tailored to each specific customer and supported by Softline's reputation in the IT service market.
The importance of audit
A tailored approach to solving customers’ IT infrastructure problems is impossible without a comprehensive audit—an integral part of any sufficiently large IT project. The audit allows you to accurately identify the source of issues and unjustified costs and develop the only correct solution to address the encountered difficulties.
According to our audit standards, we engage a team of experts capable of evaluating all aspects of the existing IT architecture. Generally speaking, the audit procedure is adjusted based on the scope and complexity of the tasks set by our customers.
Building the technological part of the infrastructure
Nowadays, there are three main approaches to data center design. The first, classic one, comprises preparing a dedicated room for hardware and appliances that will host corporate IT infrastructure and business automation tools. At the same time, the customer adopts a set of organizational measures and recruits staff to support the service level sufficient for the corporate information services. Many companies adhere to these conventional methods.
New market realities give rise to a second, radical approach to data center organization: refusing from an in-house data center. Not many companies, especially in our country, dare to abandon local hosting of business-critical IT infrastructure completely. However, the development of cloud technologies motivates more and more customers to transfer business applications to the cloud, delegating the responsibility and the burden of related technical problems from their staff to outsourcers. At the same time, customers' expenses on cloud infrastructure and applications—IaaS and SaaS—are clearly expressed in the form of subscription fees and directly correlated with the volume of consumed resources.
In fact, passing the baton for data center construction and maintenance to cloud companies, unfortunately, does not automatically solve the fundamental architectural problems mentioned above. The problems simply change their owners, leaving cozy corporate server rooms and moving to spacious data centers of service providers.
As a result, cloud service providers fighting for life and death to ensure compliance with the SLA requirements, as well as companies wishing to maintain and efficiently develop their in-house data centers, conceived and supported the third concept—modular convergent data center. In such a data center, the entire infrastructure is represented as relatively autonomous blocks interacting with each other according to a predefined procedure using a common stack of standard protocols. The modular paradigm combines the advantages of the classic and cloud approaches and is almost free of their drawbacks.
The modular data center infrastructure is still located at the customer's premises; however, it is designed to minimize the time and financial costs associated with commissioning and integration with the existing environment. Such infrastructure becomes ready for deployment in the shortest possible time, almost from the moment of module installation, connection to power supply and data transmission network. Just like with cloud data centers, support and maintenance are outsourced to the partner. The customers do not need to worry about their IT infrastructure—they just declare the necessary amount of resources and functionality, and the partner provides them.
Supported by many major vendors, this approach is steadily gaining popularity today.
A key advantage of the modular data center infrastructure, in addition to deployment speed, is their flexible scalability. When a new business process is automated or a new product is launched, additional capacities may be required. In this case, an out-of-the-box module is purchased in a pre-configured rack. If the customer needs to double the capacities, they install another module next to the previous one, achieving near-to-linear scalability.
The third advantage of modular data centers is that the infrastructure of this type is ideal for implementing a Software-Defined Datacenter (SDDC) and can become its foundation.
The concept of SDDC deserves a special mention. As the definition says, the architecture of such data centers is mostly defined by software installed in it instead of hardware.
The industry has already appreciated the benefits of server virtualization: it allows consolidating computing resources of individual physical servers into common pools and using software to allocate resources between applications precisely in the required amount. At the same time, deployment of new virtual servers requires just a few mouse clicks.
A logical consequence of the server virtualization success is applying the approach used in this technology to other IT infrastructure components—storage systems and data networks.
In the first case, we build an SDS—a software-defined storage, where storage resources are combined into similar pools with defined service levels. SDS is abstracted from specific storage equipment. In the second case, a software-defined network (SDN) is created. The topology and architecture of SDN are also set programmatically and fully unified over a flat physical structure, and the configuration of such a network is much less dependent on the functionality of specific equipment.
SDNs cannot be imagined without the use of convergent networks. The need for two independent data transfer infrastructures—one for IP traffic and one for storage traffic (classic SANs based on FibreChannel technology)—negates all the advantages of an SDN. A converged network allows you to use efficiently the same infrastructure for different types of data and for the tasks that require the highest quality and impose stringent requirements on the delay levels and transfer rates.
Altogether, these technologies shape the concept of a software-defined data center, SDDC, which, as well as modular infrastructure, relies on unification and standardization of interactions between its individual components creating a synergetic effect for these technologies. Undoubtedly, the concepts of SDDC and modular data center infrastructure are the most important and most promising to maximize the efficiency of the foundation of the modern efficient enterprise IT architecture—its technology architecture.
Softline offers the broadest range of services for the development of enterprise IT architecture at all levels. We are a cloud service provider with an excellent reputation, and we always know the optimal approach to solving the customer’s tasks.