IaaS for non-IT companies: 10 answers to director's questions

If your company's business is not related to information technology, outsourcing infrastructure to an external provider will help reduce the amount of routine operations and untie the hands of the IT department, giving it opportunities for development. The work of specialists will not be less, it will just be different: first of all, time will be freed up for the adaptation of business applications. However, management is conservative and hardly agrees to an expensive project if it does not benefit the core business.

To convince top managers of the need to move to IaaS, your arguments must lie in the economic plane. Today we will try to make out the main issues and objections of the leadership and show how to work with them.

')

Outline the disposition

IT companies in the company are often perceived as Martians, who ask for a lot of money for incomprehensible business toys. Worse, if management thinks that it understands IT, and tries to take the initiative, focusing on fashion trends. This leads to the forced introduction of supposedly innovations, which for the company are not needed at best, and dangerous at worst.

In recent years, even the most mossy leaders have recognized the need for the development of IT infrastructure, and IT specialists have developed a habit of arguing their initiatives in terms of the well-being of the company's main business. They start thinking about the transition to IaaS when their own resources are no longer enough to solve current problems or the existing IT infrastructure requires expensive upgrades.

Why do we need it, if everything is already working?

Even the most advanced managers are not well-versed in the internal IT-kitchen, and infrastructure capabilities are judged by the performance of the services with which they have to directly interact.

However, large companies often develop or at least adapt business applications to their own needs. There are often situations when developers, for the next idea to improve user services, request unplanned computing resources, which cannot be identified right now - you have to spend time and money on ordering equipment and commissioning. As a result, the time it takes to bring a product to the market grows, and the IT department is guilty.

To increase the speed of innovation within the company, it is more convenient to implement the infrastructure in the cloud and delegate its service to the service provider. If you suddenly needed an increase in computing capacity, you do not have to purchase equipment, and the contractor will be responsible for meeting the deadlines for allocating physical resources.

How will this affect capital costs?

If the time has not gone into a series or the need for additional resources has disappeared after the completion of a certain project, the equipment purchased and the infrastructure built on its site cannot be lost. In the case of IaaS, you pay only for the actually used resources of the service provider, which is quite simple to give up - this is a strong argument that the authorities like when they do not need to increase capital costs.

CAPEX differs from the work on the service model by the need to keep records, put equipment on the balance sheet, increase the taxable base, activate write-off upon expiration of the service life, dispose of equipment, etc., etc. This operating headache hinders the development of the company and no one in fact need not.

Isn't it easier to deploy the infrastructure in your own data center?

Tops of the industry are able to invest in building and maintaining compliant Tier III standards (this is the most popular level) of a data center with a reliable engineering infrastructure, but even their investments have paid off over the years. In medium and small companies, IT departments are focused on business problems, which prevents them from properly planning data center resources and leads to the creation of redundant solutions.

Large service providers are oriented towards mass demand, their data centers are certified by Tier III, and there is an SLA that regulates the level of service provision. If a company makes its data center, all costs are borne by itself - and when using the service provider’s capacity, the costs can be shared with the rest of the customers, and the operating costs will be significantly reduced. There are, of course, nuances associated with security policy, but they are also solvable.

Will switching to IaaS create security problems?

A modern data center is a secure facility with very serious access control. If necessary, the provider will provide you with your own secure perimeter inside the data center, which can be closed even with a granary lock, install your cameras, access control systems and the like.

Implement such a solution on its own site is much more complicated and much more expensive.

What to do with your own iron?

Usually, the company already has iron on its balance, which cannot simply be written off. In this case, the best approach is to combine the traditional services of the data center and the services of the IaaS service provider. You can use the infrastructure of the provider and place the existing equipment on colocation: the price of such a service is low compared to the cost of an error when creating the wrong operating conditions, significantly reducing the life of the server.

In 2013, Intel experts noticed a large number of electronic component failures in the Asia-Pacific region. Reducing the life of the servers was associated with severe environmental pollution and associated corrosion. Approximately the same effect can be achieved at home, especially when your playground is located in the industrial zone of a big city.

In commercial data centers, temperature and humidity recommended by equipment manufacturers are maintained, and the air is free from dust - servers work longer there.

Another important point: even if you have old x86, you can combine them into a common computing array through virtualization and use with actual applications until the end of their life. The idea of ​​combining traditional services of commercial data centers with IaaS is the most sensible approach right now.

Failure at the provider can create us problems?

Unfortunately, the equipment sometimes fails, but its own IT infrastructure is less resistant to accidents - the Tier III certified data center provider is much more reliable.

Corporate customers rarely have an adequate supply of computing power or a spare parts warehouse in case of breakdowns, although downtime can be very critical for them (in large retail or airline companies, for example).

If you didn’t pay an additional service for the iron manufacturer, even a simple replacement of the failed components under a limited warranty will take days or weeks. The Russian legislation does not oblige the authorized service centers of the vendor to make repairs promptly, and the businesses for which the lengthy downtime is not critical are becoming less and less. Another option - work on the client side (on premise). If you don’t want to touch the equipment at all, you can order outsourcing expertise - the company still doesn’t have it in such volume.

What to do with the premises?

If the company has already made serious investments in building its own infrastructure and this is not about the server room, but about a full-fledged machine hall (we note that this happens quite rarely), you will not have to abandon it when switching to IaaS.

To a so conscious and competent customer, the service provider will offer work on a backup site creation scheme.

Do I need to abandon their own infrastructure?

If you have rested against the ceiling of the possibilities for upgrading the machine room and equipment, there is nowhere to put the equipment, there are models of the hybrid infrastructure - it is scaled by the resources of the IaaS provider, and the basic platform continues to operate at the customer’s facilities.

When the equipment at its own site passes the obsolescence stage, it will be necessary to make a decision and either update its own fleet, or completely migrate to IaaS solutions, and use the vacated areas for other purposes (physical storage of backups, application testing, etc.). The second option is much simpler and cheaper, it does not require capital investments in the restructuring of premises, the modernization of engineering infrastructure, the purchase of IT equipment, etc.

Can we do on our own?

If the business is not connected with IT and the company is not in the first echelon, inevitably there will be problems with hiring qualified specialists in the field of developing and adapting business applications or, say, cloud architects. This is a very spoiled public, accustomed to a certain level of earnings, social packages, bonuses, etc. In addition, you simply will not have tasks corresponding to her ambitions and such specialists will not go to your staff, and it is impossible to build a quality infrastructure without them. You have to either contact the system integrator (which is expensive), or go to the IaaS provider, who has the functions of an integrator “on board” (it is cheaper).

Believe me, getting access to the best IT brains and market competencies is much easier by working with a service provider who already pays salaries to these guys.

What will IaaS give to ordinary users?

With the growth of the computer park, the ratio of the number of qualified IT staff to the number of workplaces decreases and maintaining the motley equipment in working condition becomes more difficult and more expensive.

Storing confidential data on users' computers also creates an additional headache for information security specialists. Small and medium-sized companies solve the problem using thin clients and terminal servers, but big business should take a closer look at desktop virtualization (VDI). In this case, it’s not about multiple user sessions on the same OS, but about different virtual machines on the same physical server.

This approach requires more resources and before it was chosen mainly for security reasons, as well as due to the simplicity of servicing a large number of jobs. However, in recent years, VDI began to be used to save on infrastructure, and providers have DaaS services (Desktop as a service). Earlier we described our experience in this area in detail .