How we “transferred” the entire infrastructure to the “cloud” in 5 days

A customer recently came to us and said: we need to transfer the work of the entire company to the "cloud" . Preferably yesterday. In general, like all emergency projects, this is a dream task, because there is a lot of work, but no time.

In fact, it was necessary to design the IT architecture from scratch and transfer employee data there. On the customer side, there is the executive director and his partner, about 100 users on 4 sites, on our part, there are four small working groups. One was engaged in terminal access, the second - was responsible for mail. The third was engaged in infrastructure (VPN, network connectivity), and the fourth was setting up backup.

We managed it in five calendar days - on Wednesday we conducted a full audit of the Tavros IT infrastructure, developed the basic architectural solutions and selected the technical means for their implementation. Over the next 2 days, CROC specialists actively worked on the deployment of the infrastructure that meets the requirements of the customer. On Saturday and Sunday, the little things were finished and tested. On Monday, users of the company went to work and did not notice anything, but most of the IT objects were already in our “cloud” .

History


It was necessary to have time:
  • Audit your current IT infrastructure
  • Develop basic architectural solutions
  • Choose hardware for organizing IT systems in the cloud
  • Organize a secure connection between the cloud and 4 customer sites
  • Configure a domain controller in the cloud
  • Install the Microsoft Exchange-based mail system
  • Transfer to the "cloud" several "test" user accounts.
  • Configure Terminal Servers
  • Put mail and terminal servers to support CROC
  • Configure data backup of mail and terminal servers using EMC Avamar firmware


As part of the project, we proposed a number of organizational solutions to reduce the cost of the project. The customer sought to minimize one-time costs during the move and naturally to reduce periodic payments subsequently.

At first there was an audit. Infrastructure, it must be said, was uncomplicated, but still it was necessary to assess the scale of the upcoming reform. The infrastructure consisted of corporate mail, workflow, 1C and terminal farm. The audit and infrastructure plan were handled in a day.

Prior to this, the customer actively used terminal access based on Citrix XenApp ( like this ), decided not to invent a bicycle and took this product as a basis. As the mail system used Microsoft Exchange. For the organization of mail they took it as a basis. This then allowed the customer to migrate user data to the mail server in the cloud without much difficulty. How to finish - screwed a shared backup from the cloud .

We organized the network access like this:



For employees, the composition of the application software that they used before moving to the "cloud" has not changed. Only the location of the servers and the principle of their payment has changed. Accordingly, accountants noted the time of payment, IT specialists noticed the location of the servers - by the way, they moved mailboxes themselves (during the installation process, CROC employees migrated only 5 user accounts and user mailboxes from the customer’s site, prepared instructions for migration of the rest). Ordinary ordinary user did not notice anything. Everything remained as comfortable as before. Perhaps the IT workers felt some relief, since CROC was now responsible for the platforms.

How often do we do that


For 5 days - quite rarely. But by themselves, the tasks of moving IT infrastructure to the “cloud” appear quite often. If two years ago such cases were rare, now even large companies with a thousand or more people almost completely move to public “clouds”. True, it happens that they translate everything gradually, the IT system after the IT system, step by step. They look closely at the cloud service and, making sure of its maturity, translate everything else.

The reasons for all are different. The most common is the convenience of maintaining infrastructure. Indeed, the area of ​​responsibility is shifted from supporting the data center and application software only to supporting the application level (like the same 1C, mail, etc.), because all the hardware is on our side.

The second reason is the unreliability of the infrastructure. One of our customers transferred everything because they had a power outage in their business center, sometimes for a couple of days. No one wanted a repetition of history.

Why so fast?


To be honest, at first we were a little worried about the timing. It was at the end of April, on Tuesday, and if we had not met it until Wednesday of next week, the May holidays would delay the completion of the project. But this option was not considered. It was difficult to find engineers to work on site before the May holidays. All customers needed, hands went like hot cakes. But found. Horizontal connections inside our company helped.

As a result, everything turned out to be done even faster than necessary. This was facilitated by the fact that there were plenty of computing resources and licenses in the cloud and it was possible to start using them on the same day. Skilled engineers have been found. And the will to win only contributed to success.

If the project was carried out at the customer site, everything would be fast only in theory. In practice, there are no servers (they need to be ordered, brought, arranged), there are no licenses, there are no engineers - they are on vacation, something else is missing, then the distributor does not have switches in the warehouse ... But here the resources are allocated, in fact, for five minutes.

Another interesting organizational point - the agreement between the two companies is usually quite problematic to agree and sign in one week. Especially in a situation where all the coordinators on both sides were involved in the process of moving to the cloud. Therefore, it was decided to draw up a letter of guarantee, which was quickly agreed and signed. And the agreement was agreed and signed a little later.

Tech support


The customer was offered the option of providing services of the 2nd level of technical support for the mail system and terminal servers. This meant that such routine operations as opening mailboxes, increasing their volume, restoring individual users' letters from backups, came under the control of the help desk of the customer company. CROC engineers are responsible only for the administration of server application software, for the timely increase in machine performance, and for the implementation of updates.

On the one hand, this approach has significantly reduced the cost of external technical support. On the other hand, its own technical support service is closer to ordinary employees of the customer company. For example, it’s easier for an accountant Galina to call her employees and say: “My number ran away from this file to another. And in general, I can’t find another file. Help me".

If the problems are wider or their IT specialists can’t cope with the task, they contact the second line of support, already ours. At the same time, as a whole, the first line of support manages the infrastructure itself, through the self-service portal of the cloud platform and through the management tools of the application software.

Summary


For four months now, the customer has been “located” in our “cloud”. There are almost 100% of its IT infrastructure. The customer did not regret the chosen IT development strategy, and he is constantly evolving.

What is the result? The customer has a fulfilled desire to transfer non-core activities to professionals from a third-party company and optimize IT costs. The customer’s financial director has happiness because of the transformation of capital costs into operating expenses, upon the provision of services. The company completely got rid of capital payments on IT. After all, all computing resources, software, technical support services, data backup services based on EMC Avamar were provided to him according to the monthly payment scheme. Moreover, the customer can always predict an increase or decrease in costs during the consumption of CROC services, since the services are provided within the framework of pre-agreed tariffs. When changing the number of users, the volume of mailboxes, and the capacity of servers, they always know what to expect in the service bill at the end of the month.

For users - nothing has changed. The IT department has less headaches due to infrastructure issues. And we got an excellent experience in operational implementation and the ability to go on holidays with a pleasant sense of, if not fulfilled duty, then the completed project.