What should be the ideal IT infrastructure model for an abstract enterprise?

Question

Information Security

Good day!
I'm thinking about the next question - what do you think should be the ideal IT infrastructure for small businesses. To begin with, let's assume several input values: to begin with, the number of employees is from 10 to 150, of which 90% are somehow connected with the need to use the enterprise's information infrastructure. That is, suppose that 1C should work at the enterprise, backup of user data (where would it be without it!), Corporate mail, a number of highly specialized financial and accounting programs. Well, plus office applications (for everyone) and resource-intensive applications for a small number of workers. Let's assume active work on work with clients and performance of any works - whether it be development of applications or drawing. How do you see the ideal infrastructure for such an abstract enterprise? How would such a system work in general? That is, if you plan to backup - cloud or local server (cluster of servers) and why. How users will work with it, how they will be served, how they will expand - both in terms of resources and increasing the volume of processed information, and territorially. Fundamentally, I understand that specific examples are needed to develop such plans, but in general is it possible?...

Answer 1

Taxes, if we consider the practical side, then:
1. You must have a well-built network, dividing everything into separate segments and dividing it into an access level and a core. These can be both simple gigabit switches, and something more serious (separately l2 switches, the core is on l3) solution. All Wi-Fi networks are separate, telephony is separate. Multiple providers with auto-switching. Between offices (if there are several) tunnels with IPSec and OSPF, because you always have two channels in each remote office. Inside local networks, you must have a protocol of the stp family that monitors loops and convergence, and nowhere are there simple "soap" boxes a la des1004, or what kind of dlink is being released now? All switching is done in advance and all available ports are provided with a link in advance and the entire network, most importantly, is ready for such a turn.
2. Storage. You must separate data storage from computational processing. At least at the level of abstraction (although ideally it is physically different hardware). Storage is designed for the current load plus a reserve at the time of capacity increase (roughly speaking, while additional disks are running - you did not fall under load if it suddenly stopped). You have access through separate ip networks or through fc san (IMHO, at the moment it is very close). Storage can take snapshots and thinly allocate resources. She also knows how to replicate a \ synchronously (and you have where). The vault has +1 power, +1 disk of each type, and +1 controller head. And also +1 link to calculators.
3. Calculation. You have separate identical pieces of hardware on which, thanks to virtualization, you have deployed a farm of computers for the main IC, and also turn all your variety of mail servers and domain controllers. (Someone will say that the CD is ideally twisted physically, perhaps). Servers are designed with a margin of at least 1 piece, so that you can safely migrate virtual machines and lower it into maintime.
About the software, too, I can throw if necessary.

Answer 2

Try browsing ITIL

Answer 3

The ideal business model will work perfectly. with which I congratulate you. but one thing bothers me a little. why do not write questions on automotive forums "what parts should an ideal car consist of?"

Answer 4

I am also interested in what kind of budget is initially allocated to this ideal enterprise. This is the biggest plug will be in this scheme.

Answer 5

Alexander, it is impossible to give any worthwhile answer to your question, one can only dream. The reason is in the considered abstract enterprise and insufficiency of initial data. A good example was given to you with a car. In reality, when planning an
ideal IT infrastructure, and even with a reserve for expansion, you need to have initial data, and the more
the better
: what parts of the facility can be accessed and what type of work is allowed - distribution of personnel at the facilities, their functionality and individual requirements - data protection requirements
- whether it is possible to receive communication services, which ones, with what characteristics and under what conditions
- who and what qualification will serve IT systems
- how the enterprise will expand in the future
- the possibility of ordering and delivering one or another equipment and software
further each of these points are being worked out, and for as long as a project containing:
- cable routes
- equipment location (active and passive)
- installation rules and conditions
- connection / switching schemes
- equipment and software operation algorithms
- all kinds of plans (ip, backup and etc.)
- initial initialization plan and a list of necessary settings
- the composition of spare parts warehouses
- maintenance regulations and instructions for all kinds of situations
- testing plan
If the initial data is incomplete, then when developing some sections of the project, it is only possible to rely on intuition and experience, hoping that a specific choice of solution will be justified during operation. For example, let's put 10 ports on the switch, not knowing how many network devices there will be, and after launch it turns out that there are
not enough of them. Therefore, it is always worth laying a reserve, but which one will be justified not to guess.
That. there is no specific ideal system, but it is realistic to design a suitable one only for certain tasks, taking into account the given conditions. How close to the ideal your IT infrastructure will directly depend on the detail and quality of the initial data, your resources and the deadlines set.
Also, the original data changes over time, and even errors are made in them.
ps
the ideal infrastructure from the point of view of a business and a designer is completely different, except for those cases when the budget is not limited, but, as it was written above, even without restrictions it is impossible to guess what and how will change in the future
Note that I did not even consider the choice of technologies, equipment and programs, because it depends on the tasks to be solved, conditions, experience and preferences of the designer. After all, you can design similar systems on different equipment.
ps2 what a question, such an answer :)