Common mistakes in building a small business IT infrastructure

    Which do everything, and because of which then you have to redo everything

    The vast majority of small business IT infrastructures that I have encountered have one major drawback - they are not scalable. No, no, the problem is not that you cannot connect the 51st computer to the 50 computers currently available. The problem is that the architecture laid down initially creates disproportionately more problems and costs with the growth of the business than it brings benefits.

    For a young business, the planning horizon is usually limited to a year, or even a couple of months, and building an IT infrastructure is far from the most pressing problem. If there is a quick and cheap solution, albeit with side effects in the future, the business will choose it than the more expensive, but correct from the point of view of future prospects. And there is nothing wrong with that. It’s bad when this practice continues after the business has matured and is confidently looking to the future.

    In this article, I will highlight several points that should be paid attention to in time, and which can help companies avoid the losses associated with a complete remodeling of the IT infrastructure as the business grows:

    1. "Zoo" equipment and unsystematic procurement of computer equipment

    In a small business, equipment is usually bought and changed as needed and exactly in the quantities and qualities in which it is needed right here and now. The models of computer equipment on the market sometimes change several times a year, being replaced by newer or cheaper ones. Under such conditions, the appearance of a “zoo” of computer equipment is inevitable in small companies when it is impossible to find two identical servers or even two identical computers on the network, but this is not a special problem, while the arising service tasks are exclusively “one-time”.

    As the business grows, various equipment models, in addition to a linear increase in service tasks, create additional difficulties in planning and implementing changes, creating reserves and even modernizing the computer park, requiring accurate accounting and analysis of each element. All this increases costs and negatively affects the quality of work. It may take years to standardize the equipment used, while previously acquired equipment will exhaust its resource.

    In order not to confront the choice of throwing out old equipment or overpaying for its maintenance, a growing business needs to be guided in the purchase of computer equipment primarily by the issues of total cost of ownership, and not by immediate needs. Conducting annual or at least quarterly purchases of computer equipment will significantly reduce the "variety", thereby simplifying the tasks of regular maintenance, as well as planning, and making changes to the network.

    2. Cable network made by the number of employees.

    For some reason, almost always when renting a new office, small business executives try to save on the installation of low-current (and even electric) wiring amounts that often do not exceed 10% of the monthly rental payment. The usual argument: in our office (30 square meters), only two people will be seated. It is not possible to convince business management of anything, because they know for sure that the company "if it grows, then by 1-2 people."

    Naturally, after a couple of years, all rooms are filled to capacity. There is no need to rent a new office, but the old one is overgrown with a pile of wires, switches and extension cords lying on the floor and trying to break at the most inopportune moment. Remodeling the cable network in a densely packed office, and even when the work can be done only on weekends or even at night, is an order of magnitude (yes, yes, it is an order of magnitude) more expensive than the initial installation cost.

    Admit to yourself that you are renting a new office four times the previous one, not in order to overpay for empty space. Do power and low-current wiring, based on the maximum possible stuffing, and do it right away.

    3. Lack of data storage structure on the file server

    A file server in a small business is often a “network drive” with 200 folders in the root and thousands of documents, the purpose of which few people already understand. A briefing on working with a file server for a new employee usually looks like this: go to the server in the “Ivanov” folder and find such and such a file in it, while Ivanov himself was dismissed from the company 6 years ago. And yet, it works great and solves its functions as long as the company remains small.

    When the company grows, miracles begin to happen on such a file server - someone selects some folders and transfers them somewhere, someone overwrites a document that another employee has been piling for a week, or even delete something “unnecessary” . Regular backups can fix the situation, but do not increase trust in the resource. It is not possible to distribute access rights within the existing file server structure, because everyone needs all the folders. The only way out is to design and launch a new file server, transfer the necessary data to it from the old one, and remove the old one into the archive.

    If you create a centralized storage of documents in your company, then please work out its structure and allocation of access rights immediately so that employees can enter and change only their data (or the data of their department). The structure of the file server and the distribution of access rights is not a matter of trust or distrust of employees, but primarily a matter of structuring data storage, when each employee is sure that his work will not disappear due to someone else’s error and knows who is responsible for the mess in that or another folder.

    4. Unlimited mailboxes

    It is generally accepted that mailbox size limits are a relic of the past when drives were expensive and the Internet was slow. In small companies, often there are no restrictions on the size of the stored correspondence, and users do not bother with the questions of regular cleaning of the mailbox from unnecessary letters, and the necessary letters can be safely stored only in mail, because "They can always be found there if that." Even if you have your own mail server, mailboxes of 30 GB in size do not present any problem if the company employs only 5 people.

    Moreover, the picture changes dramatically when at least 100-200 users begin to work with mail simultaneously. The size of the mail database becomes so large that creating (and storing) its regular backups requires not only huge disk space, but also productive (and expensive) backup systems. And all this is necessary only in order to store 90% of unnecessary information to anyone. Forcing users to clean their mailboxes on their own is unrealistic - not a single person in their right mind will be able to view their correspondence over the past 5-6 years (an average of 50-100 thousand letters). It’s also impossible to delete it, because there really are important documents that were left in case "if something" and which can only be remembered when this "what" happens. The only way out is to unload the entire archive,

    Restrictions on the size of the mailbox is not atavism, it is an incentive for employees to use mail only for current correspondence, timely clearing mailboxes of garbage and placing important information to where it really belongs: on a file server (in the folders of projects, contractors, contracts), into an enterprise management system or corporate portal. The restriction policy allows you to easily migrate existing mailboxes with any upgrades and, most importantly, buy equipment and licenses that will serve a really useful amount of data, rather than huge archives of seals and Friday videos.

    5. Rights of local administrators for employees

    The ability of employees to independently install “everything they need to work” at best turns a computer into a breeding ground for conflicting and counterfeit software, and at worst a virus distribution center. The presence of the very ability for the user to influence the operation of the operating system on the computer according to statistics does not add stability and speed to its work, and creates guaranteed problems. As the business grows, such excess rights will increasingly be reflected in the cost of support. To eliminate the consequences of such a policy and get rid of accumulated errors, a complete reinstallation of operating systems at workplaces may be required.

    The division of access rights was not invented in order to infringe on the feelings of those who were born in the soul as a system administrator. The ability to restrict access to system settings eliminates the influence of incompetent actions or simply user errors on the quality of the operating system. The practice of restricting access rights allows not only to save on the cost of support, but also significantly increase the stability of computer systems.

    6. Development and construction of IT systems without documentation

    To my surprise, a small business has absolutely no problems finding a brilliant programmer who will create a website, corporate portal, management system or even just write a couple of simple reports. The documentation for the code and development does not bother anyone - after all, it’s clear that everything has been done qualitatively, works like a clock and is inexpensive. Over time, the first programmer disappears somewhere and the second no less ingenious appears, who also quickly, clearly and inexpensively does everything that is asked of him and, of course, without any documentation - to hell with these formalities.

    But the larger the business, the more difficult it is for him to find brilliant programmers. The company grows and changes, business processes change in it, and here it turns out that these mediocrity for a small change in the function that you once did entirely in 2 hours and a thousand rubles, require a month of work and the amount is two zeros more. Moreover, your system, which worked perfectly on an old computer on a network of 20 users, suddenly began to slow down at 100, although you have long ago transferred it to incomparably faster servers. Sysadmins say that the servers are not loaded, and programmers mumble something uncertainly about architecture, queries, and some kind of table locks. The previously developed system is being transformed from a tool to increase business efficiency into a money vacuum cleaner, without which the business can no longer function.

    Documentation for systems in general and for development in particular is not an additional, but not a necessary attribute of a quality work. For a specialist to compile documentation is not particularly difficult. The availability of documentation makes it not only easier to transfer cases to new developers, but also to attract external experts for an independent assessment of the quality of the completed project.

    7. Belief in brands and purchase of equipment only for current tasks

    Often, after another failure, a small business buys “well, very high-quality equipment”, expecting that now nothing will break. Because the business is still small, and the equipment is expensive, models are bought more modestly, "containing all the necessary functionality." So the company has dual-HP servers with one processor, two disks and 4 gigabytes of RAM, and this is often enough to solve problems.

    As the company grows, it turns out that even “well, very high-quality equipment” sometimes breaks down and it is necessary to provide at least a replacement, and ideally hot standby. In addition, previously purchased equipment has insufficient power, and the cost of increasing it (adding RAM, installing additional disks and a processor) is comparable to buying a new, but not so much branded counterpart. At the same time, IT budgets are still limited and the question arises of how to properly dispose of them. The choice is not the easiest - hold on to a brand or build a network on more affordable devices, but with full redundancy, throwing out previously purchased devices.

    When deciding to purchase this or that equipment, one should proceed not from the functions that it will serve now, but from the functions that it will serve closer to decommissioning. Brand equipment is, of course, high-quality equipment, but you should buy it only when the quality of work of the IT infrastructure cannot be improved any other way. If you have a choice between buying a single, but branded server, and two simpler brands, then you should choose the latter - there will be more benefit.

    8. Cartridge refilling

    Small companies, when purchasing printers, are often guided by issues of aesthetic appearance, print speed and price of the printer itself, not paying attention to the cost of the print and the planned print volumes - after all, refilling cartridges solves any problems.

    As the company grows, the use of original supplies does not become cheaper, while print volumes increase so much that cartridge refueling specialists and printer repair engineers appear in the office almost every day. The concept of “dirty print” becomes the corporate identity of the company. Employees try not to come to the office in a white shirt, fearing to get it dirty when printing documents. The need to change previously purchased printers becomes apparent.

    Even if you are a super-small company, when choosing a printer, be guided by the fact that the cartridges in it should be changed no more than once a month, and the cost of the print is affordable for the purchase of original supplies. The purchase of such a printer will often pay off to you from the very first cartridge and save you from the need to throw out all previously purchased devices for the purpose of saving.

    9. Using Small Business Solutions

    The market has a huge number of solutions designed specifically for small businesses - these are software licenses, cloud services, analog telephone exchanges, these are scanners and printers that do not support networking, these are unmanaged switches and small network storages. All of them are united by one thing - their functionality and capabilities are strictly limited. And worst of all, some of these restrictions are often impossible to remove for any money. For this reason, when a business grows and begins to rest against them, he faces a choice: either put up and live with them further, or redo everything completely.

    Small business solutions are great for companies that will always be small. But if the business is growing, then, when deciding on the purchase of various systems, it is worth focusing not only on the functionality of the solution now, but also on the capabilities of this solution later. This will allow you not to spend money on the purchase of those systems, which then will have to be written off, but, most importantly, will save money on their implementation and subsequent refusal to use them.


    Software and hardware have a useful life in the range of 3 to 5 years, and business applications are designed for decades of work. Retirement of IT systems before their useful lives means lost money. Small business, making decisions about various investments in IT, is often oriented to one or two years of their use, or even several months. Such a planning horizon inevitably leads to erroneous actions and, as a result, lost money in the future. The cost of correcting the mistakes made at the start can significantly exceed the cost of properly organizing the IT infrastructure at the very beginning.

    Good luck!

    Ivan Kormachev
    IT Department Company

    Also popular now: