Today we conclude our Business in Search of Software
series with an article devoted to cloud-based solutions. Read previous articles here:
PWA – Browser Under Hood of Application
Nowadays, storing your data in a cloud is considered normal. But if you said that to someone a century ago, that person would probably think you were not quite right in the head. In truth, cloud based solutions got their name because of their remoteness. Essentially, these are servers whose capacities can be accessed from another location. Many of them also provide ready-made solutions that can range from mail services to machine-learning environments.
For start-ups that don’t have the resources to build in-house computing capacities, clouds are the only means of storing and processing data. At the same time, large companies are also increasingly distributing data between their own servers and remote ones. Finally, IT giants, such as Google, Amazon and Microsoft (the same goes for Russian majors like Mail.ru and Yandex) actually lease the capacities of their data centres.
Cloud capacity
One of the biggest advantages of cloud solutions is that they allow businesses to rent as much storage space as they need. For a corporate portal, these capacities will be very different compared to high load systems like 1C – obviously, the rental costs will vary as well. In any case, it’s still cheaper than developing your own infrastructure, especially if you take into account the relatively short equipment life. Large corporations can buy such equipment in bulk and thus optimise their expenses, but for smaller companies it’s much cheaper to use cloud services instead. For instance, developing a neural network may require very high computing power. Instead of buying a server for a hypothetical million dollars and configuring it for your own needs, it’s more reasonable to rent the necessary space for some $20,000–$30,000 and switch the entire development to a cloud.
Besides being an infinitely cheaper, convenient and quicker option, the cloud also eliminates the need to support the server. In other words, if something goes wrong, the load is automatically and seamlessly rebalanced without the customer even noticing the problem.
Naturally, it’s not only complex systems that can benefit from cloud-based solutions. For example, platforms like Yandex 360 or Google Workspace (formerly G Suite) offer extensive toolkits to facilitate remote work, such as mail services, documents, shared presentations and spreadsheets, video conferencing solutions, etc. Such services will be equally useful for large companies and start-ups alike.
Risks and localisation
Another advantage of cloud systems is that they usually use different servers for data processing and backups. This essentially gives you an additional safety net, and you will not lose your data if something happens to the equipment.
Sadly, you can't completely eliminate all risks. We once had a situation where the building that used to house our servers – both main and backup ones – was subject to an illegal takeover, and the power in the entire data centre went out. Luckily, everything wound up fine in the end: though the client did experience some downtime, we managed to successfully retrieve all data. Admittedly, this entire situation is obviously more of an exception than a rule, and we could have avoided the problem entirely had we stored backup servers separately from the main ones at the time.
To err on the side of caution, many companies are also transferring data from foreign cloud servers to domestic ones, allowing them to have more control over their assets. Indeed, localisation of computing powers is a very valid and reasonable step that allows you to safeguard your business from other companies’ decisions which you cannot influence.
The other side of the coin is that transferring information from foreign clouds to domestic ones can be a very complex, lengthy and expensive process. To what extent? Well, that depends entirely on the volume. If the company rents only one server, the process will take hours or days. But if we're talking dozens or hundreds of servers, data migration may drag on for months. In any case, that’s still a necessary safeguard measure, and the sooner you do it, the better.