How to treat a malfunctioning cloud provider
Difficult situations and behavior of people in a critical situation - a litmus test, which will allow you to understand everything that is necessary for making a decision about further cooperation.
For many years we have been leasing our virtual infrastructure and are well aware of all the concerns and myths that have developed around the topic of reliability of IaaS providers. Let's try to deal with some of them.
/ Photo bobcat rock CC
')
In our Russian community, high-profile stories are constantly being discussed about how a particular western company worked on solving a critical situation. Such stories happen often enough and give us space to draw conclusions and learn from the mistakes of others.
Failures on the side of the IT infrastructure provider can occur due to a variety of circumstances. For example, even with a point error during the firmware update of one of the technological elements. Even such a trifle can lead to huge troubles - the attention of the media to the problem - is a multiple multiplier, which makes it clear to your potential clients that you could not keep the situation under control.
On the other hand, any errors can be corrected by excellent work later. We all know about hundreds of news stories with stories of failures in major corporations. But who remembers what really caused a particular failure? The way the company emerges from this situation is remembered much brighter than many failures.
Cloud providers are often faced with an incorrect understanding of the essence of their work from customers. The irritation of those who have undergone the provider’s error has its own meaning, but in order to adequately assess the situation (including when choosing a provider), sometimes there is not enough basic knowledge.
For example, the simplest separation of IaaS providers into UDP and TCP providers. The first - provide more opportunities for scaling, but pay less attention to reliability. The second - on the contrary. Of course, we speak in very simple words, but in fact there is no knowledge hidden from the average person.
A TCP provider is almost the complete equivalent of using your hardware and enterprise products available on the market. UDP is less reliable, although problems can happen to everyone.
As we said above, there are failures at all. At the level of software updates, hardware problems or managerial errors. There are also more funny situations. For example, problems with cooling and automatic triggering of defense mechanisms.
Mass coverage of the problems of large IaaS-providers blurs a little the real picture, which is that there are failures on the side of ordinary companies that use their IT infrastructure. It is worth noting here that people write and tell about this very, very rarely due to the secrecy of such information.
Thus, it is very difficult to assess the real distribution between providers and their own infrastructure. One can only say that the provider has a specialized expertise to solve even the most difficult emergency situations. A business that is not directly related to this field (for example, a digital agency) cannot always boast such knowledge and the presence of specialized specialists.
When choosing and working with your IT infrastructure or renting virtual capacity from an IaaS provider, you should pay attention to the guarantees that the service provider provides you. It is important to be aware of the possible risks and prepare for the worst. This approach will allow you to understand the probability of occurrence of failures and prepare for their occurrence.
For many years we have been leasing our virtual infrastructure and are well aware of all the concerns and myths that have developed around the topic of reliability of IaaS providers. Let's try to deal with some of them.
/ Photo bobcat rock CC')
In our Russian community, high-profile stories are constantly being discussed about how a particular western company worked on solving a critical situation. Such stories happen often enough and give us space to draw conclusions and learn from the mistakes of others.
As it usually looks
Failures on the side of the IT infrastructure provider can occur due to a variety of circumstances. For example, even with a point error during the firmware update of one of the technological elements. Even such a trifle can lead to huge troubles - the attention of the media to the problem - is a multiple multiplier, which makes it clear to your potential clients that you could not keep the situation under control.
On the other hand, any errors can be corrected by excellent work later. We all know about hundreds of news stories with stories of failures in major corporations. But who remembers what really caused a particular failure? The way the company emerges from this situation is remembered much brighter than many failures.
Cloud providers are often faced with an incorrect understanding of the essence of their work from customers. The irritation of those who have undergone the provider’s error has its own meaning, but in order to adequately assess the situation (including when choosing a provider), sometimes there is not enough basic knowledge.
For example, the simplest separation of IaaS providers into UDP and TCP providers. The first - provide more opportunities for scaling, but pay less attention to reliability. The second - on the contrary. Of course, we speak in very simple words, but in fact there is no knowledge hidden from the average person.
A TCP provider is almost the complete equivalent of using your hardware and enterprise products available on the market. UDP is less reliable, although problems can happen to everyone.
There are problems
As we said above, there are failures at all. At the level of software updates, hardware problems or managerial errors. There are also more funny situations. For example, problems with cooling and automatic triggering of defense mechanisms.
Mass coverage of the problems of large IaaS-providers blurs a little the real picture, which is that there are failures on the side of ordinary companies that use their IT infrastructure. It is worth noting here that people write and tell about this very, very rarely due to the secrecy of such information.
Thus, it is very difficult to assess the real distribution between providers and their own infrastructure. One can only say that the provider has a specialized expertise to solve even the most difficult emergency situations. A business that is not directly related to this field (for example, a digital agency) cannot always boast such knowledge and the presence of specialized specialists.
Conclusion
When choosing and working with your IT infrastructure or renting virtual capacity from an IaaS provider, you should pay attention to the guarantees that the service provider provides you. It is important to be aware of the possible risks and prepare for the worst. This approach will allow you to understand the probability of occurrence of failures and prepare for their occurrence.
Source: https://habr.com/ru/post/261403/
All Articles