7 problems of Russian cloud hosting

photo (s) HP Creative Express

This post is an attempt to put together and take a closer look at the main problems of the Russian “cloud” hosting. He does not in any way claim to be an exhaustive reflection of the problems or the ultimate truth. We will be happy if representatives of hosting companies, cloud hosting users, equipment manufacturers, representatives of data centers and in general all who are interested in this topic join the discussion.
')
So, we consider the main problems of Russian cloud hosting:

Vagueness of definitions

It often happens that visually explaining what a “cloud” is and how it differs from the usual virtual hosting, even the employees of the provider cannot. And if they can, they explain it differently from company to company. Someone calls the established American standards for help, someone invents his own definitions. But to the consumer of cloud services, in our case - hosting, by and large, it doesn’t matter anyway, according to which of the standards the cloud is built. He first needs to get a service. In this case, very often, in order to receive it, it is required of him to understand the intricacies of marketing and technical terms. Hence the second problem.

Accessibility to understand

The rapid development of cloud technologies has generated a lot of new terms and concepts: instances, cores, nuclear cores, time slots. This often causes confusion among people accustomed to thinking in categories of gigahertz and gigabytes, as well as paying money for months, not hours. On the one hand, due to a misunderstanding of the mechanism, the main marketing advantage of the clouds, on which the main advertising message is built, is the payment for the amount of resources actually used does not work. On the other hand, instead of imposing terms, providers could more tightly tie in cloud technology and the usual parameters for evaluating the power of virtual machines. A different approach to the description of the actual capacity received by the client often leads to the following problem.

Price range

Prices for virtual machines (instances, slots) of comparative performance from provider to provider differ significantly. The main pricing factors here are the virtualization technologies used, the hardware on which the machines are running, the data storage systems, and, of course, the positioning of the provider itself. The situation with virtualization technologies is extremely ambiguous - while “open” systems do not require any costs from the provider, proprietary systems imply licensing of each running machine. Because of this, there are frequent situations when the provider has the cost price of one virtual machine higher than the “selling” price of a similar machine from another provider. This situation has been around for quite some time, and the manufacturers of virtualization tools, in general, have embarked on liberalizing their own licensing systems, but the changes adopted are not enough. The equipment also plays its role in the final price - some providers build their “clouds” on a dozen simple single-unit servers and model switches, others buy titanic terabit switches and complex storage systems. There is nothing bad about it, the main thing is that service consumers understand what it is they pay for. But there is another risk factor - expected

The arrival of the "giants"

The giants of the scale of the Big Three mobile operators turned their attention to the cloud computing and services market. And foreign players are carefully studying the situation, promising good prospects. Of course, the main focus here is not on cloud hosting, but on mass business services, but the direction of movement is obvious. The arrival of new large and very large players is expected to reshape the market and will require new prioritization. At the same time, the effect of scale works better for large companies - the cost price of the service is, as a rule, lower, and it will be difficult to compete with them. Especially because

Consumers are oriented to the west

Local consumers of cloud hosting services are increasingly paying attention to the services of large providers providing capacity in the US and Europe. Especially if the hosted project focuses on Western audiences. There is nothing bad in the phenomenon as such; in theory, it is most effective to place the service closer to potential consumers and rent three sites at once for Russia, Europe and the USA. But after all, for decades the emerging stereotype “foreign means good” continues to work even after the death of the country in which it took shape. Perhaps this is partly the reason that many potential customers of Russian service providers providing quality services (no matter what they say) are trying to transfer all their tasks to Western “clouds”. At the same time, neither increased delays, nor at least a similar level of reliability, and, often, even less, or problems with the accounting of payments are taken into account. In addition, serious projects, as a rule, require serious customization of hosting parameters, to provide which mass service is unable because of its mass character, which is also often not taken into account. “Success stories”, of course, are also quite a few, otherwise no one would use the services of Western providers. This problem, in our opinion, has much in common with the problem of insufficient consumer awareness, who simply does not know that a service of similar, if not higher quality, can be obtained in Russia at a comparable price. In this case, the level and speed of technical support in most cases will be higher by default. One of the factors that exacerbate this problem is the fear of the mythical “grooves” of servers, this problem is too broad and deserves a separate detailed post. In the meantime, the next problem -

"Iron" factor

The market is young, only in the last year several active players have appeared, plus traditional hosting providers are implementing the transition to solutions of varying degrees of “cloudiness”. Most often, the “running in” is already being done live, on the “combat” systems - simple is fraught with the loss of users, and so far not numerous. And here, of course, problems begin to appear. Most often, the advertised autoscaling of capacities fails, the main marketing trump card of the clouds - the expected allocation of resources does not occur, or it does not happen, virtual machines "stick" or even go to Valhalla. We must pay tribute, in the States the same problems. As it turns out, for the time being, the autoscaling mechanism is so complicated from a software point of view that it’s not a single solution on the market that provides truly reliable and transparent operation. Disk fails a little less often - for “clouds”, high-performance and distributed storage is required, with hundreds of logical drives spread across the storage area network. Obviously, the more complex the decision, the higher the likelihood that something will go wrong. It is impossible to solve the problem radically, it only remains to gradually increase reliability, catching bugs. And, of course, we must not forget, perhaps, the main problem -

Human factor

This topic is truly inexhaustible, because compared to it, all other problems fade into the background. In any business, it is people who unfortunately create the bulk of the problems, where the first places are occupied by failures, information leaks, loss of profits and unplanned expenses, which is often the same thing. The saddest cases are that everyone has heard of “forgotten” air conditioners in data centers, unexpected repairs in the form of switched off switches, incorrect settings of storage systems, resulting in the loss of user data. But besides these, without exaggerating the catastrophic failures, there are other problems. For example, technical support. Very often, people "on the phone" do not have the necessary qualifications to quickly solve problems and pass on the request further, wasting time on unnecessary communications. It happens that the line specialists themselves are poorly aware of the specifics of the company's services and cannot properly convey it to the potential consumer, losing it in favor of the competitors. And with marketing, too, not everything is clear - high-quality and expensive services by default are transferred to the rank of fast food, forming the perception of cloud services as something cheap and insignificant. In general, the young Russian cloud hosting industry still has a lot to learn from other types of businesses that are more mature and have already debugged their business processes.

Something like this we imagine the main problems and challenges facing Russian providers of cloud hosting and services. We deliberately did not use trademarks in the post, because in one form or another all of the above questions apply to most providers, and we should not talk about the problems of some, keeping silent about the shortcomings of others.

We believe that awareness of the problem is half the solution, and the free exchange of information on how to deal with these or other negative factors in the work will help the establishment and improvement of the market. What, in the end, will be useful to all - service providers and their consumers, not to mention the general IT-climate in Russia.