Guaranteed and non-guaranteed channel: what to choose?

Important for any web project is used channel of access to the Internet. The normal functioning of any modern web service is impossible without good bandwidth. The lack of attention to the issues of speed and quality of connection to the Internet can have serious consequences: the outflow of users, the undermining of reputation, the shortfall in profits ...

As you know, there are two main types of channel: guaranteed (English guaranteed bandwidth) and unguaranteed (in professional jargon, it is also called spherical - from the English shared bandwidth). Consider each of these types in more detail.

')

Non-guaranteed channel

A non-guaranteed channel is a channel in which there is no constant connection speed, and all resources are constantly dynamically redistributed between users.

How is the connection organized via the non-guaranteed channel? The internet service provider has a channel whose bandwidth is limited. This channel is connected to some uplink. Channel resources are shared between users. At the same time, there are always more users than the bandwidth allows. This practice is called oversubscription (English oversubscription). Quite real is the situation when a channel with a bandwidth of, for example, 200 Mb / s (this figure is no more than a conditional one) can be used by several hundred, or even more than a thousand clients. Why is this possible?

All of these clients never use the channel at the same time, so it is never loaded at full capacity. There always remains a free reserve, which is constantly redistributed among active users. The undoubted advantage of this practice is the ability to provide users with an Internet connection for a low price.

But at the same time, a constant connection speed is by no means guaranteed, since the channel resources are constantly redistributed between users. At peak times, the speed may drop to critical limits.

Payment for using an unguaranteed channel is usually charged on the basis of actually consumed traffic. This is exactly what attracts many users who are guided by the following logic: why pay extra money for a guaranteed band, if I still have a small amount of traffic? In theory, all this looks quite convincing, but in practice the situation is completely different. Users who have chosen a non-guaranteed channel will inevitably have to face a number of drawbacks.

First, the connection speed over the non-guaranteed channel is almost always lower than the one stated by the service provider. Of course, certain numbers are indicated in the contracts, but they are only the maximum possible: due to the fact that the channel’s resources are constantly redistributed among numerous “neighbors”, in most cases they cannot be achieved in actual practice (in all legal documents this point is often described in the help of streamlined formulations that not every user will understand the first time). Some companies almost openly warn that servers connected to the Internet via an unguaranteed channel are not recommended for storage and distribution of “heavy” content.

Secondly, there is always a risk of channel degradation : the connection speed changes constantly, and if overloaded, it can fall to a critical minimum, which will make it impossible to transmit over a channel even small amounts of data.

Thirdly, one has to pay for very uncomfortable and unprofitable tariffs for using the non-guaranteed channel. More precisely, at first glance, these tariffs seem low and profitable, but in reality everything is completely different.

The principle of payment for the actual consumption implies that for any excess of traffic you need to pay extra. An unexpected surge in network activity can result in serious expenses.

Some budget hosters use a tricky move, stating that they do not charge customers for using a non-guaranteed channel if they do not exceed the established volumes of consumed traffic. But free cheese, as we know, happens only in a mousetrap.

When traffic consumption is exceeded, the connection speed is automatically reduced. This practice is found among many hosting providers: as soon as the volume of outgoing traffic goes beyond the established boundaries, the speed decreases, for example, from the stated 200 Mbps to 10 Mbps. To remove the restriction, the client will have to pay an additional amount. The described restrictions are more or less tolerable: some companies set limits not only on outgoing, but also on incoming traffic. So all favorable prices are beneficial only on paper. Strict cost planning in the case of a non-guaranteed channel turns out to be difficult.

Guaranteed channel

A guaranteed channel is a channel with a guaranteed (i.e., constant) data transfer rate. How is this guaranteed speed ensured?

Imagine two points connected by a physical communication channel. On its basis, with the help of special equipment, an electrical communication channel is obtained that has a certain bandwidth. The result is an information channel with a certain boundary speed. When one physical channel is used by several users, the common information channel is divided into several subchannels. Inside each of these subchannels, a certain speed is guaranteed. In some cases, the boundary speed inside the subchannel can even be increased if the resources of other subchannels are free.

A guaranteed channel is usually symmetrical. This means that the speed of incoming and outgoing traffic When transferring large amounts of data in both directions, the speed never decreases.

The guaranteed channel has the following undoubted advantages:

1. Stability. Internet connection speed is always constant, even at peak times.
2. Guaranteed quality. When organizing an Internet connection, the user and the service provider enter into an agreement on a guaranteed service level (Service Level Agreement, SLA), which specifies the required connection speed, which should always be provided. The SLA also clearly specifies the time to restore access to the Internet in case of malfunctions.
3. Fixed payment. The fee for using the guaranteed channel is charged at a fixed rate. In such a situation, it becomes much easier to plan the costs of project development. It seems to some inexperienced users that they overpay without loading the guaranteed channel “to the full”. It is hardly possible to agree with this opinion: as attendance grows and loads increase, it will become clear that a fixed tariff is much more profitable than payment for actual consumption, which in many cases can turn into unpleasant surprises (see above).

What to choose?

Based on the above arguments and facts, it is possible to make an unequivocal conclusion: the non-guaranteed channel is completely inappropriate for a modern developing web project. The pricing policy of many Internet service providers offering connection through an unguaranteed channel is dishonest: the stated speeds are almost never achieved in reality, and information about overpayments and traffic restrictions is often not communicated to the client in clear text.

Finally, nowadays, when the share of “heavy” multimedia content makes up most of the global Internet traffic, the choice of non-guaranteed channel seems a rash and short-sighted step.

The distinctive features of a guaranteed channel are stability, as well as an honest and transparent pricing policy: you will always have a connection at guaranteed speed for a fixed amount. The cost of connection via a guaranteed channel fully pays for itself as the project develops and expands.

Readers who can not leave comments here are invited to our blog .