In the past article we looked at security issues in terms of virus threats. In this article, we will consider the problem of data safety and dispel the myth that it is expensive and not reliable.
The service and data backup system is a little more than just a backup. The term "redundancy" implies both file copying (backup copies) and duplication of equipment and services, which entails a reduction in business downtime. Data security is an important component in any business organization. However, there is one factor that should not be "discarded" - this is the time to restore backups. The inability to quickly recover data after any attack and / or loss affects business processes in different ways. For one business, daily “downtime” will not bring losses, while for another, every lost minute costs a huge amount. Frequently encountered shortcomings in the data backup system in small and medium-sized businesses:
-
There is no backup. There are many reasons: there is no one and / or no time to do it, the fear of starting to do it, because it supposedly will take a lot of time and money.
-
Backups are written to the same hard drive and to the same hardware that acts as a server. In this case, we can assume that all efforts are equal to "0", since in case of equipment failure, the data will not be recoverable.
-
The backup plan is missing or incorrect. The plan is considered invalid if the backups have a short retention period.
-
The time factor for restoring services is not taken into account. The backup plan is correct, the copies have a long shelf life, but the business does not have time to restore the health of the services.
-
The allowable amount of lost data has not been defined. Lost data is data that was entered into the database, etc., before the crash, but it cannot be restored from backups.
The first three points are quite clear, but points 4 and 5 are worth considering in more detail. Both paragraphs are similar to each other, but differ in that the fourth paragraph describes the aspect of the service availability time, and the fifth - the allowable amount of lost data.
So let's look at the fourth point. It says that if you need a quick recovery of business processes, then you should consider a hardware or software cluster or online replication of services. The choice of solution depends on the requirements and capabilities of the business. If 10-15 minutes of downtime is not critical for a business, then online replication will do. There are businesses for which even a day of downtime will not be problematic, then services are restored manually in the event of a crash.
Fifth point. The allowable amount of lost data should be determined. If the business does not expect any data loss, then a failover cluster is the best solution. If the cluster is configured correctly, in the event of any failure, users will not even feel any changes, and the data will remain absolutely safe. If the amount of lost data entered in the last 5 minutes is not critical, then online replication is sufficient. If the amount of data lost during the working day is not critical, then it is enough to set up a backup once a day.
For very tiny businesses, it is ideal to store backup copies of data and services in various cloud storages. This option is good because you can get high-quality service and storage reliability for free or almost free of charge, you only need a reliable Internet connection. However, this method is not suitable for larger companies due to paid licenses in case of using paid cloud services and/or large amounts of data. In this case, it can be advised to use 2 different computers as storage, provided that the loss of data for a day of work is not critical for the business, and the recovery time of 24-48 hours does not affect business processes. The use of 2 different computers is extremely important for the safety of data and the speed of restoring services, because if the main computer fails, the duplicating one will be able to fully or at least partially ensure the smooth operation of services. It's even more convenient to use 2 NAS storage instead of computers, but this is a more expensive and somewhat limited solution. For a simpler backup, System Administrator LLC recommends NAS from dlink, but for better and more complex it is better QNAP.
Larger companies (from 50 people) or companies for which long downtime is critical should take a closer look at renting virtual machines, or buying and setting up their own servers. However, when deciding to buy your own servers, you should remember that this will also require the involvement of highly qualified specialists (whether in-house or outsourced) to maintain and maintain servers, and therefore increase costs. Good servers with high-speed disks are not cheap, and you need to buy them from at least 2 pieces (otherwise you will not be able to make the required fault tolerance), and you should also remember that servers need to be provided with a separate dust-free room with a constantly low temperature, an uninterruptible power supply system and grounding, smoke/temperature/humidity sensors. Indeed, without this, any servers will last much less than the manufacturer guarantees. For this level of business, our company offers various optimal solutions:
-
rent of virtual servers on our hardware servers in the data center;
-
"hybrid solution" - placement of services on the customer's servers and online replication to our virtual servers;
-
complex support of services on the customer's servers.
When planning IT costs for your business, don't forget the sheer number of free services you can run for your business. Paid cloud services guarantee you a high level of fault tolerance, at a low cost. And only if it is impossible to use cloud services, plan to create your own IT infrastructure.
Our experts will be happy to help you set up an infrastructure of any complexity and budget based on your needs and capabilities.