Cloud Server Technical Conditions

In order to apply Arvancloud Cloud Server infrastructure (IaaS) for critical and high-risk cases, you should use it with a special configuration and alongside other Arvancloud services (or alternative solutions). Also, keep in mind that in some technical conditions, these services are not recommended to users at all.
Based on their sensitivity levels, each of your services can be placed in one of the main categories below:

Using the default configuration of Arvancloud’s computing infrastructure services is recommended only to the services with a sensitivity level of Moderate or Low. In other cases, you need to take into account the risk levels, requirements, and proceedings that are mentioned below:

Internet Low Speed or Disruption Risk

Currently, the internet connection in all the data centers of Arvancloud (in Cloud Server or IaaS infrastructure) is supplied by the upstream operators. As a result, the slightest fault, such as disruption or low internet speed from the operators is out of the control of Arvancloud, and it can cause problems in the performance of servers. This is a subject that you need to keep in mind when building a server.

Storage Speed Risk

Cloud storage can be divided into Network Storage and Local Storage, both of which can be fast (Hot storage or SSD), average-speed (Warm Storage or SAS), slow (Cold Storage or HDD), or very slow.
When using Local Storage, due to the direct connection of the disk with a physical server, the risk of data loss is high. However, the process of writing and reading is very fast with almost no latency. On the other side, Network Storage ensures data security by writing on 3 different servers and disks. But, this will cause latency in writing and reading.
Currently, Arvancloud only offers Network Storage in the form of SDD and HDD. Notice that SDD disks can be used with a minimum latency of 8 milliseconds. and a maximum of 30 ms. Therefore, if a database needs a latency of below 8 to 30 ms for writing and reading, using the current Arvancloud Block Storage will not be recommended as it is based on Network Storage.

Hardware Failure Risk

In Arvancloud architecture, when the hardware is taken out of service for updating, repair, or maintenance with prior notice, the users and all the instances connected to the server will be migrated to other servers without any disturbances. However, if there is an abrupt unavailability due to unexpected issues like a hardware failure of the server or a power fluctuation in the data center, it will result in disruption of the instance.
Currently, the typical time taken for all the instances to be transferred from one physical server to another is approximately 60 minutes.

Instance Resizing Risk

At the moment, Arvancloud has managed to successfully execute 99.2% of instance resizing. However, in certain cases, disruption may lead to irreparable damage. So, to resize the instance, it is advised to keep the following points in mind beforehand:

• Turn off the instance before resizing it.
• Create a backup and a snapshot of the instance before resizing it.

Local Storage Usage Risk

When it comes to using Local Storage disks, remember that they might not be resistant to failure and data loss because these disks are physically attached to physical servers.
Data loss can take place in the following cases:

• If for any reason, there is damage in a physical disk at the infrastructure layer.
• If the instance is migrated from the physical server for any cause. (e.g. resizing)
• If an issue appears on the physical server or the disk. (In this case, the data will not be accessible until the problem on the server is fixed.)

Critical/High Sensitivity Service Requirements

In services with a sensitivity level of critical or high, you should take the following measures:‌

• The service must be established and deployed in more than one cloud data center.
• Snapshots and backups of the server should be taken at regular intervals to ensure that the service can be restored in case of any failure.
• High Availability (HA) should be fully implemented in every data center at any level, and at no point should only one instance be responsible for a function.
• Use Arvancloud CDN and cloud security features (or any other alternative solution) to distribute the load across data centers and check the health of servers and services.
• Use solutions for Horizontal Scaling of important services.
• Use Arvancloud Object Storage (or any alternative solution) for regular backups.