As service is arranged
Service consists of server and client part. "Agentny" and "agentless" technologies, depending on infrastructure are used. On the client computer or the host server of virtualization is established the agent whose task is connection of the client computer / host with the Acronis Backup Cloud server and implementation of problems of backup and recovery.
About client part
Here what Agents (Clients) are developed:
- Clients of backup for Windows, Linux and Mac – are responsible for data backup by machines under control of Windows/Linux/Mac OS.
- The client of backup for VMware – executes backup of virtual computers on the VMware vSphere platform (ESX/ESXi) without installation of the corresponding clients in guest systems. This client is established by "screw" machines with access to vCenter Server and storage in which backup copies of virtual computers remain.
- The client of backup for Hyper-V – executes backup of virtual computers on the MS Hyper-V platform without installation of the corresponding clients in guest systems. This client is established on Hyper-V host.
- The client of backup for MS Exchange Server – Microsoft Exchange Server 2003, 2007/10, 20013 versions are supported.
- The client of backup for MS SQL Server – Microsoft SQL Server versions 2005, 2008, 2008 R2, 2012, 2014 are supported.
About server part
Let's consider how the server part is arranged. The server part of service consists of two components – system management and storage of backup copies.
The managing component is available on the Internet and allows to manage by means of the web browser backup of remote machines and already created backup copies; to create, edit and delete politicians of backup and policy of storage; to configure enciphering of the created backup copies, using AES or GOST the standard, and in case of need to save separate backup copies locally; to monitor a status of remote machines; to recover separate files / folders, disks/sections or entirely machines from a cloud directly on "naked iron". One of the most distinctive and useful opportunities is creation of hierarchy of the subordinated administrator and user accounts within which are distributed data access and to remote machines. Administrators can monitor a status of the subordinated accounts and in case of need give help.
The component of storage allows to unroll scalable, cheap and at the same time secure storage location of data. The storage for backup copies consists of server group in which client data register. For ensuring sufficient level of reliability of all stored user data each entering file breaks into "K" of blocks and then "N-K" (where N – some number bigger K) redundancy blocks with use of algorithm of error correction of Read-Solomon are added. All blocks are stored independently from each other, and safety of any "K" of blocks from written "N" guarantees recovery of the stored user data.
The data storage system of Acronis is under construction on physical servers. But at the same time server roles are assigned directly to disks.
There are three server roles: Server of Metadata (Metadata Server (MDS)), Server of Storage (Storage Server (STS)) and Front-end Server (FES).
The server of Metadata is responsible for information storage about fragments into which the file, and an arrangement of these fragments on servers is broken. It is the most critical component of system.
For providing the high level of availability and fault tolerance of storage it is recommended to have several servers with MDS role. One of servers becomes the basic, and metadata are periodically replicated with other servers with MDS role.
Besides, on each server having MDS role also the component of system management (MGMT) is installed. If the primary server of MDS ceases to work, the system management component automatically joins on other server with MDS role, thus, the web console of management of storage is constantly available.
The server of Storage (STS) is intended for storage of fragments of data.
Front-End the server allows clients of the Acronis Backup Cloud service to get access to storage and to perform data transmission between the user party and the data storage of Acronis.
Work with service
And now we will consider how it works at practice. And the companies of ultimate users access to management of service via the web console is provided to administrators of the company IT Lite (the supplier of service). On the chart given below the standard architecture of service of backup is shown. Blue shooters designate interaction between software components. Black shooters show how administrators and ultimate users get access to backup service.
Announce the rights, please!
Administrators IT Lite have access to management of groups and accounts of users.
Administrators of the companies of ultimate users have the rights allowing to manage the users who are only in their group! And ultimate users in turn get access to the console in which it is possible to add computers and to create the schedule of automatic backup. Service is integrated with the website of the supplier of service that allows users at once on filling of a form on this page to pass automatic registration.
Acronis Backup Cloud user's eyes
For the ultimate user work with system is simple and evident. After creation of the user account and logging into the personal account it is necessary to specify the computer for which the task for backup will be configured. For this purpose it is necessary to click on "+" and to select OS on which the agent will be set. The following step it is necessary to configure a task for backup.
In the console of management of the user all Tasks which were carried out for the specific computer earlier are displayed, in addition there is a Create a New Task and Recover Data from a Cloud function. Besides, viewing of a current status of a backup is available to the set computer.
Comparison with local backups
Data of users of the BaaS service are stored on servers which are located in the certified class TIER-3 data-center.
Thanks to the used architecture protection against failure at the level of separate servers and separate disks is provided that, we will notice, it is impossible when using RAIDs which are used for creation of fault-tolerant storages. In system complete check of integrity of data is also used. Level of redundancy is configured in the console of management of storage. The selfrepair design used when developing storage allows to avoid losses, typical for RAIDs, in system performance.
In case of failure of one of disks or even the whole server the system will make automatic rebalancing that allows to avoid immediate replacement of the failed components.
To Bekapit locally – yes, quicker, but it is less reliable
Forgot to mention about feature of use of backups as service – in case of emergency it is necessary to recover the computer (virtual computer) entirely. Yes, it is valid if to compare the speed of recovery of the server, then, of course, to make a local backup quicker. But in this case it is necessary to make a choice – or for benefit of speed, but without guarantee from jambs, or for benefit of reliability of a solution, but it is necessary to offer speed a little.
I will sum up so – I came to a conclusion that the question of reliability/simplicity of local backups and use of BaaS, is more likely a question valuable. The one to whom information is really valuable will select service rather, and the others – the first option. But it already my personal opinion, directly does not belong to business.
Summarizing everything stated above, would like to add that, as they say, a picture is worth a thousand words. Therefore it is better to work with service independently and to draw the conclusions, the benefit is the free test.
This article is a translation of the original post at habrahabr.ru/post/271047/
If you have any questions regarding the material covered in the article above, please, contact the original author of the post.
If you have any complaints about this article or you want this article to be deleted, please, drop an email here: firstname.lastname@example.org.
We believe that the knowledge, which is available at the most popular Russian IT blog habrahabr.ru, should be accessed by everyone, even though it is poorly translated.
Shared knowledge makes the world better.