System administrators share on those who do not do backups yet and those who already do.
This folk wisdom appeared not at its own, it is a fruit of numerous cases which led to loss of the most valuable that business has – information. According to the researches, those companies which lost working data for the considerable period of time or hose which could not manage to restore an operating mode of work of the enterprise in a short time, in the majority lost the critical mass of customers or in general came to bankruptcy.
To understand which steps shall to be undertaken for the effective data security, first of all it is necessary to give definition for the next moments and concepts:
WHAT– what data are of commercial value and demand backup as they are operated, what specific actions shall to be undertaken for this purpose that the created backup copy of data could be restored consistent and without additional labor costs.
RPO (Recovery Point Objective) – how often backup processes have to be passed to narrow “a vulnerability window” of data as much as possible. We have to remember that excessively frequent backup raises technical requirements to a complex and leads to increase in the budget of the decision.
RTO (Recovery Time Objective) – a point in time (PIT – Point In Time) when your data will be restored from backup copies, and the infrastructure will come to an operating mode of service of business. This concept has a direct bearing on restoration time from a backup copy and the accompanying actions for restoration and restart of infrastructure from a backup copy.
The optimum solution of a problem of backup represents strict balance between these parameters. Only the careful analysis and a sober assessment will allow to create the decision possessing one more parameter, important for business — the acceptable price.
let’s CARRY OUT DETAILED ANALYSIS OF THE LISTED ABOVE PARAMETERS
WHAT. Modern business in its work uses very different types of data served by different applications each combination of which demands an individual approach to it for achievement of desirable result.
The most frequent are the main categories and problems related to them:
- File storages (file services). For the file service a problem point are the files opened on editing which cannot be copied without creation of specific conditions from outside the file service.
- Databases. In case of databases it is necessary to take measures for ensuring “completeness of transactions” and to consider in the course of backup structure of a DB and its distribution according to various tables/files.
- Mail services. The problems connected with mail services are similar along with the file services and databases: open objects, preservation of infrastructure of a mail service. Also recovery of the concrete letter or group of letters, but not all e-mail server can be often necessary. Existence of similar functionality important as it allows to considerably save time for recovery work.
- Means of collaboration. The most part of modern means of collaboration represent an application server and a database in which objects are stored. The difficulties connected with recovery of data have no fundamental differences from the DB servers.
- Virtual computers. It is necessary for effective backup of virtual computers support by service of backup of integration with a hypervisor and an operating system of the virtual computer functioning inside. The hypervisor at the command of means of backup has to provide cleaning of caches of operations of record on a disk and freeze a condition of the computer for the purpose of creation of a picture of its state which will be used for creation of a backup copy.
RPO. To determine this parameter for business, it is necessary to understand that backup is the instrument of creations of copies of data and there is a border of “a rationality of its application” as from the point of view of requirements to the equipment, and business, for achievement of a continuity of functioning of systems serve a cluster. In case of large volume of data, unreasoned architecture, small productivity of a complex there can occur “overlap” — the previous operation of backup did not end yet and the following already has begun.
RTO. Time spent for reduction of system in operating state after accident depends on:
- existence of the fulfilled plan of maintenance of a complex — Disaster Recovery Plan (DRP). It is important that such plan existed not only on paper, but also passed regular tests for compliance of the plan to a real condition of IT structure;
- the qualified personnel in IT division possessing necessary skills;
- integrity of data structure in backup copies in case of which violation it is necessary to spend an extra time on editing/recovery of data by technicians;
- productivity and readiness to accept the restored data of the equipment. There were some cases when the customer at the time of accident tried to pile up processing of the main data files on “iron not of the first freshness” that led to impossibility of work with data: there is service, it functions, but the resources available to it, are insufficient and the first serious inquiry leads to long expectation of result;
- backup service productivity which consists of the computing resources allocated under service and productivity of storage of backup copies. In case of allocation under problems of service of backup of small resources we ourselves doom ourselves to the big RTO window and losses. It is also necessary to remember that difficult schemes of backup with a large number of intermediate points raise requirements to resources of service of backup and increase RTO.
We already talked about WHAT we want to copy and about time intervals. It is high time to pay attention to WHERE we will copy our data.
At the moment as means of storage of backup copies the following technologies are used: tape drives, disks in RDX cartridges, the systems of data storage of general purpose filled with economic disks, specialized DSS and external storages of “cloudy type”. To be fair it should be noted that very big complexes of the software and specialized DSS are make off the term “cloud”.
TAPE. The key element is the cassette drive, exactly its model determines productivity, the interface of connection and type of the cartridge. At the moment the suppressing share of the market was received by technology of LTO cartridges (Linear Tape Open), the existing devices with DLT/SDLT (Digital Linear Tape/Super DLT) tapes can be incidentally met and AIT/S-AIT (Advanced Intelligent Tape/Super AIT).
LTO drives differ in the interface of connection (USB, SCSI, FC, SAS), a drive case form factor (half, full, double – the bigger size of the drive allows to place more productive and durable mechanism of the drive) and the supported standard of the cartridge. Considering long life cycle of technology it was put a lot of effort for ensuring compatibility: in most cases the drive of N generation can work with cartridges of its generation at a full speed, with the previous generation — at the reduced speed, and to read cartridges of the pre-previous generation at the minimum speed.
|Generation of devices|
|Year of emergence||2000||2003||2005||2007||2010||2013|
|Physical capacity||100 GB||200 GB||400 GB||800 GB||1,5 TB||2,5 TB|
|Maximum speed (MB/s)||15||15-40||15-80||60-120||80-140||120-160|
Physically tape solutions may happen as follows:
- separate drives for 1 cartridge (USB, SCSI, SAS);
- automatic loaders for 6-12 cartridges (SCSI, SAS, FC);
- tape libraries supporting from 24 to hundreds of cartridges and from 2 to tens at the same time working drives (SAS,FC).
As a product with a long-term experience of operation, the tape on the one hand got rid of all children’s diseases, and on the other hand all its disadvantages became clear:
- low cost of the separate carrier;
- mobility of the carrier, possibility of transfer of copies of data in the protected bank storage;
- low energy consumption.
- low speed of direct access to data because of consecutive access;
- lack of support of a deduplication;
- difficulties with simultaneous carrying out several parallel sessions of backup;
- need of duplication of copies for achievement of high reliability of storage;
- strict requirements to ensuring speed of a source of backup copies in connection with requirements of mechanical components of the drive, violation of this rule results in premature wear of the mechanism;
- big copies can be divided into some cartridges and thus to be less reliable;
- low maximum speed of backup.
RDX is an attempt to create the simple decision for small tasks without problem points of tape drives. Represents the head module — the receiver of cartridges with hard drives. This product was not widely adopted in our realities because the majority of small enterprises do not pay due attention to a backup perspective, and for those organizations which are concerned by safety of information, such product cannot satisfy requirements any more.
- mobility of copies;
- decent speed of reading/record, both consecutive, and free-hand.
- rather high price;
- it is necessary to have specific software of backup.
DSS of general purpose. It is admissible to use as a storage of backup copies in case of observance of the following parameters:
- the existence of the software of backup supporting a disk as the target device (very often this functionality is a separately licensed option);
- DSS has to support RAID 6 of the group of large volume or technology, similar by ideology;
- for comfortable operation of DSS has to support allocation of volumes/exports of large volume.
All listed parameters together do not represent anything difficult, but violation of any of them leads to impossibility of comfortable operation. Also it should be noted that use of the reserve software without functionality of a deduplication seriously increases the cumulative cost of storage of backup copies, thus reducing appeal of use of DSS of general purpose as storage of backup copies. At modern DSS of general purpose there can be technologies of a compression and a deduplication of blocks of data that in certain cases allows to apply them under a problem of backup even with traditional software without expanded functionality. Each such case demands very detailed planning for achievement of desirable result.
Specialized SHD for storage of backup copies represent the optimized hardware-software complexes. They can be presented both in the form of monolithic decisions (EMC Data Domain, HP Store Once), and in the form of appliance (IBM ProtecTIER, Fujitsu ETERNUS CS) the using DSS of general purpose.
Fundamental difference of specialized decisions from DSS of general purpose are the compression and a deduplication of data in real time. The data flow getting into a complex, contracts immediately and thus the useful capacity of disks of a specialized complex has very high virtual capacity.
Pluses of the specialized DSS:
- very high resultant capacity of the solution;
- opportunity to always use backup copies like “full” – backups of this kind demand least of all time for restoration from the copy because the software of the backup has no need to carry out difficult data processing and to choose the necessary blocks of information from incremental/differential backup copies;
- regardless of what types of backup copies we use and what applications we copy if at this backup copy there are repeating blocks of data, they will be deduplicated and squeezed;
- the best representatives of specialized DSS of storage of backup copies can replicate backup copies between platforms taking into account a deduplication and a compression of data, thus requirements to a throughput strip between platforms decrease in tens/hundreds of times;
- specialized drivers for software packages of backup allow to distribute loading between the server, the client and specialized DSS increasing decision productivity;
- flexibility of introduction — opportunity to present system as a file resource, virtual tape library, acceleration device;
- high speed of record/reading.
Minuses of the specialized DSS:
- examination at a stage of planning and introduction is necessary for achievement of the necessary results;
- the maximum coefficient of efficiency/ the minimum TCO is reached at observance of a set of factors.
The design of the correct BRS solution demands an assessment of parameters both computing systems, and the network environment, definition of types and parameters of the reserved systems and volumes of information. The balanced characteristics of the listed components provide achievement of desirable result.
Investigate together with our experts your situation in the field of BRS, receive qualitative recommendations, and store the information reliably!