Disaster recovery methods in Russia
span tape backup to cloud disaster recovery, but for businesses in the most
critical sectors in-house provision is the favoured method
In recent years, many Russian companies have recognized the
importance of data protection and disaster
recovery schemes. And in some sectors of the economy,
especially finance and telecoms, they place a special emphasis on the issue.
Today, most companies' preferred method is still to
duplicate their entire server estate, while more contemporary methods such
as the tools in
server virtualization environments are not yet commonly used.
However, there are providers that offer more advanced
disaster recovery systems.
"Based on our experience and customers' requests, the
most popular service is disaster
recovery as a service [DRaaS]," says Maxim Berezin,
head of cloud computing practice and director of virtual data centers at IT
company Croc.
"The idea of the service is the replication of
copies on several separate clusters of a cloud platform, which helps to quickly
move data from one location to another in case of a disaster.
"Tape backups are
still used by many companies, too, as they are relatively cheap and can be
physically moved from one data processing centre to another," he adds.
"Their main drawback is that they are not intended for automatic transfer
to a backup data processing centre."
Meanwhile, many companies prefer to run disaster
recovery schemes in-house rather than outsource them, and talk
about using locally developed systems, which might be in line with government
officials' recent statements about lowering dependency on imported software.
Disaster recovery testing is standard procedure at Uralsib Bank
Uralsib Bank, which is in the top 25 in terms of assets, has
its own disaster recovery scheme. The scheme comprises several components based
on products from various suppliers, with software from IBM, HP and Symantec.
"The risk of losing data is there all the time,"
says Alexander Shestakov, senior executive director for IT at Uralsib.
"Data is regularly lost as a result of equipment malfunction and due to
users' mistakes.”
At Uralsib, the disaster recovery scheme uses regular
backups to separate servers, as well as traditional tape backups.
"We receive data recovery requests on a regular basis,
roughly one or two times a week," says Shestakov. "In most cases,
those requests concern technological system updates and users' requests for
recovery of online data.
"Using software from several vendors allows the bank to
flexibly use the advantages of different products for different tasks," he
adds.
The disaster recovery scheme at
Uralsib is regularly
tested for its operability and integrity of backup copies.
Tests are implemented as part of existing standard procedures, as well as
during updates of banking systems' technological data, for instance by means of
recovery from a backup copy to a test environment or a development environment.
"By doing that, two tasks are
implemented simultaneously," says Shestakov. "On the one hand, the
test environment or the development environment is being updated. On the other
hand, integrity of backup copies is tested and staff
disaster recovery skills are also tested."
However, he says testing the
integrity of backup copies cannot be achieved in all cases. “For example,
a NAS device
doesn't have a test environment and there is no need to update it.”
Vimpelcom
classifies recovery data according to backup frequency
At Vimpelcom, one of Russia's top
three mobile phone operators, data recovery is also handled in-house by a
department focused specifically on that task, says a company spokesperson.
All data is divided into four
recovery classes – Class 1 to Class 4 – and for each class there is a backup
procedure. Class 1 data is backed up every two hours, and to two backup
platforms. Class 4 data is backed up once a day, with a second platform being
optional.
Disaster recovery tests are run
regularly for Sarbanes-Oxley-critical
systems, with each system being tested once a year.
RusHydro
enhances disaster recovery capabilities for the future
RusHydro, a major hydroelectricity
company with branches in the country's far-flung regions, has a complex,
multi-level system for data storage and disaster recovery which has been in
operation since 2005 and is outsourced to an IT provider.
It is constantly updated, based on
the size and content of data in storage, and has been developed using products
from Hewlett-Packard.
"Over the past 10 years there
have been all kinds of situations, including malfunction or failure of the
equipment, and we have had to restore the system's functionality quickly,"
says Garald
Bandurin, director of the IT department.
The most recent problem occurred
when a storage subsystem malfunctioned, but data was restored quickly.
"Overall, the system was
designed is such a way that we haven't had any fatal case of data loss in all
these years, and disaster recovery is a regular procedure built into it,"
says Bandurin.
But, although the existing disaster
recovery scheme has worked well, the company is looking to improve and develop
it in-house as RusHydro begins to use new applications, such as SAP, which is
to roll out in test mode in July 2015.
Another reason to set up an in-house
data storage and recovery scheme is a new project aimed at multi-dimensional
modelling and designing of power facilities.
According to Bandurin, the new
scheme is to be based on local products that have begun to arrive in the
market. A test ground for new systems is currently in place.
"In-house development doesn't
mean we are going to be goofing around, breaking existing rules for information
system design or stealing work from systems integrators," says Bandurin.
"We have a very traditional
approach, but at the same time we want to be at the same level as leading
companies,” he says. “Sometimes we have to deal with tasks that no one has yet
implemented. We have to look for our own solutions when it comes to information
systems and design," he adds.
Bandurin is confident about the new
scheme and in-house handling of disaster recovery. "I believe that in
three years' time we won't have a single critical part in our distributed IT infrastructure,
the disappearance of which from the operative IT scheme could lead to
disaster," he concludes.
No comments:
Post a Comment