The Four Challenges of Backup and Recovery

Q. What are some of the

challenges faced by organizations with respect to backup and

recovery?

A. It's a well known fact that data is growing tremendously in every

organization, but even more interesting is the fact that if data

grows from say 1 to 2 TB, then its backup data grows many times this

value! This was one of the findings of a recent Forrester research

report on the state of backup and recovery amongst organizations in

the APJ region. It was based on a survey of 550 organizations across

APJ, out of which 70 were large Indian organizations. The logic for

this is simple. Companies do weekly, daily, monthly and annual

backups, and are required to archive 7-8 years of data. As a result,

backup data ends up being 6-7 times of the actual data we backup.

The challenges faced by

Indian organizations with respect to backup and recovery are

different from the rest of Asia. While the primary challenge for

Indian organizations is related to optimization of the storage and

backup infrastructure itself, for the rest of Asia, the primary

challenge is managing data growth. 52% of organizations in rest of

Asia said this, while only 27% of the Indian respondents gave this as

their key challenge.

The number 3 challenge

related to backup and recovery is dealing with natural disasters,

thanks to the slew of disasters that struck the world in 2011.

Apparently, 2011 was the biggest year of natural disasters, which has

changed a lot of things. A lot of customers who used to backup data

on tapes were not able to restore them, not because the tapes were

corrupt, but because they couldn't even send people to get the

tapes as their data centers were in the basement. Japan for instance,

was the worst hit. Due to the Tsunami and earthquakes, even roads

were broken, due to which companies couldn't send their people

to get the tapes. Due to this, online disaster recovery has become

very popular since the last 6-8 months.

The 4^th

challenge being faced by organizations is related to backup as a

service. While we have been talking a lot about cloud computing, but

it's only over the last 4-5 months that people have seriously started

considering it and started putting non-critical data on the cloud.

This is happening because over time, people are finding it

increasingly difficult to manage large volumes of data. For instance,

the Oracle databases and their instances that were maintained by

companies used to be a terabyte or half. Now, their size has reached

double digit TBs. EMC for instance, has customers who have single

Oracle instances of 40 TB. Due to this, backing them up within a 24

hour window is just not possible.

Q. Advancements in

technology have improved data replication and backup speeds. Are they

not sufficient to deal with growing data volumes?

A. People have been doing backups since 1942 when backup meant tapes,

and unfortunately, most people still people follow the same pattern.

This just doesn't help because now, 90% of data in a lot of

organization is in digital format and not paper. Even your credit

card and mobile bills for instance, are in digital form. The volumes

have grown so much that the next production cycle starts even before

the previous backup has been done. Under such a situation, tape

becomes insufficient. People want quick backups. Moreover, now it is

recovery that has become extremely importance. The focus is on how

fast we can recover. So, people are looking for online replication.

There are products in the market that will start the replication as

soon as the backup starts. So your DR is ready instantly as a result.

Moreover, it's backed up to a drive and not tape, using a technology

called called disk to disk de-duplication.

Q. What's the benefit

of using disk over tape for backup and replication?

A. Earlier, tape was not able to keep more than one week's data.

Whatever was your weekly data, you would put it in the tape and ship

it out. You couldn't keep the tape for more than a week, but now with

de-duplication on disk backup, one can keep 30-90 days data on the

disk itself and replicate it to the DR side and then apply whatever

policies are there for backing it up to tape. So the impact on your

primary side backup is nothing. That's what lot of banks are

doing these days, and some of the large banks are even going

completely tapeless and are moving to Tier-1 to Tier-2 disks.

Q. Please comment on

the RoI of disk vs tape based backup and replication.

A. From

a RoI perspective, it would be expensive in short term, but if you do

total 3-5 years TCO and also look at the soft benefits like bandwidth

advantage and green IT, the savings are immense. If you backup 60 TB

of data to tape per day for three years, you end up generating 50.4

tonnes of CO2, but if the same amount of data is de-duped, you only

generate 34.2 tonnes of CO2.

We had this customer who

had 45,000 tapes and replaced them with 4 big tapes. The customer

would soon be going completely tapeless and will put 3 of our Data

Domain appliances into their racks. Data Domain was a company

acquired by EMC in 2009. As a result, they save the space of storing

their 45,000 tapes, along with the cooling and power. Their weekly

backup capacity is 35TB. They are moving from mainframes to open

systems, with their first tape migration, converting 45,000 tapes to

disk and total could be 100-120 TB. This is the power of

de-duplication. 8/10 banks in New York, 7/10 banks in Singapore, 7/10

banks in Hongkong, 8/10 banks in Australia are using our Data Domain

de-duplication technology.