Tech Roundup: Disaster recovery tools

LoveJunForever

A
disaster recovery plan (DRP)
refers to an organization's procedures for preventing data loss after a
catastrophic event by copying data to a remote facility.
But before investing in a DR site and buying equipment, a company must
ask itself two questions: What data is so vital that, if it were lost,
the company would go out of business? How much time can the company
afford to lose if the local site went down?
Only then can hardware and software components be purchased and a DRP implemented.
This tech roundup discusses how DR tools are used, their evolution and what the competitive landscape looks like.
               
                       
                        By Shane O'Neill and Beth Pariseau
                         13 Jul 2005 | SearchStorage.com
                         A
disaster recovery plan (DRP)
refers to an organization's procedures for preventing data loss after a
catastrophic event by copying data to a remote facility.
But before investing in a DR site and buying equipment, a company must
ask itself two questions: What data is so vital that, if it were lost,
the company would go out of business? How much time can the company
afford to lose if the local site went down?
Only then can hardware and software components be purchased and a DRP implemented.
This tech roundup discusses how DR tools are used, their evolution and what the competitive landscape looks like.
Definition
The main aspect of a DRP -- sometimes referred to as a
business continuity plan (BCP)
-- is the off-site replication of data. There are a number of
replication tools, and they're growing more common as the cost of disk
and bandwidth decrease.

There are two main types of off-site disaster recovery replication:
synchronous (SR)
and
asynchronous
replication (ASR).
With synchronous replication, data from site A is backed up
simultaneously and continuously at site B. While ideal, synchronous
replication is often too expensive for most except the largest
companies.
Asynchronous replication, in which data from site A is backed up at
site B, but not immediately, is available at significantly lower costs
and has made DR more accessible to more businesses.
According to Arun Taneja, consulting analyst and founder of the Taneja
Group in Hopkinton, Mass., a company must know its recovery point
objective (RPO) and recovery time objective (RTO) before it can know
what kind of replication it should be using.
"If you only think about equipment, you will just be spending lots of
money and still be naked, so to speak," Taneja says. "You first need to
know what data is mission-critical and how long your main data center
can afford to be down."
Key vendors and products
There are three types of off-site DR replication products: array-based, third-party and managed services.
Array-based replication employs software that allows users to replicate
between two proprietary storage arrays. An example of this is SRDF
offered by EMC Corp., which replicates between two EMC Symmetrix
systems. Other products in this area include IBM Corp.'s Peer to Peer
Remote Copy (PPRC), between IBM arrays, and Hitachi Data Systems' (HDS)
TrueCopy, between two HDS arrays.
Third-party products, which replicate data from a site A heterogeneous
environment to site B, include FalconStor Software's IPStor replication
software, NSI Software's DoubleTake and GeoCluster, Topio Inc.'s Data
Protection Suite, SANRAD's Global Disaster Recovery (GDR), and Kashya
Inc.'s Data Protection Appliance, which is distributed by Unisys and
Xiotech.
Managed services are newest in this arena and are gaining popularity
with mid-sized companies. With managed services, an outsourcer operates
a company's remote site or the whole DR operation.
Among companies offering such services are: WilTel Communications, with
a SONET network with space for rent; StorageTek, with services through
Arsenal Digital Solutions Inc.; Evault; IBM Global Services; and
Hewlett-Packard Co. (HP), which has at times taken over entire data
centers for DR purposes.
Trends and innovations
Tony Asaro, senior analyst at Enterprise Strategy Group in Milford,
Mass., says that one widely used method of data replication is
installing host-based software applications on individual servers and
then replicating all the data to a remote server running the same
software.
However, Asaro says most companies are still doing DR the old-fashioned
way: sending tapes off-site. "Naturally, this is not the most efficient
method, but it does provide a low-cost solution," Asaro says.
Yankee Group senior analyst Stephanie Balaouras agrees that despite
technical innovations, users have been slow to implement a
sophisticated DR plan.
A survey conducted by Balaouras last year showed that a high number of
users still do not deploy DR strategies to the extent that they could.
"The highest was financial services with 54%," she says. "I was kind of
expecting a little bit higher at this stage of the game."
According to Balaouras' study, the higher the company revenue, the more
likely they were to have long-distance DR replication products in use.
"I think it's totally been a cost issue," she concludes.
Balaouras adds that the availability of modular, affordable mid-tier
storage, the dropping cost of bandwidth and a wider variety of options
for network transport is raising both awareness and use of DR
replication technologies.
Going forward, Asaro predicts that innovations such as data
de-duplication, data compression and encryption will help increase DR
plans. "Through de-duplication, companies can only send unique data
over the WAN, thereby minimizing bandwidth requirements and network
costs," Asaro says.
Asaro highlighted Asigra Inc., Avamar Technologies, Inc. and Data
Domain as vendors providing de-duplication and data compression.


本文来自ChinaUnix博客，如果查看原文请点：http://blog.chinaunix.net/u/8368/showart_35645.html