Storage, DR and BCP

Storage matters to each and every business enterprise, its just they don't

realize its true value. There has been an explosion in the growth of storage. So

now, companies need to look more closely at storage security as storage

requirements are growing and becoming more complex. Quickly but securely finding

and accessing the information needed is now a challenging task for the IT

administrators. Traditional storage methods are unable to keep the pace with the

current enterprise needs and thus fail to provide required flexibility.

Importance of storage



The business value of data has increased dramatically over the past few years.
The amount of storage in enterprise data centers will continue to grow at

phenomenal rates. Actively managing and securing the stored data has become the

core process for many enterprises. The way in which new technologies are

reaching the market, it is becoming harder to understand and deal with the

scalability and manageability issues caused by this growth.

As data centers grow and become more complex internally, the level of skill

required to manage storage and data increase as well. Because of the rapid

changing technology there arises a critical shortage of skilled personnel. The

ever increasing business value of data has now made it the most valuable

corporate asset. As a result, storage has now gained importance as the only IT

resource of the company which assures data availability, integrity and

recoverability.

Value of data backup



Data is one of the most valuable asset for a company, also one of the most often
overlooked things when it comes to maintaining computer systems is the

importance of data backups. But a business cannot operate without data. Rather,

only a business that can manage its data can manage its costs, risk and customer

service. Delay in accessing data, or loss of proprietary data equates to loss of

revenue, reputation, diminished stock prices and worth of customers and

investors. Personnel, equipments and buildings can be replaced, but not lost

data. Further, you may lose as much as less than 50% of data due to hardware

problems, 30% due to human error, software failure accounts for 13%, virus,

theft and natural disaster



accounts for 15% (IDC reports). The value of lost data varies widely, depending
on incident. Lost data can possibility be recovered by recovery specialists but

it is VERY expensive and time consuming. And even if the lost data is recovered

there is no guarantee that it will be in a completely usable state. So a daily

data backup program should be priority for any business enterprise.

Several emerging storage trends are driving the development of new

technologies that organizations will need to implement in the next two to four

years. Technologies that will hold the most immediate promise for storage

managers include the following :

Virtualization



The process of representing physical storage in logical form is what we call
virtualization. There are two types of virtualization : storage based and

network based. Storage based virtualization is most commonly used. Network based

virtualization is a recent development in storage industry and is gaining its

market share rapidly.

Advantages of virtualization include simplified management, increased

scalability and availability, reduction in downtime, more efficient usage and

allocation and reduced management costs. There is no preferred standardization

yet, so it can be configured with different standards and schemes.

It all depends on which vendor product you've chosen as different vendors

design and build their product as per their understanding of virtual storage

system. The architectural constructs being developed by major vendors include :

SAN appliance, SAN in-a-box and distributed enterprise solution. For choosing

the right architecture which fits your enterprise requirements, you'll have to

consider your existing network infrastructure and types of storage devices used.

Grid storage



The term GRID makes it clear that it's a large meshed network technology with
coordinated management, where no single centralized switch controls routing. GSM

networks or telephone networks can be stated as an example for Grid Technology.

It has multiple nodes connected to several switches. Each node is redundant and

its own storage medium. To enable grid storage services, you need a file system

which supports grid storage service models and goals and a system directory

service to manage common namespace and virtualized data. There are three main

advantages of grid storage. First, because of its new feature of fault tolerance

and redundancy, the network can reroute the path another way, if a storage node

fails or if there is any congestion between two nodes.

This results in reduced downtime and reduced cost for online maintenance.

Secondly, unlike the older storage methods, Grid Storage is more flexible as it

can automatically recognize the new storage node, if added in the existing grid.

As a result, implementation cost and downtime issues are reduced considerably.

Thirdly, it maintains high level of performance under fluctuating conditions

because of existence of multiple paths between each pair of nodes. Thus, a grid

storage system is consistent, flexible, scalable and reliable.

Intelligent storage switches



There's a new generation of storage switches on the rise, and many vendors have
already started heralding them as the future of storage networking. SANs, NAS

and virtualization each help you manage storage more easily. However, when used

in isolation they simply make bigger islands of storage, which still need to be

managed separately. This could be accomplished with intelligent storage switches

as it has the advantage of managing through a centralized pool of disks. An

intelligent storage switch provides Ethernet interfaces for access to LANs, MANs

and WANs, and file-level access (NAS) or block storage over IP (iSCSI, for

example). The switches' Fibre Channel ports provide block-level access to disk

arrays or hosts. Main advantages offered by intelligent storage switches are

modular design, multiprotocol support, In-line virtualization, integrated load

balancing, separate disk arrays and high speed data mover at the core.

Continuous Data Protection (CDP)



One of today's most talked about storage management topic is Continuous Data
Protection due to its ability to recover almost infinite data very quickly in a

matter of seconds. As with any new technology there is not yet universal

agreement in how CDP is designed, deployed or even an exact definition of the

technology. SNIA defined it this way: “A methodology that continuously

captures or tracks data modifications and stores changes independent of the

primary data, enabling recovery points from any point in the past."

CDP is very different than traditional backup, even backup to disk. CDP

system provides protection against data corruption and preserves a record of

every recent transaction. In case of any hardware failure or virus attack, if

the data is corrupted, CDP system has the ability to recover the most recent

clean copy the affected file.

NAS accelerator



With the development of recent storage over TCP/IP technologies, NAS accelerator
technology has gained momentum. A NAS Accelerator is a kind of a network circuit

card which offloads TCP/IP processes from a microprocessor.

It is also know as the TCP/IP accelerator network interface card. Often fiber

channel technologies require highly skilled personnel but this is a file level

storage technology which allows administrators who are familiar with Ethernet

communications. Moreover it also reduces latency, and increases throughput.

Performance obtained by doubling the number of processors is equal to the

performance obtained by installing one NAS accelerator. This reduces initial

cost of implementation and maintenance. As the microprocessor is not burned with

TCP/IP processing it allows network applications to run faster. A high-end NAS

accelerator can process Internet functions, file transfers, data backup and

archiving at high speed.

InfiniBand



Infiniband technology utilizes a high-speed switched fabric communication
architecture interconnect, providing multi-use interface with minimal CPU

utilization. It was previously know with a working name of SYSTEM I/O. It is a

point-to-point bidirectional link topology. This technology is intended to

connect with high speed processors. It address greater performance, low latency,

high bandwidth and is designed to reduce total cost of ownership. It is more

efficient and reliable as it creates multiple redundant paths between nodes. It

has the potential to be a replacement for existing legacy shared bus approach (PCI/PCI-X)

as it has many advantages over it.

Storage Resource Management



The cost of storage is getting cheaper, which encourages people to store more
and more data. The problem is that much of the data is stale or redundant, which

is wasting valuable space. SRM tools provide a means to identify which data

should be deleted or archived. It refers to software that manages storage from a

capacity, utilization, policy and event-management perspective. Its the process

of optimizing the efficiency and speed with which the available drive space is

utilized. As a result, companies can free up space on the disk, reduce

administrative costs in managing it and save time and money in what it would

take to back it all up.

All SRM tools provide two basic functions, that of monitoring and alerting.

IT organizations that implement these tools can better monitor the rate at which

the capacity is being used, by whom and by how much. Complexity in the computing

environments, and rise in the value of corporate data at a rapid rate has lead

to the increased importance of SRM. There are lot of SRM solutions in the

market, both as stand-alone products, or as part of an integrated program suite.

Most businesses rely heavily on data in some form or another. And with the

availability of computers, a lot of this data is now found in digitized form.

But what do you need to do to protect this data-sometimes critical for the

survival of the business itself-from any disaster?

Disasters are classified into two types at the top level-natural and

man-made. Natural disasters include earthquakes, fires, tsunamis etc. Manmade

disasters include viruses, equipment malfunctions, strikes and even terrorist

attacks. And making sure that your business remains functional means having to

deal with these if they happen. This is where Business Continuity Planning (BCP)

comes in. BCP includes Disaster Recovery (DR) as one of the steps in the whole

process. We shall take a overview look into the process of ensuring business

continuity.

Awareness



One of the most important aspects of BCP is that the top management of the
business or organization be aware about the importance of BCP and DR. Only then

will the implementation of BCP and DR processes be filtered down throughout the

organization for taking care of the adverse circumstances.

Identify critical PoF



One of the initials steps in BCP is to first identify all critical PoF (Points
of Failure) in the business. These PoFs are what require to be planned for

exigencies and have DR protocols laid in. All things critical to the running of

the business come into this. For instance, this includes critical documents like

financial and legal information, etc. , critical processes like auditing,

quality control etc, critical equipment like data servers and others.

Identify threats and impact



There are a large number of ways that can affect a business, a list of

threats, risks and potential fallouts if that scenario occurs need to be

identified. For instance, if earthquake is a potential threat, work out what

parts of the PoFs can be affected by it.

Create solutions



Each scenario outlined in the sections above must have a solution designed for
ensuring minimum or no downtime. There are two parts to this. One, the PoFs

and/or the threats are non-technology related. For instance, a bird flu epidemic

or a government sealing drive affects people and not technology. But you still

need to plan for this. Two, technology gets affected as a PoF and/or threat. A

terrorist strike can break down your data center and so can a malicious computer

virus. You must have DR strategies in place to quickly get back to running

business.

Implement and test



If the DR strategy requires certain new or modified processes, ensure that these
processes are implemented. Creating a document that says 'Backups must be

moved off site into a secure vault to recover from earthquake.' does not do

much good if the backups are piled up in the local closet. A successful

implementation is one that is tested thoroughly. Making sure that all scenarios

are catered for is essential in this phase.

Maintain and tweak



Regular maintenance and tweaking the processes is an essential task too. As time
goes on, the threats and PoFs also change and making sure that your BCP

addresses these changes is key.

Disaster Recovery Tiers

The following tiers are taken as an unofficial standard for disaster recovery scenarios. Tier 0-No off-site data and no BCP: There is no recovery plan and data loss can happen Tier 1-Data backup off-site but not hot site: This means that the DR plan takes care of storing the data remotely, but there is no backup “hot” site where the data can be quickly restored for starting work again. Tier 2-Hot site: Data backup is taken off-site. There is also a provision to quickly restore this data either in the remote site or another backup site to start business processing from there. Tier 3-Electronic Vaulting: Tier 2 + ability to move data using extremely high speed transmission to electronic vaults so that data is more current than using the standard backup/restore technique. Tier 4-PiT: Point-in-Time copies are used for BCP where data recovery has to be much more current as well as faster. Disk based solutions are generally used in this case. Tier 5-Transactional: Using standard transactional integrity scenarios ensures almost nil data loss between the production and recovery centers. Tier 6-Disk mirroring: Having the best state to maintain current data, this requires some form of disk mirroring technique combined with Point in Time snapshots. Different vendors provide different synchronous and/or asynchronous implementations of this. Tier 7-Tier 6 + Automation: This is the best possible DR scenario where data concurrency, quick recovery combined with completely automated procedures exist.