Home > Articles > Networking > Storage

  • Print
  • + Share This
This chapter is from the book

Data Loss and Business Risk

Risk is a measure of potential economic loss, lack of return on an investment or asset, or material injury. Another way to state this is that risk is a measure of exposure to harm. Some common risks are material loss (for example, damaged equipment, facilities, or products), risk to sales and revenue, lawsuits, project failure, and market risk. Risk is associated not only with hard assets, such as building or machinery, but also with revenue, customer loyalty, and investments in projects.

How risk is measured depends on the assets deemed to be at risk. In computer security circles, risk is usually a measure of threats (the capability and willingness for malicious behavior), vulnerability (the holes in the system that can be exploited), and harm (the damage that could be done by a threat exploiting a vulnerability). No matter how you measure risk, the most important component is harm. Without harm, there is no risk.

Insurance, locked cabinets, background checks, and currency hedges are ways that companies seek to minimize harm to their assets and the profitability of the business. If one thinks of information as being a corporate asset, protecting the underlying data is necessary to ensure the value of the asset and prevent its loss. Ultimately, data protection is about mitigating business risk by reducing the ability of some threat to do harm to mission-critical data.

The Effect of Lost Data on Business Operations

Companies recognize that data loss represents a business risk. Even if a monetary value is not assigned to the data, the negative effects on operations can be significant. In many cases, corporate operations can be so adversely affected that companies feel the need to mention the risk in regulatory filings and shareholder reports.

Three types of damage may occur because of data loss. First, data may be unrecoverable. In this case, important business records may be lost forever or available only in hard-copy form. Any business process that is dependent on that data will now be considerably hindered. This is the worst form of damage that can occur.

Next, data may be recoverable but may require considerable time to restore. This scenario—the most likely—assumes that data is backed up in some other place, separate from the primary source. This is a better situation than irrecoverable loss, but the data will be unavailable while recovery operations take place. In some cases, not all the data may be recovered. This is a common problem with data restored from nightly backups. Any data created during the day when the primary data was lost is not on the backup tapes and is lost forever.

Finally, while data is unavailable, either permanently or temporarily, applications not directly related to lost data may fail. This is especially true of relational databases that reference other databases. Loss of a central database of customer information, for example, may cause problems with the sales system because it references customer information. A loss of this type can result in cascade failures, in which several applications fail because of their dependence on another application's data.

Risk to Sales

A company may suffer measurable harm when data loss makes it impossible for it to interact with customers. The result is that the company will not realize sales and revenue.

E-mail has become a primary form of corporate communication. Losing an important e-mail or attachment may mean that a customer may not be serviced correctly; thus, sales are lost. This is especially true of companies that sell capital equipment to other companies. A hard drive crash on the e-mail server may cause an important bid to go undelivered. The salesperson may not even know that the bid was not received by the customers (because it is sitting in the Sent folder stored on a local hard drive) until the sale is lost.

As large companies have become more dependent on call centers, they have become equally dependent on the customer relationship management (CRM) systems that help them track customer issues and orders. This represents a risk to sales, revenue, and profitability. If this risk is realized—if the worst-case scenario comes true—the harm done to the business may be severe enough to propel it into bankruptcy.

Inability to Operate

Extreme data loss such as loss of an entire database, even temporarily, has been known to cause organizations to fail. A company may not be able to fulfill orders, update employee records, produce financial reports, manufacture goods, or provide services. It may not even have an operating phone system. Computer technology and the data associated with it are integrated into all aspects of an organization's operations. Because of this dependence on information technology, there is a clear risk that data loss can make it impossible for an organization to perform properly.

Even partial data loss can disrupt business operations and produce negative effects. Employees may be idled for long periods of time while data is re-created or recovered, reducing productivity. Applications may fail unexpectedly when referencing data that is no longer available. Essential reporting may be incomplete because component data is not available.

Loss of data also makes it difficult for managers to measure company operations. Most modern businesses rely on financial, market, and manufacturing metrics. Without the ability to gather and report on key business indicators, managers are running blind as to the health of the business. Destroyed, damaged, or altered data skews metrics and disrupts decision-making. The overall effect of this type of disruption is reduced revenue and higher expenses, leading to loss of profitability.

Lawsuits and Fines

There is potential for lawsuits and fines when a company experiences data loss. With shareholder lawsuits fairly common, failure to protect data could easily lead to litigation, especially if data loss can be tied to a negative change in the share price of the company's stock. A more likely scenario is that data loss will affect operations and sales, causing the business to underperform. This can then trigger shareholder suits.

Other types of legal action can result in adverse judgments for companies. Companies may be sued for failure to perform duties outlined in contracts or the inability to produce goods and services that have been paid for. A lost order record may result in a customer's suing for direct and collateral damages.

Regulators now have the power to impose data retention requirements on companies. Data retention requirements tell a company what data must be kept and for how long. Fines can be levied when these requirements are not met.

It is not enough simply to have good policies; the policies have to be followed up with good practices. In 1997, Prudential Insurance was fined heavily because it did not properly implement existing electronic document retention policies. This led to the destruction of electronic documents needed as evidence. There was no indication that employees willfully destroyed evidence—only that the company did not take sufficient action to ensure that it was preserved. Though Prudential had a good electronic document retention policy in place, its inability to implement it properly cost the company $1 million in fines. [2]

Damaging legal situations can occur when data loss causes financial information to be released late. Regulators, markets, and shareholders expect certain reporting to occur at previously announced intervals. When a company fails to meet these expectations, that failure often leads to fines, lawsuits, drops in price of the company's stock, or even delisting from financial markets.

All these situations represent financial harm to the business. As such, steps need to be taken to protect the business against the risk of lawsuits and fines.

Theft of Information

Another type of harm that requires data protection is theft of corporate information. This may take the form of theft of secrets or a violation of private data. Theft of secrets happens when a thief is able to access internal company information vital to current and future operations. Some examples of this these secrets are product plans, product designs, and computer source code. The economic impact of theft of secrets is difficult to ascertain, because the harm is indirect and manifests itself over long periods of time.

Theft of private information, such as customer information, may have three effects:

  • Lawsuits may arise when it is known that this information has been stolen. Customers may sue for damages that result from the use of this confidential information.
  • Regulators in some countries may be empowered to take criminal and civil action against a company that suffers such a breach. The European Union, for example, requires that "Member States shall provide that the controller must implement appropriate technical and organizational measures to protect personal data against accidental or unlawful destruction or accidental loss, alteration, unauthorized disclosure or access, in particular where the processing involves the transmission of data over a network, and against all other unlawful forms of processing." [3]

    Other political entities have similar laws that require the safeguarding of information from destruction or breach.

  • Customers may refuse to do business with a company that allows such a theft of private information. It is reasonable to assume that a customer would not want to continue to do business with a company that has not taken adequate care to safeguard private information.

Reasons for Data Loss

As one might expect, there are many reasons why a corporation might lose important data. Broadly, they can be broken into the following categories:

  • Disasters
  • Security breaches
  • Accidents or unintended user action
  • System failure

Some data protection techniques can be applied to all these causes of data loss; others are better used for specific categories.


Disasters are the classic data-loss scenario. Floods, earthquakes, hurricanes, and terrorists can destroy computer systems (and the data housed on them) while destroying the facilities they are kept in. All disasters are unpredictable and may not behave as forecast. The goal of data protection is to create an environment that shields against all types of disasters. What makes this difficult is that it is hard to predict what type of disaster to guard against, and it is too costly to guard against all of them. Companies guard against the disasters most likely to occur, though that is not always good enough. Until just a few years ago, most U.S. companies did not take into account terrorism when planning for disasters.

There are two classes of disasters: natural and manmade. Natural disasters are often large in scope, affecting entire regions. Earthquakes and hurricanes, with their ability to do widespread damage to infrastructure, are especially worrisome; they rarely provide enough time to develop a plan for data protection if one is not already in place. After the disaster begins, it is too late to try to save data.

Manmade disasters are often more localized and generally create much less damage. Fires are the most common manmade disaster, although many other manmade incidents can cause data loss, too. The worse manmade disaster resulting in widespread loss of data (and life) was the September 11, 2001, terrorist attack on the World Trade Center in New York City. The destruction of key computer systems and the harm that it wrought to the economy of the United States led the U.S. Securities and Exchange Commission and the Comptroller of the Currency to jointly issue policies [4] requiring that data be adequately protected against regional disasters.

Security Breaches

When an intruder breaches the network, server, or storage defenses of a company, he usually has one of three goals: to look at information he shouldn't look at, to deny the company the use of its data, or to damage and destroy data. Because the harm is intentional, an intruder can do more selective damage aimed at long-term harm.

Intruders come in two types: insiders and outsiders. The press tends to accentuate the problem of outsiders, yet insiders are as big a problem. Insiders can do more damage because they already have access to vital systems (and don't have to work as hard to get at important data) and know what type of damage can do the most harm. Insiders also have the advantage of less scrutiny. Most IT departments have sophisticated methods of detecting outsiders trying to break in. Fewer companies monitor activity inside their network. For this reason, insiders can go undetected until they do damage, whereas outsiders are often stopped at the network perimeter.

Security concerns affect data protection strategies in two ways. First, it is important to keep backups or copies of data, in case a security breach results in damage or destruction of critical data. Second, part of the data protection strategy needs to be securing vital data and information assets against harm. Although network and server security is well formed and understood by IT professionals, storage system security is much less mature, in terms of both technology and best practices.

Accidental Data Loss

Accidental loss represents one of the most common data loss scenarios. End-users are often the culprits; they delete, overwrite, and misplace critical files or e-mails, often without knowing they've done so.

In the 1980s and early 1990s, it was not at all unusual for the help desk to get frantic calls from end-users who had reformatted their hard drives. Fortunately, changes in desktop operating systems have made accidental reformatting of a hard drive much more difficult, and it is now a rare event. Damaged or reformatted floppy or Zip drives are still a common problem, though this usually destroys only archive data. As other forms of mobile media, such as solid state memory devices, are used by more people, the likelihood of loss of data on these devices grows. And yes, people sometimes drop their smart media cards in their coffee.

Though IT personnel may feel frustrated by the silly errors end-users make that result in data loss, they are responsible for quite a few errors themselves. Botched data migrations, hastily performed database reconfigurations, and accidentally deleted system files are everyday occurrences in the IT world. One of the most common and most damaging IT errors occurs when a backup tape is overwritten. Not only is the previous data destroyed, but there is no good way to recover much of it. Also, quite a few backups are damaged due to sloppy storage practices.

The risk that the end-user represents is usually a recoverable one. Although it's a hassle to dig out backups and pull off individual files, it is still something that can be done if the data in question is important enough. Good habits, such as backing up files to file servers or automated backups and volume shadow copying (now part of the Windows operating system), can alleviate many of the effects of end-user data loss.

IT mistakes represent much greater risk. The effects of an IT accident are not limited to individuals; instead, they affect entire applications and systems, many of which are mission critical. Strict policies and controls are necessary to prevent these types of errors.

System Failure

System failures often cause data loss. The most famous type of failure is a hard drive crash. Although hard drives don't fail with the frequency that they used to, failures are still a major problem for many system administrators. This is especially true of drives in high-use servers, in which drive failure is inevitable. Data can also be corrupted or destroyed because of spurious errors with disk array hardware, Fibre Channel and SCSI host bus adapters (HBAs), and network interface cards (NICs). Fluctuations in electricity, sudden power outages, and vibration and shock can damage disks and the data stored on them.

Failures in software are also a source of data loss. Updated drivers and firmware are notorious for having bugs that cause data to be erased or corrupted. The same can happen with new versions of application or database software. The failure of IT to properly back up and verify the integrity of a backup before installing new software is an age-old problem leading to irrecoverable data loss.

System failures cannot be completely prevented, but steps can be taken to reduce the likelihood of losing data when they occur. One of the most common steps is to buy high availability (HA) devices for mission-critical applications. HA units offer better protection against shock, flaky electricity, and link failures that can corrupt data. They also have software protection that ensure that I/O is complete and that bad blocks do not get written to disks. Good backup and archive procedures are also important parts of a plan to protect against system failure.

  • + Share This
  • 🔖 Save To Your Account