CHAPTER
INFORMATION IN THIS CHAPTER
How Denial of Service Works Dangers of Denial of Service Defense against Denial of Service The Future of Denial of Service On April 26, 2007, the nation of Estonia was hit with a denial-of-service (DoS) attack. The attack lasted, off and on, until May 18th of the same year. The attack effectively cut off Internet access for much of the country. Members of the Parliament could not access their e-mail, people were unable to access their online banking accounts, Estonian news agencies could not communicate outside the countryâs borders, ATMs ceased to work, and citizens traveling abroad discovered their debit cards no longer worked.1
Estonia was not overcome because of outdated infrastructure. It was (and is) one of the most âwiredâ countries in Europe, thanks to their TiigrihĂźpe (Tigerâs Leap) project. In Estonia, as in France and Greece, Internet access is regarded as a basic human right, and the Estonian government has invested heavily in information technology (IT).
One might also be tempted to dismiss an Internet outage as nothing serious. Of course, if your business depends on the Internet, you may feel differently. Estoniaâs largest bank, Hansabank, is estimated to have lost around $1 million as the result of the attack. Banks are increasingly dependent on Internet banking and foreign money transfers, and thus an âalways onâ Internet. If the Internet is your business, as with Amazon.com and eBay, the effect can be disastrous.
Was this attack the result of careful planning by a foreign government? It now seems likely that the attack was organized and coordinated by one man: a 22-year-old Russian named Konstantin Goloskokov. He apparently carried out the attack in protest of the Estonian governmentâs decision to move the Bronze Soldier, a war monument in Tallinn erected by the Soviet Union in 1947. At the time of writing, the Estonian government has arrested and convicted just one person: Dmitri Galushkevich, who took part in the attack working from his laptop.
DoS attacks are on the rise and can be perpetrated by large-state actors, experienced hackers, or even by novices (âscript kiddiesâ) following any of the âhow-toâ manuals found on the Internet. DoS attacks can be launched for any number of reasons, from political protest to espionage and even extortion. These attacks can be intentional, like the one just described, or unintentional, like the âSlashdotâ effect.
As an example of unintentional DoS, suppose several aggregators, including SlashdotA and Digg,B pick up your essay on why Data was the best acting captain in Star Trek history. Now, thousands of people are visiting your site every minute, and the bandwidth allocated to you by your Internet service provider (ISP) is quickly used up. Now nobody can get to your site, not even you. Worst of all, you canât post the adorable video of your cats dressed as the crew of the enterprise. Youâve been the victim of unintentional DoS. You may even get a bill from your ISP for the extra bandwidth.
This chapter will focus on intentional DoS â a denial-of-service attack. DoS attacks can be launched for a number of reasons; the Estonia case was a sort of protest but they can be used to damage competitors for financial gain. In 2004, businessman Saad âJayâ Echouafni allegedly hired computer hackers to launch a DoS attack on three of his competitors. Another application of DoS attacks for financial gain is extortion. A company receives a threat that they will be subjected to a DoS attack unless they wire money to an offshore account. In many cases, the company will simply pay. In 2004, Carnegie Mellon University surveyed 100 companies. They found that 17% of medium-size businesses had been the target of some form of cyber-extortion.C
HOW DENIAL OF SERVICE WORKS
DoS requires two elements: a resource of finite capacity, and the means to acquire or âuse upâ the resource faster than it can be replenished. Although we generally think of these attacks in terms of computers, DoS attacks do not have to be network-based. It is possible to have âreal-worldâ DoS attacks, provided you have the above two elements.
Real-world examples include the practice of âland blockingâ where a company purchases the land around a store to prevent competitors from opening nearby, and many of the methods used by DeBeers to control the diamond market in the twentieth century.D These are examples of a single entity that is powerful enough to consume enough of the available resources (land, diamond production, or any resource your competition needs) to disrupt or starve others. In general, this requires deep pockets or significant resources, which makes it much less likely to occur than DoS attacks in the virtual world.
Distributed Denial of Service
To conduct a successful DoS attack, you usually need a lot of help. This is the origin of the distributed denial of service attack (DDoS). As an example of a DDoS attack for the physical world, consider the following: you admire the cool-headed reasoned approach of Captain Picard over the random cowboy style of Captain Kirk, and decide that what the world needs is a statue of Picard in San Francisco, the (future) home of Starfleet. To this end, you start a campaign to raise money, and people begin sending you checks.
Sadly your âfriendâ Mike does not agree and makes his mission to stop you. He recruits friends to send you hate mail, and soon your mailbox is stuffed with angry letters about why Kirk is better than Picard. Sorting through the mail takes longer and longer, and you only find a few checks in every batch of letters. Soon youâve got friends involved to sort the mail, but sending the same angry letter multiple times is easier (and cheaper) than sending a new check, so the volume of hate mail far outstrips the volume of checks. You need more and more time to sort the mail, for fewer payoffs.
Everyday your mailbox is stuffed full. The post office begins to hold your mail because it cannot deliver it. Now you must drive to the post office to collect boxes of (mostly) photocopied hate mail, and you have to open every letter because you canât easily tell which ones contain checks. Ultimately you may have to abandon your quest, noble though it may be.
Although a DDoS attack is hard to pull off in the physical world, DDoS is the most common â and disturbing â form of DoS attack in the virtual one. Utilizing DDoS techniques and the Internet, small groups (or even a single individual) can conduct massive DoS attacks. The rest of this chapter will focus solely on DDoS attacks.
Overview of a Denial of Service Attack
Suppose you want to conduct a network DDoS attack against a particular victim. You are taking the role commonly called the intruder. To conduct a DDoS attack, you need to be able to âuse upâ some resource needed by the victim. You can target any resource likely to interrupt your target. For instance, if you are targeting an online retailer, you might do any of the following.
⢠Overload the Web servers.
⢠Overload any network link.
⢠Crash servers.
⢠Attack a dependency.
In the last case you donât attack the victim directly, but you might attack their bank, their credit card clearing company, their accounting firm, or one of their suppliers, making it impossible for them to conduct business.
It is likely that your target, especially if it is a bank or online retailer, has a lot of capacity you must use up before you are successful. Your computer just cannot generate enough network traffic to launch a successful DDoS attack. You need help.
In some cases, it may be possible to use social engineering to accomplish your goal. For example, if you can start a successful Internet rumor about your victim, you may be able to get others to do your work for you. For example, you might post a long and official-looking ânewsâ story about how your âfriendâ Mike was caught raising Dalmatian puppies for their fur. Soon activists are calling him at all hours, filling his mailbox with hate mail, and even stopping by his house. Mike can no longer go to the store, let alone continue his DDoS attack against you. Sadly, his friends may carry on in his absence.
Most often you will instead attempt to gain control of a large number of computers from which to conduct the attack. During the 1990s, you might have targeted universities because they had large numbers of always-connected machines with fast connections and (typically) low security. Today, university networks are better protected and monitored, but the rapid growth of the Internet means you can find a large number of always-connected machines with reasonably fast connections and low security in peoplesâ homes and small businesses. If you can compromise enou...