Viruses, Worms, Zombies, and other Beasties (Based on Susan Whittemore, Sanjeev Arora, Alex Halderman, and Steve Shenfield)

The Threat Landscape Impact: Costly regulatory inquiries and penalties, consumer and shareholder lawsuits, loss of consumer confidence Motivation: Financial gain 2017 Outlook: Cyber-extortion will continue to rise Impact: Competitive advantage, trade secret disclosure, operational disruption, brand and reputation Motivation: Personal advantage, monetary gain, professional revenge, patriotism 2017 Outlook: More organizations will implement insider threat mitigation programs and processes Impact: Disruption of business activities, brand and reputation, loss of consumer confidence Motivation: Negatively impact reputation, drive attention to a cause, pressure for change 2017 Outlook: Expected to escalate attack methods with high-profile data breaches Impact: loss of competitive advantage, disruption to critical infrastructure Motivation: Economic, political, and/or military advantage 2017 Outlook: Will continue to strengthen their defensive and offensive cyber skills

Cybersecurity The cost and risks of cyber attacks are increasing Cyber Threat Landscape Cybersecurity events and costs are increasing: 79% of survey respondents detected a security incident in the past 12 months1 Average total cost of a data breach increased 23% over the past two years2 Average cost paid for each lost / stolen record increased 6%1 Industry Outlook Data breaches are expected to reach $2.1 trillion globally by 20193 76% of survey respondents1 were more concerned about cybersecurity threats than in previous 12 months: Increase from 59% in 2014 Reputational Risk An IT security breach can have serious implications in how a company is perceived: 46% of companies suffered damage to reputation & brand value due to a security breach4 19% of companies suffered damage to reputation & brand value due to a third-party security breach or IT system failure4 The risk of losing customer trust is significant and rising: 82% of customers would consider leaving an institution that suffered a data breach5 Source: 1U.S. State of Cybercrime Survey, 2Ponemon Institute, 3Juniper Research , 4Forbes

1 million cybersecurity job openings globally in 2016* 4 *Cisco, 2016; 1Information System Security Certification Consortium, 2015;PricewaterhouseCoopers, 2015; Burning Glass, 2015; ISC2, 2015; Cybersecurity Ventures, 2016

Encryption Encryption strongly protects data en route You Amazon.com Encryption strongly protects data en route Today’s story: Attacker can compromise your computer without breaking encryption.

Encrypted ≠ Secure You Amazon.com Break into your computer and “sniff” keystrokes as you type

Spoofing Attacks Attacker impersonates the merchant (“spoofing”) Amaz0n.com’s key Amaz0n.com You Emails to update your account information (ebay, bank, …) Attacker impersonates the merchant (“spoofing”) Your data is encrypted… …all the way to the bad guy!

Breaking into a Computer What does it mean? How is it done? Can we prevent it?

What’s at Stake? Other fears: cybercrime, terrorism, etc. Kinds of damage caused by insecurity Nuisance: spam, … Data erased, corrupted, or held hostage Valuable information stolen (credit card numbers, trade secrets, etc.) Services made unavailable (email and web site outages, lost business) Hostage (break into your computer and encrypt files) Information = $ Bring down a web site like Amazon or airline - cost them $ Other fears: cybercrime, terrorism, etc.

Main themes of today’s lecture Self-reproducing programs: viruses, worms, zombies Other threats to computer security Internet = Today’s Wild West There is no silver bullet against cyber crime, but follow good security practices Wild west - very little policing of internet -> citizens have to look out for themselves Little police results from architecture of net that also leads to little censure (so there is an upside).

Breaking into a Computer What? Run unauthorized software How? Trick the user into running bad software (“social engineering”) Exploit software bugs to run bad software without the user’s help Social engineering

Example of “social engineering”: Trojan Horse CoolScreenSaver.exe

Viruses and Worms (Recall self-replicating programs: Automated ways of breaking in; Use self-replicating programs (Recall self-replicating programs: Print the following line twice, the second time in quotes. “Print the following line twice, the second time in quotes.” )

Computer Viruses Must fool users into opening the infected file Self-replicating programs that spread by infecting other programs or data files Cool Screen Saver Notepad Solitaire Paint Payload Payload Payload Payload Explain that word files can contain programs (macros) and probably good idea not to run them. Word virus that Hooks into open and save dialog Check see if this computer is infected If not, add itself to MS master template When you save file it checks to see if file contains virus and if not add it So this spreads but doesn’t do anything bad Could check to see if date is after X and if so delete all files on disk (time bomb) Must fool users into opening the infected file

Email Viruses Infected program, screen saver, or Word document launches virus when opened Use social engineering to entice you to open the virus attachment Self-spreading: after you open it, automatically emails copies to everyone in your address book Other forms of social engineering: downloadable software/games, P2P software, etc. Examples of social engineering -- urgent title, looks like it’s from your friend

David L. Smith Aberdeen, NJ The Melissa Virus (1999) Social engineering: Email says attachment contains porn site passwords Self-spreading: Random 50 people from address book Traffic forced shutdown of many email servers $80 million damage 20 months and $5000 fine Intel, lucent, microsoft -- all shut down corporate email for a while while they cleaned up Deal with FBI to work with them reduced sentence David L. Smith Aberdeen, NJ

Computer Worms Self-replicating programs like viruses, except exploit security holes in OS (e.g., bugs in networking software) to spread on their own without human intervention Payload Payload Payload Payload Payload Payload Payload

“Can we just develop software to detect a virus/worm?” [Adleman’88] This task is undecidable. (so no software can work with 100% guarantee) Current methods: (i) Look for snippets of known virus programs on harddrive (ii) maintain log of activities such as network requests, read/writes to hard-drive and look for “suspicious” trends (iii) look for changes to OS code. No real guarantee

A losing battle? Constant battle between attackers and defenders Example: Anti-virus software finds “signature” of known virus Attacker response: Polymorphic virus – to thwart detection, change code when reproduced Anti-virus software adapts to find some kinds of polymorphism But an infinite number of ways to permute viruses available to attackers Halting problem

Example of how worms spread: Buffer Overflow bug From: COS 116 Staff Subject: Welcome Students! Return address Space reserved for email subject Memory … W e l c o m e S t u d e n t s ! 1 2 6  memory address: 100000 Buffer overflow bug: Programmer forgot to insert check for whether email subject is too big to fit in memory “buffer” From: Bad Guy Subject: <evil code . . . . . . . . . . . . . . . . . >100000 … < e v i l c o d e . . . . . . . . . . . . . > 1 1 2 6

The Morris Worm (1988) First Internet worm Created by student at Cornell Exploited holes in email servers, other programs Infected ~10% of the net Spawned multiple copies, crippling infected servers Sentenced to 3 years probation, $10,000 fine, 400 hours community service Wanted to measure the size of internet Robert Tappan Morris

The Slammer Worm (2003) Fastest spreading worm to date Only 376 bytes—Exploited buffer overflow in Microsoft database server products Spread by sending infection packets to random servers as fast as possible, hundreds per second Infected 90% of vulnerable systems within 10 minutes! 200,000 servers No destructive payload, but packet volume shut down large portions of the Internet for hours 911 systems, airlines, ATMs — $1 billion damage! Patch already available months previously, but not widely installed Gridlock Author never caught

Why do people write worms and viruses? Sometimes because they are curious / misfits / anarchists / bored… Once was glory Now increasingly for profit Possibly related to military goals Botnets…

Main reason: Botnets Virus/worm payload: Install bot program on target computer Bot makes target a zombie, remotely controlled by attacker Many zombies harnessed into armies called botnets – often 100,000s of PCs

Zombies Attacker’s Program Bot Bot program runs silently in the background, awaiting instructions from the attacker

Why go to the trouble of creating a botnet? Can rent these out for $ Can infect them with virus Can cover your tracks for other activity

Reason 1: DDOS Attacks “Distributed Denial of Service” Objective: Overwhelm target site with traffic. Example: Wikileaks incidents 2010 “Attack www.store.com” Extortion -- if you don’t give me $ I’ll launch this attack

Reason 2: Sending Spam Messages are hard to filter because there are thousands of senders “Forward this message: Subject: Viagra! …”

Other reasons Click fraud. Commit other cybercrime that is hard to trace

Storm Botnet Created via email scam in 2007 spread to a million computers Owners unknown (believed to be Russian) Used for DoS and Email spams, available for “rent” Fiendishly clever design distributed control, similar to Kazaa, Gnutella rapidly morphing code; morphs every hour or so seems to detect attempts to track/contain it and “punishes” its pursuers

And if you weren’t scared enough already…

Spyware/Adware Hidden but not self-replicating Tracks web activity for marketing, shows popup ads, etc. Usually written by businesses: Legal gray area Download kazaa the installer puts some extra stuff on your computer

International warfare by other means Stuxnet: Computer worm allegedly created by US and Israeli intelligence to target Iranian nuclear processing faciltiies.

Attackers are Adaptive Defenders must continually adapt to keep up

Can we stop computer crime? Probably not! Wild West nature of the Internet Software will always have bugs Rapid exponential spread of attacks But we can take steps to reduce risks… Theory: halting problem Practice: 1 security bug / 20k lines of code (high quality) 40 million lines == 2k bugs (handful corrected / month)

Protecting Your Computer Six easy things you can do… Keep your software up-to-date Use safe programs to surf the ‘net Run anti-virus and anti-spyware regularly Add an external firewall Back up your data Learn to be “street smart” online

Learn Online “Street Smarts” Be aware of your surroundings Is the web site being spoofed? Don’t accept candy from strangers How do you know an attachment or download isn’t a virus, Trojan, or spyware? Don’t believe everything you read Email may contain viruses or phishing attack – remember, bad guys can forge email from your friends

First Line of Defense For Users For Businesses Install system security mechanisms Protect yourself from being a zombie For Businesses Security companies can guard a client’s network ex) Prolexis Technologies

System Security Mechanisms Firewalls Switches & Routers Blackholing Sinkholing Clean Pipes Intrusion Prevention Systems (IPS)

Defenses Firewalls Pros Will prevent simple flood attacks ex) SYN flood Able to allow or deny protocols, ports, or IP addresses Cons Unable to prevent more complex attacks

Defenses Switches & Routers Pros Both have the ability to limit data rate Both have network Access Control Lists ACLs are custom router filters Able to filter both inbound and outbound traffic Cons Most can be easily overwhelmed

Routing Technique 1 Blackholing Attempts to mitigate the impact of an attack Redirects traffic from attacked DNS or IP address to a “black hole” Then all traffic will be dropped Must know IP address of attacker or else legitimate traffic will be dropped as well

Routing Technique 2 Sinkholing Routes suspicious traffic to a valid IP address where it can be analyzed Capturing traffic and analyzing it can be done with a sniffer Traffic found to be malicious is rejected Cons Unable to react to severe attacks as effectively as blackholing

Defenses Continue Clean Pipes Best used when deployed inside Internet Service Providers (ISPs) When an attack occurs, traffic is diverted to a cleaning center in the ISP Here the traffic is “cleaned” by specialized filtering devices and malicious activity is removed Only legitimate traffic is passed to the destination

A Final Defense Intrusion Prevention System(IPS) Monitors network traffic for malicious activity Scans both inbound and outbound Searches for suspicious patterns known as signatures or rules System logs malicious activity and will attempt to stop it

Sources http://cisco.com/web/about/ac123/ac147/archived_issued/ipj_7 -4/dos_attacks.html http://docs.google.com/viewer?a=v&q=cache:Gs5vmKHFfpUJ:pat hmaker.biz/whitepapers/CSISurvey2009.pdf http://i.cmpnet.com/v2.gocsi.com/pdf/CSISurvey2007.pdf www.tik.ee.ethz.ch/~ddosvax/talks/ddos_td.pdf http://en.wikipedia.org/wiki/Denial-of-service_attack http://www.csoroundtable.org/knowledge/there-business-case- it-security http://en.wikipedia.org/wiki/Intrusion_prevention_system http://csdl2.computer.org/comp/mags/ic/2009/06/mic20090600 10.pdf