Independent Submission D. Lazanski Internet Draft Last Press Label Intended status: Informational March 8, 2020 Expires: September 9, 2020 A User-Focused Internet Threat Model draft-lazanski-users-threat-model-t-00 Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html This Internet-Draft will expire on September 9, 2020. Copyright Notice Copyright (c) 2020 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Lazanski Expires September 9, 2020 [Page 1] Internet-Draft A User-Focused Internet Threat Model March 2020 Abstract RFC 3552 introduces a threat model that does not include endpoint security. RFC 3552 is 15 years old and in those 15 years the threat landscape has changed. Security issues and cyber attacks have increased and there are more devices, users, and applications on the endpoint than ever. This draft proposes a new approach to the Internet threat model which will include endpoint security, focus on users and provide an update to the threat model in RC 3552. Table of Contents 1. Introduction...................................................2 2. A History of Data Breaches.....................................3 3. Botnets........................................................5 4. Emerging Threats...............................................6 5. Impacts........................................................7 6. Guidelines.....................................................7 7. A New Internet Threat Model....................................7 8. Way Forward....................................................8 9. Security Considerations........................................9 10. IANA Considerations...........................................9 11. Conclusions...................................................9 12. References....................................................9 12.1. Informative References...................................9 13. Acknowledgments..............................................11 1. Introduction Data breaches are on the rise: personal data is stolen and often leaked or sold on a never-before-seen scale. Malware and ransomware attacks impact the most vulnerable in our global societies today. Better security results in better privacy through prevention of these breaches. However, even though the IETF is privacy-focused, policy and design decisions taken by the IETF have radically changed the architecture of the Internet, arguably without due consideration to cyber defence implications or outcomes. In recent years, this has obsoleted many systems, technologies and programmes which use Internet protocols for prevention, detection and mitigation of cyber attacks. RFC 7258 established that "Pervasive Monitoring" is an attack on privacy that needs to be mitigated where possible. Furthermore, RFC 3552 assumes that the endpoints involved in a communications exchange have not been compromised, but that the attacker has near complete control over the network between the endpoints rather than the endpoints themselves. These assumptions have led to a focus on communications Lazanski Expires September 9, 2020 [Page 2] Internet-Draft A User-Focused Internet Threat Model March 2020 security and the development of protocols that place this kind of security above all else. Ironically - or coincidentally - as the development of these protocols have taken place over the last several decades, there has been and continues to be a sharp rise in cyber attacks. The Internet threat model in RFC 3552 does not even mention that the greatest threat to the Internet is the growing scale and variety of cyber attacks against all types of endpoints that is resulting in significant data breaches. This now needs to change. 2. A History of Data Breaches A data breach is an incident where data is inadvertently exposed in a vulnerable system, usually due to insufficient access controls or security weaknesses in the software.[1] In the first six months of 2018 alone, Gemalto reported that there were 945 data breaches resulting in 4.5 billion records being compromised.[2] This section describes some recent cyber attacks on the Internet that led to data breaches. In October 2013, Adobe announced that hackers had stolen nearly 3 million encrypted customer credit card details and the IDs and encrypted passwords of 35 million customers.[3] In December 2013, the retailer Target announced that 40 million credit card records and personal details for a further 70 million customers had been compromised. A report from Verizon indicated that after one week, 86percent of passwords used by Target had been cracked and Verizon's security consultants were able to move about with complete freedom on Target's internal network.[4] In May 2014, eBay notified 145 million users to change their passwords following a cyber attack that compromised encrypted passwords, customer names, email addresses, mailing addresses, phone numbers and dates of birth.[5] In July 2015, a commercial website that enabled extramarital affairs (called Ashley Madison) was breached; a month later, more than 25GB of company data, including user details, was leaked. The ethics and impact on human rights of this breach are particularly notable, as it resulted in at least one confirmed suicide.[6] In 2016, Uber was breached, giving hackers access to the names, email addresses and phone numbers of 57 million riders and drivers. 600,000 US drivers had their names and license plate numbers stolen. The current assessment is that other personal data, including trip location history, credit card details, social security numbers and dates of birth were not downloaded. [7] Also, in August of 2016, Lazanski Expires September 9, 2020 [Page 3] Internet-Draft A User-Focused Internet Threat Model March 2020 Dropbox was hacked to release over 68million user email addresses and passwords onto the Internet. [8] In March 2018, as part of a coding review, Google uncovered a coding glitch that potentially exposed the personal data of up to 500,000 Google Plus users, including their names, email addresses, occupations, genders and ages.[9] Google could not confirm which users were affected by the security flaw as they keep API log data for only two weeks (and, presumably, log data analysis was lacking or insufficient to detect the breach as it was happening). In May 2018, Twitter advised all 330 million of its users to change their passwords after a software exposed them in plaintext. [10] Additionally, in September 2018, British Airways announced that personal and financial details of up to 380,000 customers who had booked flights over a 16-day period had been stolen. This breach was traced to a rogue script that had been installed on the third-party payment supplier used by British Airways.[11] Also in September 2018, Facebook suffered its worst security breach ever; the exploitation of several simultaneous software bugs gave login access to as many as 50 million accounts.[12] April 2019, a 146GB data set containing over 540 million Facebook records were found exposed on AWS servers, as two third-party companies had collected Facebook data on their own servers.[13] In November 2018, 500 million Marriott International customers had their details stolen in an ongoing breach since 2014. Approximately 327 million hotel guests had a combination of name, address, phone number, email address, passport number, date of birth, gender and arrival/departure information stolen.[14] In January 2019, the personal data of more than 3500 people living with HIV in Singapore was leaked in Singapore, allegedly by an insider with access to sensitive records.[15] Also in February 2019, a file containing 2.2 billion compromised usernames and passwords was found on the dark web. This 600GB file was a collation of previous data breaches, truly demonstrating the scale and severity of the data breach and cyber defence problem in totality.[16] And these are only a handful of breaches that have been made public. So many more go unreported in the public. Data breaches are one of the singular most important issue in cybersecurity today. IBM's 13th "Cost of a Data Breach" study found that the global average cost of a data breach in 2018 was $3.86 million.[17] That is the average cost of one - not many - data breaches. It is unthinkable and unrealistic that any revised Internet threat model does not highlight the large and ongoing threat from data Lazanski Expires September 9, 2020 [Page 4] Internet-Draft A User-Focused Internet Threat Model March 2020 breaches, whatever their cause. Threat actors are making full use of the Internet technology that allows them to hide on endpoints and perform such large data hacks that mostly go undetected. Internet security researchers and developers must accept the reality of all the security issues in the Internet ecosystem. Decisions being made in the name of privacy are sometimes leading to larger inadvertent security and privacy losses. 3. Botnets A botnet is a string of connected computers used, in this case, to perform a malicious function against an end user, organisation or series of users.[18] Though computers working together to increase computing power for functions does not constitute a botnet in itself (and is used often in data centres for chat rooms or email services, for example) botnets are a specifically used for malicious intent. There have been a number of recent, high profile botnet attacks and only a few will be described here as examples. In 2000, EarthLink Spammer sent 1.25 million phishing emails over a year and made $3 million in profits by using fake websites and domain names to accomplish this. Subsequently the spammer was convicted and Earthlink won $25 million in damages.[19] Created in 2007, Cutwail was the biggest botnet on the Internet by 2009 by number of infected computers or hosts sending email. It was sending 51 million emails every minute.[20] By 2010, however, it started a DDoS attack to nearly 300 major sites including PayPal and US federal agencies. By 2013 it was the botnet to use for sending spam, but over time its use declined through targeted attempts to take it offline as well as the expiration of email addresses. Though not as popular and sending far less than it once did, Cutwail still sends spam to this day.[21] A more recent botnet was the centre of one of the biggest outages of the Internet network. The Mirai botnet was first identified in 2016. The Mirai botnet as well as variants infect Internet of things devices and those infected devices scan the Internet for IP addresses of other Internet of Things devices, thus creating a multiplication of IoT devices which are infected. Though the bot still exists in various forms, the most serious attack took place on 21 October 2016 when the Domain Name System (DNS) provider Dyn was attacked by DDoS using a coordinated system of infected IoT devices. Much of the Internet was unreachable after three attacks occurred during the day. Though eventually resolved on that day, the sheer size and scale of the attack is still viewed as one of the biggest attacks on the Internet to this day.[22] Lazanski Expires September 9, 2020 [Page 5] Internet-Draft A User-Focused Internet Threat Model March 2020 According to Kaspersky Labs, there were just over 15,000 botnet attacks in 2018.[23] Worryingly, of those attacks, approximately 40 percent were new in both type and the target. Again, as IoT devices increase and as networks expand coverage and ability to handle even more devices and data, it is likely that botnet attacks willcontinue to be seen on such a scale. 4. Emerging Threats Older methods of cyber attacks are still happening and causing breaches, as endpoint security remains incomplete and not up to date. Servers remain unpatched and vulnerable and client devices become legacy or unsupported, to name just a few issues. In parallel, new categories of attacks are emerging. Software updates are a new attacked vector. In March 2019, Kaspersky uncovered the ShadowHammer supply-chain attack which injected malicious code into the ASUS Live Update Utility. This attack involved signing malicious code using stolen certificates and was estimated to have affected half a million users.[24] As a result of the ShadowHammer attack, public focus turned to how and what could be the point of infection. Suggestions were that the IP addresses could have been the point of origin of the attack while others suggested that the malware itself was dormant and inactive until a certain update triggered the malware. In July 2019, Godlua became the first publicly known malware to use DNS-over-HTTPS to avoid DNS-based malware protection security systems. [25] The malware uses DoH requests to determine where the active URL originates and where it will connect. The malware takes advantage of this information in order to initiate a DDoS attacks. The malware attacks both windows and linux systems and takes advantage of a backdoor exploit. [26] Attacks on individual consumers have dropped by nearly 40 percent, due to the fact that attacking one person is largely not financially viable, but attacks on business organisations have increased year on year.[27] Ransomware is on the rise, motivated by economic gain and the weaknesses in endpoints. Malware is freely available and the vulnerable attack point of an endpoint can be found. Botnets are increasing in size and scale as well as ease of use. There are other emerging threats that require more research to collate fully and this section is a starting point. Lazanski Expires September 9, 2020 [Page 6] Internet-Draft A User-Focused Internet Threat Model March 2020 5. Impacts As noted in draft-arkko-farrell-arch-model-t-02 there is a difference between user interaction endpoints and system endpoints. Acknowledging that the end-to-end model supports permissionless innovation, it is imperative to ensure that the open and innovative nature of the Internet continues. However, a taxonomy of endpoints and agreement on those which have had the most security impact in the last 15 years in necessary to continue this work. Work is underway in to attempt to catalogue the most well-known threats and considerations to be taken for protocol designers in light of these threats. This is a taxonomy of a sort along with basic guidelines and information. 6. Guidelines TODO: insert list of guidelines for protocol designers. 7. A New Internet Threat Model Many endpoints are vulnerable; CLESS begins a much needed research programme to demonstrate what capabilities and what limitations can be expected at the endpoint and from a variety of types of endpoints.[28] Endpoints have changed over the last 10 years, but assumptions about endpoints in the IETF hasn't changed in that time. Draft-iab-for-the-users-03 discusses that end users are beneficiaries of the IETF standards. End users use endpoints which have new and emerging threats. Even the user is not often in full control of what happens on their endpoint and what security protections apply to their own data a model where the Internet is user-centric would give more control to the user. The user is both the home Internet citizen and the organisation administrator seeking to protect against data breaches; both need the power to control where their data goes and choose their security protections. So while endpoints are the focus now, does the Internet need to be user-centric in the future? Won't that give users even more assure privacy? ATTACK versions of methods, when categorised by type, show that endpoint methods of compromise are increasing faster than network attacks.[29][30] This may be due to more variety in endpoints, substandard security in many endpoints or the difficulty of attacking a network compared to an endpoint. Whatever the reason, the logical conclusion is that the current Internet design is not stopping cyber attacks. Perhaps a fresh approach is required. Lazanski Expires September 9, 2020 [Page 7] Internet-Draft A User-Focused Internet Threat Model March 2020 As more power and control has shifted to endpoints - and even to only a select few applications on endpoints network defences can protect fewer and fewer endpoints; concurrently, attacks have increased and attacks have increased. The diagram above shows the proliferation of attacks on end points increase over a 3 and a half year timescale while network an physical attacks remain largely unchanged. Whether this is correlation or causation requires thorough research, essential to changing the existing threat model approach from its current approach. The existing Internet Threat Model of RFC3552 makes the general assumption that end-systems have not been compromised and that while end-systems are difficult to protect against compromise, protocol design can help minimise the damage.[31] Revisiting this general assumption in the light of the magnitude of an increase in data breaches and their subsequent negative results is a good starting point for a new Threat Model which can result in protocol design that helps mitigate end-system compromise. 8. Way Forward Additional Internet drafts on the same topic were published include this one and draft-arkko-farrell-arch-model-t-02. The logical way forward would be to this draft and the concurrent other two. RFC 3552 will need to be revised in light of the development of the threat landscape that has changed and grown in the 15 years since RFC 3552 was published. This draft highlights a selection of attacks and data breaches over the last decade and a half. A revision to RFC 3552 would need to include all known and potential attack surfaces taking into account mobile network development, new and emerging devices which are connected to the Internet and the proliferation of users, devices and applications on and over the Internet, as mentioned above. Next steps would be to identify the threat landscape and the attack vectors that have appeared in the 15 years and are well known since RFC 3552. Though threats have focused on the network side of the security landscape, a combined Internet draft should include an identification of different threats, definitions and classifications, examples, assumptions about the endpoints and an update to this Internet draft and RFC 3552 as well. Such a combined draft would supersede RFC 3552 and be more representative of the current state of play of the threat landscape. Lazanski Expires September 9, 2020 [Page 8] Internet-Draft A User-Focused Internet Threat Model March 2020 9. Security Considerations This document proposes a new way of thinking about developing Internet security protocols and does not create, extend or modify any protocols. The intent is to continue discussion and bring in a cyber defence viewpoint. 10. IANA Considerations Upon publication this document has no required actions for IANA. 11. Conclusions The Threat Model indeed needs revisiting and changing, because cyber defence threats and attacks are increasing, yet the responsibility to help mitigate these threats and attacks is largely unrecognised in the IETF community. These threats and attacks should be given the attention they deserve and a way forward is proposed. Further, it is imperative that new conclusions and recommendations from a revisited threat model are backed up by research, case studies and experience, rather than bold assertions. Research and evidence is important to achieve effective security, unsubstantiated guesswork is not. This draft continues to highlight the importance that any threat model must be based in evidence about data breaches. This draft continues the discussion which focuses on the user, identifies the current threats and proposes mitigation of those threats. 12. References 12.1. Informative References [1]https://haveibeenpwned.com/FAQs/ [2]https://www.cbronline.com/news/global-data-breaches-2018 [3]https://krebsonsecurity.com/2013/10/adobe-to-announce-source- code-customer-data-breach/ [4]https://krebsonsecurity.com/2015/09/inside-target-corp-days- after-2013-breach/ [5]https://www.businessinsider.com/cyber-thieves-took-data-on-145- million-ebay-customers-by-hacking-3-corporate-employees-2014-5 Lazanski Expires September 9, 2020 [Page 9] Internet-Draft A User-Focused Internet Threat Model March 2020 [6]See https://digitalguardian.com/blog/timeline-ashley-madison- hackfor a timeline of the breach. [7]https://us.norton.com/internetsecurity-emerging-threats-uber- breach-57-million.html [8]https://www.theguardian.com/technology/2016/aug/31/dropbox-hack- passwords-68m-data-breach [9]https://www.experian.com/blogs/ask-experian/google-data-breach- what-you-need-to-know/ [10]https://www.theverge.com/2018/5/3/17316684/twitter-password-bug- security-flaw-exposed-change-now [11] https://medium.com/asecuritysite-when-bob-met-alice/the- british-airways-hack-javascript-weakness-pin-pointed-through-time- lining-dd0c2dbc7b50 [12]https://www.nytimes.com/2018/09/28/technology/facebook-hack- data-breach.html [13]https://www.databreachtoday.co.uk/millions-facebook-records- found-unsecured-on-aws-a-12337 [14]https://www.bbc.co.uk/news/technology-46401890 [15]https://www.straitstimes.com/singapore/2400-singaporeans- affected-by-data-leak-contacted-by-moh [16] https://mobilesyrup.com/2019/01/31/collection-2-data-breach- 600gb-leaked-emails-passwords/ [17]https://securitytoday.com/articles/2018/07/17/the-average-cost- of-a-data-breach.aspx [18]https://us.norton.com/internetsecurity-malware-what-is-a- botnet.html [19]https://www.bizjournals.com/atlanta/stories/2002/07/22/story4.ht ml [20]https://www.whiteops.com/blog/9-of-the-most-notable-botnets [21]https://www.wired.co.uk/article/infoporn-rise-and-fall-of-uks- biggest-spammer Lazanski Expires September 9, 2020 [Page 10] Internet-Draft A User-Focused Internet Threat Model March 2020 [22]https://www.theverge.com/2016/10/21/13362354/dyn-dns-ddos- attack-cause-outage-status-explained [23]https://securelist.com/bots-and-botnets-in-2018/90091/ [24]https://www.vice.com/en_us/article/pan9wn/hackers-hijacked-asus- software-updates-to-install-backdoors-on-thousands-of-computers 13. Acknowledgments This document was prepared using 2-Word-v2.0.template.dot. A previous version of this draft was posted in July 2019 with a name of draft-lazanski-users-internet-model-t-00 - this version is intended to update that document with further information. Authors' Addresses Dominique Lazanski Last Press Label London United Kingdom Email: dml@lastpresslabel.com Lazanski Expires September 9, 2020 [Page 11]