Cloud Security with Object Storage

AWS logoMany cloud providers are often criticized for the security provided with object storage services. Even more after the disclosure of private information that occurred in 2017 by using these services. These security breaches were also from well-known organizations such as Verizon, Accenture, Booz Allen Hamilton, Viacom, National Security Agency, National Credit FederationAustralian Broadcasting Corporation, Department of DefenseRepublican National Committee, etc. There are often new organizations to add on this list but they are the main one from the last months. These organizations were mainly using the object storage service S3 from AWS.

Object Storage

This is not a technology only provided by AWS with the S3 service. There are many services provided by other well-recognized cloud providers to store files in the cloud such as Azure, Google Cloud, DigitalOcean, IBM, etc. However, AWS S3 is definitely the object storage service that is the most used by many organizations. The service was also first released in 2006 before other services from competitors. The statistics are a little bit old but as of April 2013, AWS mentioned that S3 has more than 2 trillion objects stored with 1.1 million requests per second. In 2018, it is possible to assume these numbers are even higher.

Amazon S3 is often wrongly targeted by the media. It is simply the most popular service used by many organizations of all size. We have to keep in mind that object storage is only a way to store files, often with a cloud provider but it could also be with a private infrastructure.

IT Administrators

I often read some IT professionals and even information security professionals to have doubts on these services. Mainly doubts on the security measures available to protect the information stored. It is important to understand that security breaches related to object storage are often not related to the underlying technologies. Cloud providers such as AWS, Google and Azure are able to provide secure environment for your files. The configuration for such spaces, or buckets in S3 terms, is secure and private by default. How is it possible in this case to have public files on the Internet?

Simply ask your IT administrators. It is more a question of misconfiguration. In order to authorize a public access to the stored files, someone actually needs to perform a manual action to change the default behaviour. The approach would be different for each service but the principle is the same. It is possible to manage accesses on S3 with rules but other services could be simpler with an option to be set at “Private” or “Public”. This is often a configuration available for the space and/or per file.

Maybe it is the time to review the accesses implemented for your files stored in the cloud? From object storage services like S3 but also on services like Office 365, Dropbox, Google Drive, etc. It is so easy to forget about a file that should not be available for all on Internet.

Third-Party Vendors

Are you aware of your third-party vendors who could use object storage with your information? For example, with Verizon and Republican National Committee, in both situations, third-party vendors were involved i.e. Nice Systems and Deep Root Analytics. Organizations easily trust more and more third-party vendors and share confidential information. This data can be about the organization operations but often on clients. Nevertheless, it is important to evaluate the information sent to external vendors and to understand how this one uses the data.

NIST and the Digital Identity Guidelines

The NIST published last June the final version of the Digital Identity Guidelines also known as SP 800-63. This publication was a draft since 2016 and they even asked for comments from the community on GitHub during the summer 2016. All these comments were inputs for the final publication. Many posts on the Internet mention these changes. But I think it is still important to reiterate them since they are not necessarily well known by everyone who is not in information security.

Who is the NIST?

The National Institute of Standards and Technology (NIST) is a non-regulatory government agency of the US Department of Commerce responsible, among other things, to publish standards for federal agencies. The Special Publications (SP) 800 series are well known to be important guidelines in the information security field for private and public organizations. Worldwide professionals value these publications and they are often used to structure their information security strategies.

New Requirements

Since the past two decades and more, we all saw the result of these requirements. A lack of user experience where most users were often able to circumvent the rules. Many studies have demonstrated these requirements were adding little value on the security side. Furthermore, users were often able to find a way around these requirements thus reducing the security goals. There are mainly 3 requirements updated:

Password complexity rules

You know the rule where you have to put at least a lowercase character, an uppercase character, a number and a special character? It’s not a requirement anymore. Studies shown that users were simply using different patterns to respect this requirement. For example, one trick was to replace some letters by numbers or even simpler, to add an exclamation point at the end. These patterns are all well known by hackers and these passwords were not more secure because of the complexity rule. Oh, all characters should also be available, even emojis!

Password Expiration

This is even something audited during common external audits. That moment at work when you receive a notification and you have to change your password. And this, often every 90 days. No more! We all know users were keeping the same password and adding a character at the end. I was the first one to do it because I always thought it was not efficient. It’s better to have one good password for the service than having a weak one changed every X days. However, it could still be possible to force a user to change a password in certain situations. For example, it should be possible to request users to reset their passwords if the service suspect a compromise. So, it is still important to keep a password history.

Password Hints and Knowledge-Based Questions

Financial services are really good with this requirement, mainly for the knowledge-based questions. This is when questions are also asked with a password to complete the authentication process. The main problem with this is the fact that most answers are now freely available online with social media. For the password hint, I never understood this one. I always saw this one as “let’s just give more clues to hackers on my password” so I never used this one.


The next two are interesting recommendations from the NIST:

Common passwords and usernames

With the new requirements, users are not forced to choose specific characters or to change password. However, the NIST suggests that passwords should be validated against a dictionary of well-known passwords and/or a list compromised passwords. This recommendation makes so much sense. For example, if a user is trying to use, “Test123!” would hopefully fail the validation against the dictionary. There are many dictionaries available online. The logic behind that is the fact that hackers are using these dictionaries to find passwords. The situation is also applicable to common usernames such as “admin” or “root”.

Multi-Factor Authentication (MFA) With SMS

At one point, the NIST completely removed SMS as a valid method to implement with the multi-factor authentication. But with the final release, SMS is still supported, but not necessarily recommended. It is theoretically possible to intercept an SMS. However, it is still more secure to implement MFA with SMS than having no MFA at all. The alternative would be to have an app such as Google Authenticator or a solution with push notifications such as Duo.

October 2017 : Security Breaches

The data security breaches occurred/disclosed in October 2017.


DISQUSThe popular commenting system was breached in 2012. Disqus got notified by Troy Hunt, a security expert, who obtained a copy of the data. According to the company, the data exposed are from 2007 and involve 17.5 million users. Among the user’s information stolen include email addresses, usernames, sign-up dates and last logins. However, about one third or approximately 5.8 million users, also got their passwords in the wild. At least, the passwords were not in clear text but hashed with a salt with the now weak SHA-1 algorithm. They seem to have handled the situation well with a public disclosure in 24 hours and they asked the affected users to reset their password account. They have also mentioned that they are now using the bcrypt algorithm which is now the best practice.

Far Eastern International Bank

Far Eastern International BankA malware infected this Taiwanese bank which instructed the SWIFT terminal to move $60 million into different bank accounts based in Sri Lanka, Cambodia and the United States. SWIFT is the main global banking network where it is possible for banks to exchange funds between them. It is not the first time this situation occurs and a well-know breach occurred in 2016 with a Bangladesk bank where the attempt was to steal $951 million. The Far Eastern International Bank was able to retrieve most funds. Mostly since the breach in 2016, the SWIFT organization has developed a more stringent security requirements for their customers with the Customer Security Programme (CSP) but many banks are still in the process of getting certified.


AccentureThis is another big name in the IT consulting industries. Accenture offers consulting services for the largest organizations and often seen as a leader in cloud consulting services. UpGuard reported that AWS S3 buckets were configured for public access. In total, 4 buckets were available for everyone. These buckets contained confidential API data, customer information, private keys, 40 000 passwords mainly in clear text and even logs from a monitoring solution. One bucket contained more than 137 gigabytes of data.


Remember the data breach that occurred in 2013 at Yahoo? It was first disclosed by the company that someone had access to information on one billion accounts. This number was revised by Verizon, the now parent company of Yahoo, at 3 billion accounts. It was possible to retrieve the usual information such as names, email addresses and hashed passwords. Some hash would still be with the weak MD5 algorithm.

Hyatt Hotels

It was possible to obtain the information from cards manually entered or swiped at the front desk. This situation occurred between March 18, 2017, and July 2, 2017, in 41 properties across 11 countries. As expected, it was possible to get the cardholder name, card number, expiration date and verification code. This is the second security breach for this company.

Pizza Hut

About 60 000 customers might have been impacted by a security breach that would have occurred from the morning October 1, 2017, to midday October 2, 2017. Data including customer names, billing postal code, delivery addresses, email addresses, and payment card information. Pizza Hut notified by email customers impacted only 2 weeks after the situation and they are offering a free credit monitoring service for a year.

South Africa

66 million records were obtained on South African. What, wait, the population is only about 56 million people? The obtained database also included 9 million people with a deceased status. The database was openly available on a web server owned by Jigsaw Holdings and was probably bought from a credit bureau in 2014. Information available include South African ID number, name, gender, age, location, marital status, estimated income, address, phone numbers, employers, etc.

Patient Home Monitoring Corporation

An estimated 150,000 American patient files were available through an unsecured AWS S3 bucket. It is hard to know for how long this bucket was available with public access but Kromtech Security Researchers have discovered the breach on September 29, 2017. 47.5 gigabytes of data with about 316,000 PDF files including mainly blood test results. These documents contained names, addresses, contact information, dates of birth, diagnoses and names of physicians. All this information is strictly regulated by the Health Insurance Portability and Accountability Act (HIPAA).


When you are a major company who is developing software and hardware, you have a central database somewhere to track and document all vulnerabilities related to your products. Of course, this database contains critical information about your products and you probably prefer to keep it secret. Well, this database at Microsoft was hacked in 2013.

iDNS: Scam Going On for More Than 15 Years

iDNS renewal letterYou probably already received one of these letters if you have registered a domain name in the past few years. The company behind these letters is Brandon Gray Internet Services Inc. The worst part is the fact this is a legitimate organization registered and operating in Canada (Markham, Ontario). I thought for a long time it was only a scam here, but I recently discovered they also operate in the United States, Europe and Australia.

Operations Under Many Names

I got my first domain name in 2003 so I don’t exactly remember the first time I received one of these letters. However, I believe it was under the name, “Domain Registry of Canada”. They now seem to use more often iDNS as shown on the image. Over the years, they used many different reseller names under the parent company “Brandon Gray Internet Services” such as:

  • Namejuice
  • Domain Registry of Canada (DROC)
  • Internet Domain Name Services (iDNS)
  • Domain Registry of America (DROA)
  • Domain Renewal Group
  • Liberty Names of America
  • Domain Registry of Europe (DROE)
  • Domain Registry of Australia

Deceptive Message

The main concern is the deceptive message in these letters sent by mail. It is possible since postal addresses are freely available for each domain name with a WHOIS query. There is always the situation where the domain owner is using a privacy protection service but it is not always the case. The main objective is to trick the owner to renew the domain name with them. Nevertheless, this renewal also means the transfer of the domain to the new registrar. A situation that will definitely lead to future problems. The reseller names used can easily mislead the recipient to think they are an official government authority.

The business is totally unethical, but there is a grey zone worth mentioning. I believe the wording was updated throughout the years to be more… compliant with the law. However, it is still a deceptive message and can surely mislead a neophyte in the universe of domain names management.


In my opinion, you have all elements for a perfect scam. The message that would protect them: “This notice is not a bill, it is rather an easy means of payment should you decide to switch your domain name registration to Internet Domain Name Services” and “As a courtesy to domain name holders, we are sending you this notification of the domain name registration that is due to expire in the next few months“. A message to generate fear to the recipient: “Failure to renew your domain name by the expiration date may result in a loss of your online identity making it difficult for your customers and friends to locate you on the Web“. Finally, a possible new opportunity even if it is not true: “Privatization of Domain Registrations and Renewals now allows the consumer the choice of Registrars when initially registering and also when renewing a domain name“.


They have a long history of lawsuits and investigations following various complaints since the beginning with different regulators e.g. Competition Bureau of Canada, Advertising Standards Authority (ASA), Federal Trade Commission, ICANN and the Canadian Internet Registration Authority (CIRA). But also some lawsuits with other companies such as Tucows, and Deinternetman.

What Should You Do?


Their prices are even higher than the competition and you simply don’t want to write down your credit card information on a piece of paper. Be careful to the details. They are not even able to use a well-known TLD such as a .com. They are using a country code top-level domain .as (American Samoa) which is a redirection to the ccTLD .to (Tonga).

Unfortunately, this scam seems to be working since Brandon Gray Internet Services is still in operation and the scam is going on after more than 15 years. You will receive these letters a few times a year if you own more than one domain. The only solution for now is to throw it away. You have nothing to do. You could always complain to some authorities but I personally think it is not worth the time after so many years… However, be sure to still renew your domain name on time with your current registrar. In fact, you should simply activate the auto-renewal offer by most registrars.

Septembre 2017: Brèches de sécurité

This post was published when this blog was also in French. This post is available in English.

Septembre 2017 a été un mois intéressant pour plusieurs brèches importantes de sécurité. Nous avons tous appris la valeur de nos informations personnelles. À partir de maintenant, je vais publier un billet mensuel au sujet des brèches importantes de sécurité du mois précédent.


Equifax est un des plus importants bureaux de crédits et ils ont eu un accès récurrent non autorisé à leurs systèmes du 13 mai au 30 juillet 2017. Les équipes techniques étaient même avisées de la principale vulnérabilité exploitée puisqu’une note a même été distribuée à l’interne le 9 mars afin de corriger celle-ci (Apache Structs CVE-2017-5638). L’équipe de sécurité a détecté la situation uniquement le 29 juillet. Le PDG a appris la situation le 31 juillet et les administrateurs ont obtenu les informations le 24 et 25 aout. Finalement, c’est seulement le 7 septembre que Equifax a divulgué publiquement la brèche de sécurité.

143 millions (143 000 000, oui, avec six zéros) enregistrements sur des citoyens américains ont été obtenus, incluant noms, numéros d’assurance sociale (NAS), dates de naissance, et même certains permis de conduire. Après l’investigation par une firme de sécurité, le nombre final est de 145.5 millions, et inclut maintenant des numéros de cartes de crédit pour environ 209 000 clients. Au Canada, on est un peu plus chanceux puisque le nombre initial était de 100 000 clients, mais après investigation, le nombre révisé est de “seulement” 8 000 clients.

Le PDG a pris une retraite anticipée avec plusieurs exécutifs incluant le Chief Information Officer (CIO) et le Chief Security Officer (CSO). Equifax sera également la cible de plusieurs poursuites judiciaires, autant aux États-Unis qu’au Canada. L’ancien PDG devra même témoigner devant le Congrès américain. Il y a aussi quelques interrogations des régulateurs suite à la vente des actions de certains exécutifs lors de la détection de la brèche de sécurité. Toutefois, ces transactions ont été effectuées avant la divulgation publique de la situation et ces personnes pourraient donc faire face à des accusations pour délits d’initiés.

US Securities and Exchange Commission (SEC)

Cette agence fédérale américaine est principalement responsable de faire respecter les lois sur les valeurs mobilières et de réglementer cette industrie. La Commission a découvert une vulnérabilité applicative en 2016 et a été “rapidement” corrigée. Toutefois, la SEC a divulgué un incident possible puisqu’il aurait possiblement eu un accès non autorisé avant de mettre en place le correctif. Cette fois-ci, aucun accès à des renseignements personnels, mais à des informations sensibles n’étant pas encore publiques sur des compagnies. Une déclaration officielle a été publiée le 21 septembre.


Une des firmes comptables parmi les “Big 4” a aussi été la cible ce mois-ci. La nouvelle a été publiée par le Guardian le 25 septembre. Deloitte est souvent la firme, parmi les “Big 4”, la plus reconnue pour ses services en cybersécurité. Les clients de la firme incluent 80% des organisations du Fortune 500. Il y a eu un accès non autorisé sur le serveur global de courriels de la firme hébergé avec le service infonuagique de Microsoft Azure. Et ce, probablement depuis octobre ou novembre 2016.

Rien de trop compliqué cette fois, les pirates informatiques ont simplement obtenu les informations de connexion concernant un compte administrateur. Par la suite, il était possible de se connecter directement au serveur de courriels ayant un accès aux courriels des 244 000 employés de Deloitte. Plusieurs courriels devaient probablement contenir plusieurs informations sensibles sur leurs clients et même des pièces jointes intéressantes. Le système n’a pas été compromis avec des connaissances techniques avancées, mais bien par des techniques d’ingénierie sociale dans le but d’obtenir les informations de connexion. De plus, sans authentification à deux facteurs (2FA), il était simple de se connecter à distance.


Sonic est une chaine importante de restauration rapide aux États-Unis avec près de 3600 restaurants. Brian Krebs a été le premier à signaler cette brèche de sécurité le 26 septembre. En fait, le processeur de paiement pour les cartes de crédit de la chaine a informé celle-ci des activités inhabituelles en lien avec leurs transactions. Le détail de cette brèche de sécurité n’est pas encore connu. Toutefois, il a été possible de retrouver au moins 5 millions de comptes de cartes de crédit et débit en vente en ligne. Ces comptes sont fort probablement en lien avec la brèche chez Sonic.

Whole Foods Market

Whole Foods Market, qui a été acheté par Amazon, a aussi divulgué le 28 septembre que des informations sur les cartes de paiement ont été volées. L’investigation est toujours en cours et on devrait en savoir plus bientôt. Il y a un point très intéressant qui a été mentionné dans le communiqué de presse et c’est le fait que les systèmes en lien avec ne sont pas connectés à ceux de Whole Foods. J’espère bien pour eux, mais bon, ça valait la peine de préciser cette information.