I’ve lost track of the number of times I’ve been asked this question over the last 12 months. Everyone from CIO’s of large organisations through small startups and entrepreneurs, academics and even family members has asked me this when I tell them what I do. Not surprisingly it gets asked a lot more when hacking is on the 10 o’clock news as it has been a number of times over the last year or so with attacks on companies like TalkTalk, iCloud, Fiat Chrysler and, most infamously, Ashley Madison.
I’ve decided therefore to research the facts around cloud and security and even if I cannot come up with the definitive answer (the traditional answer from an architect about any hard question like this usually being “it depends”) at least point people who ask it to somewhere they can find out more information and hopefully be more informed. That is the purpose of this post.
First of all it helps to clarify what we mean by “the Cloud” or at least cloud computing. Let’s turn to a fairly definitive source on this, namely the definition given in the National Institute of Standards and Technology (NIST) Definition of Cloud Computing. According to the official NIST definition:
“Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.”
Note that that this definition makes no statement about who the cloud service provider actually is. This definition allows for clouds to be completely on premise (that is, within a companies own data centre) and managed by companies whose business is not primarily that of IT just as much as it could be the big ‘public’ cloud service providers such as Microsoft, IBM, Amazon and Google to name but four. As long as there is network access and resources can be rapidly provisioned then it is a cloud as far as NIST is concerned. Of course I suspect the subtleties around this are lost when most people ask questions about security and the cloud. What they are really asking is “is it safe to store my data out on the internet” to which the answer very much is “it depends”.
So, let’s try to get some hard data on this. The website Hackmageddon tracks cyber attacks around the world and publishes twice monthly statistics on who is being hacked by whom (if known). Taking at random the month of August 2015 there were 79 recorded cyber attacks by Hackmageddon (which as the website points out could well be the tip of a very large iceberg as many companies do not report attacks). Of these there seem to be no attacks that are on systems provided by public cloud service providers but the rub here of course is that it is difficult to know who is actually hosting the site and whether or not they are clouds in the NIST definition of the word.
To take one example from the August 2015 data the UK website Mumsnet suffered both a distributed denial of service (DDoS) attack and a hack where some user data was compromised. Mumsnet is built and hosted by the company DSC a hosting company not a provider of cloud services according to the NIST definition. Again this is probably academic as far as the people affected by this attack are concerned. All they know is their data may have been compromised and the website was temporarily offline during the DDoS attack.
Whilst looking at one month of hacking activity is by no stretch of the imagination representative it does seem that most attacks identified were against private or public companies, that is organisations or individuals that either manage their own servers or use a hosting provider. The fact is that when you give your data away to an organisation you have no real way of knowing where they will be storing that data or how much security that organisation has in place (or even who they are). As this post cites the biggest threat to your privacy can often come from the (mis)practices of small (and even not so small) firms who are not only keeping sensitive client information on their own servers but also moving it onto the cloud, even though some haven’t the foggiest notion of what they’re doing.
As individuals and companies start to think more about storing information out in the cloud they should really be asking how cloud service providers are using people, processes and technology to defend against attackers and keep their data safe. Here are a few things you should ask or try to find out about your cloud service provider before entrusting them with your data.
Let’s start with people. According to IBM’s 2014 Cyber Security Intelligence Index 95% of all security incidents involve human error. These incidents tend to be security attacks from external agents who use “human weakness” in order to lure insiders within organisations to unwittingly provide them with access to sensitive information. A white paper from the data security firm Vormetric says that the impacts of successful security attacks involving insiders are exposure of sensitive data, theft of intellectual property and the introduction of malware. Whilst human weakness can never be completely eradicated (well not until humans themselves are removed from data centres) there are security controls that can be put in place. For example insider threats can be protected against by adopting best practice around:
- User activity monitoring
- Proactive privileged identity management
- Separation-of-duty enforcement
- Implementing background checks
- Conducting security training
- Monitoring suspicious behaviour
Next cloud providers need to have effective processes in place to ensure that the correct governance, controls, compliance and risk management approaches are taken to cloud security. Ideally these processes will have evolved over time and take into account multiple different types of cloud deployments to be as robust as possible. They also need to be continuously evolving. As you would expect there are multiple standards (e.g. ISO 27001, ISO 27018, CSA and PCI) that must be followed and good cloud providers will publish what standards they adhere to as well as how they comply.
Finally what about technology? It’s often been said that internet security is a bit like an arms race where the good guys have to continuously play catch up to make sure they have better weapons and defences than the bad guys. As hacking groups get better organised, better financed and more knowledgable so security technology must be continuously updated to stay ahead of the hackers. At the very least your cloud service provider must:
- Manage Access: Multiple users spanning employees, vendors and partners require quick and safe access to cloud services but at the same time must have the right security privileges and only have access to what they are authorised to see and do.
- Protect Data: Sensitive data must be identified and monitored so developers can find vulnerabilities before attackers do.
- Ensure Visibility: To remain ahead of attackers, security teams must understand security threats happening within cloud services and correlate those events with activity across traditional IT infrastructures.
- Optimize Security Operations: The traditional security operations center (SOC) can no longer operate by building a perimeter firewall to keep out attackers as the cloud by definition must be able to let in outsiders. Modern security practices need to rely on things like big data analytics and threat intelligence capabilities to continuously monitor what is happening and respond quickly and effectively to threats.
Hopefully your cloud service provider will have deployed the right technology to ensure all of the above are adequately dealt with.
So how do we summarise all this and condense the answer into a nice sentence or two that you can say when you find yourself in the dreaded elevator with the CIO of some large company (preferably without saying “it depends”)? How about this:
The cloud is really a data centre that provides network access to a pool of resources in a fast and efficient way. Like any data centre it must ensure that the right people, processes and technology are in place to protect those resources from unauthorised access. When choosing a cloud provider you need to ensure they are fully transparent and publish as much information as they can about all of this so you can decide whether they meet your particular security requirements.
Ping. Floor 11.