Just after the New Year, I caught up with a work colleague of mine and I started to chat about all the good work we are doing in Dell with regards Zero Trust and the broader Zero Trust Architecture (ZTA) space. Clearly he was very interested (Of course!!). We talked about the Dell collaboration with MISI (Maryland Innovation Security Institute) and CyberPoint International at DreamPort, the U.S Cyber Command’s premier cybersecurity innovation facility. There, Dell will power the ZT Center of Excellence to provide organisations with a secure data center to validate Zero Trust use cases in the flesh.
Of course, me being me, I was on a roll. I started to dig into how this will be based on the seven pillars of the Department of Defense (DoD) Zero Trust Reference Architecture. Control Plane here, Macro-segmentation there, Policy Enforcement Points everywhere!
Pause… the subject of a very blank stare…. Reminiscent of my days as a 4 year old. I knew the question was coming.
“But Why Zero Trust?”
This forced a pause. In my defense, I did stop myself leaning into the casual response centered on the standard logic: Cyber attacks are on the increase, ransomware, malware, DoS, DDoS, phishing, mobile malware, credential theft etc., ergo we must mandate Zero-Trust. Clearly this didn’t answer the question, why? Why are we facing more cyber related incidences and why shouldn’t I use existing frameworks such as ‘Defense in Depth’? We have used them for decades, they were great then, why not now? What has changed?
Of course a hint lies in the title of this post, and in particular the very first line of the DoD Reference Architecture guide.
“Zero Trust is the term for an evolving set of cybersecurity paradigms that move defenses from static, network-based perimeters to focus on users, assets, and resources. Zero Trust assumes there is no implicit trust granted to assets or user accounts based solely on their physical or network location (i.e., local area networks versus the Internet) or based on asset ownership (enterprise or personally owned)”
So the goal is to move from ‘static, network-based perimeters’ to ‘focus on users, assets and resources’. However, as you may have guessed, the next question is……
“But Why?”
I think we can formulate a relevant coherent answer to this question.
The Problem of De-Perimeterisation
Traditional approaches to network and infrastructure security are predicated on the idea that I can protect the perimeter. Stop the bad stuff at the gate and only leave the good stuff in leveraging firewalls, ACL’s, IPS and IDS systems and other platforms. ‘Defense in Depth’ has become a popular framework that enhances this network perimeter approach, by adding additional layers on the ‘inside’, another firewall here another ACL there. Just in case something gets through. Like a series more granular sieves, eventually, we will catch the bad stuff, even if it has breached the perimeter.
This approach of course has remained largely the same since the 1990’s, for as long as the Network firewall has existed. ( in fact longer but I choose not to remember that far back!)
The ‘noughties’ were characterised by relative simplicity:
- Applications all live in the ‘Data-Center’ on physical hardware. No broad adoption of virtualisation just yet. What’s born in the DC stays in the DC for the most part. Monolithic workflows.
- Hub/Spoke MPLS based WAN and Simple VPN based remote access. Generally no split tunnels allowed. In other words to get to the internet, when ‘dialed-in’ you needed to reach it via the corporate DC.
- Fledgling Internet services, pre SaaS.
- We owned pretty much all our own infrastructure.
In this scenario, the network perimeter/border is very well defined and understood. Placing firewalls and defining policy for optimal effectiveness is a straightforward process. Ports were opened towards the internet but the process was relatively static and manageable.
Interestingly, even back then we could possibly trace the beginnings of what we now know of Zero-Trust movement. In 2004, the Jericho Forum, which later merged into the Open Group Security Forum, remarked rather prophetically;
“The traditional electronic boundary between a corporate (or ‘private’) network and the Internet is breaking down in the trend which we have called de-perimeterisation“
And this was almost 20 years ago, when things were….. well, simple!
Rolling on to the next decade.
Things are beginning to change, I had to put a little thought into where I drew my rather crude red line representing the network perimeter. We now have:
- The rise of X86 and other types of server virtualisation. All very positive but lending itself to proliferation of ‘virtual machines’ within the DC. Otherwise known as VM sprawl. Software Defined Networking and Security ‘Defense in Depth’ solutions soon followed such as VMware NSX to manage these new ‘East-West’ flows in the Data Center. Inserting software based firewalls representing the birth of micro-segmentation as we know it.
- What were ‘Fledging’ Web based services have now firmly become ‘Business Critical ‘ SaaS based services. How we connected to these services became a little bit more complicated, indeed obfuscated. More and more these were machine to machine flows versus machine to human flows. For instance, my internal app tier pulling from an external web based SaaS database server. The application no longer lived exclusively in the DC nor did we have exclusive ownership rights.
- More and More, the remote workforce were using the corporate DC as a trombone transit to get to business SaaS resources on the web. This started to put pressure on the mandate around ‘thou must not split-tunnel’, simply because performance was unpredictable at best, due to latency and jitter. (Unfortunately we still haven’t figured out a way to speed up the speed of light!)
Ultimately, in order for the ‘Defend the Perimeter’ approach to be successful we need to:
- ‘Own our own infrastructure and domain.‘ Clearly we don’t own nor control the Web based SaaS services outlined above.
- ‘Understand clearly our borders, perimeter and topology.’ Our clarity is undermined here due to the ‘softening’ of the split-tunnel at the edge and our lack of true understanding of what is happening on the internet, where our web based services reside. Even within our DC, our topology is becoming much more complicated and the data flows are much more difficult to manage and understand. The proliferation of East-West flows, VM sprawl, shadow IT and development etc. If an attack breached our defenses, it is difficult to identify just how deep it may have gotten or where the malware is hiding.
- ‘Implement and enforce our security policy within our domain and at our perimeter’ Really this is dependent on 1 and 2, clearly this is now more of a challenge.
The Industry began to recognise the failings of the traditional approach. Clearly we needed a different approach. Zero Trust Architectures (ZTA), began to mature and emerge both in theory and practice.
- Forrester Research:
- 2010: John Kindervag coined the phrase ‘Zero Trust’ to describe the security model that you should not implicitly trust anything outside or inside your perimeter and instead you must verify everything and anything before connecting them to the network or granting access to their systems.
- 2018: Dr. Chase Cunningham. Led the evolution into Zero Trust eXtended Framework (ZTN). ‘Never Trust always Verify’
- Google BeyondCorp:
- 2014: BeyondCorp is Google’s implementation of the Zero-Trust model. Shifts access controls from the network perimeter to individual users, BeyondCorp enables secure work from any location without the need for a traditional VPN
- Gartner:
- 2014: Adaptive Security Architecture
- 2018: Continuous Risk and Trust Assessment (CARTA): Advocates that no user or device, even those already within the network should be inherently trusted
And so the the current decade:
Because the perimeter is everywhere, the perimeter is in essence dead…….
I refrained from the red marker on this occasion, because I would be drawing in perpetuity. The level of transformation that has taken place over the last 4-5 years in particular has been truly remarkable. This has placed an immense and indelible strain on IT Security frameworks and the network perimeter, as we know them. It is no longer necessary to regurgitate the almost daily stream of negative news pertaining to cyber related attacks on Government, Enterprise and small business globally, in order to copperfasten the argument, that we need to accelerate the adoption of a new fit for purpose approach.
In today’s landscape:
- Microservice based applications now sit everywhere in the enterprise and modern application development techniques leveraging CI/CD pipelines are becoming increasingly distributed. Pipelines may span multiple on-premise and cloud locations and change dynamically based on resourcing and budgetary needs.
- Emerging enterprises may not need a traditional DC as we know it or none at all, they may leverage the public cloud, edge, COLO and home office exclusively.
- The rise of the Edge and enabling technologies such as 5G and Private Wireless has opened up new use cases and product offerings where applications must reside close to the end-user due to latency sensitivity.
- The continued and increasing adoption of existing established enterprises of ‘Multi-Cloud’ architectures.
- The emergence of Multi-Cloud Data mobility. User and application data is moving, more and more across physical and administrative boundaries based on business and operational needs.
- The exponential growth of remote work and the nature of remote work being ‘Internet First’. More often than not, remote users are leveraging internet based applications, SaaS and not leveraging any traditional Data Center applications. Increasingly a VPN less experience is demanded by users.
- Ownership it shifting rapidly from Capex to dynamic, ‘Pay As You Use/On-demand’ Opex based on-premise cloud like consumption models, such as Dell APEX.
So, if you recall, the three key controls required to implement a ‘Perimeter’ based security model include:
- Do I own the Infrastructure? Rarely at best, more than likely some or increasingly none at all. Indeed many customers want to shift the burden of ownership completely to the Service Provider (SP).
- Do we understand clearly our border, perimeter and topology? No. In a multi-cloud world with dynamic modern application flows our perimeter is constantly changing and in flux, and in some cases disappearing.
- Can we implement security policy at the perimeter? Even if we had administrative ownership, this task would be massively onerous, given that our perimeter is now dynamic at best and possibly non existent.
So where does that leave us? Is it a case of ‘out with the old in with the new’? Absolutely not! More and more security tooling and systems will emerge to support the new Zero Trust architectures, but in reality we will use much of what already exists. Will we still leverage existing tools in our armoury such Firewalls, AV, IPS/IDS, and Micro-segmentation? Of course we will. Remember ZTA is a framework not a single product. There is no single magic bullet. It will be a structured coming together of the people, process and technology. No one product or piece of software will on its own implement Zero Trust.
What we will see though emerge, is a concentration of systems, processes and tooling in order to allow us deliver on the second half of the first statement in the DoD Reference Architecture Guide.
“Zero Trust assumes there is no implicit trust granted to assets or user accounts based solely on their physical or network location (i.e., local area networks versus the Internet) or based on asset ownership (enterprise or personally owned)”
If we can’t ‘grant trust’ based on where something resides or who owns it, then how can we ‘grant trust’ and to what level?
The answer to that lies in a systematic and robust ability to continuously authenticate and conditionally authorize every asset on the network, and to allocate access on the principle of ‘least privilege’. To that end, Identity and Access Management systems and processes (IAM) will step forward, front and center in a Zero Trust world. ( and into the next post in this Zero Trust series…..)
DISCLAIMER
The views expressed on this site are strictly my own and do not necessarily reflect the opinions or views of Dell Technologies. Please always check official documentation to verify technical information.
#IWORK4DELL
Cablestoclouds.blog 
[…] Storage Direct introduced in Release 19.17. I blogged about this feature back in late July. Click here to view…