Archive

Posts Tagged ‘Virtualization Security’

Patching the (Hypervisor) Platform: How Do You Manage Risk?

April 12th, 2010 7 comments

Hi. Me again.

In 2008 I wrote a blog titled “Patching the Cloud” which I followed up with material examples in 2009 in another titled “Redux: Patching the Cloud.

These blogs focused mainly on virtualization-powered IaaS/PaaS offerings and whilst they targeted “Cloud Computing,” they applied equally to the heavily virtualized enterprise.  To this point I wrote another in 2008 titled “On Patch Tuesdays For Virtualization Platforms.

The operational impacts of managing change control, vulnerability management and threat mitigation have always intrigued me, especially at scale.

I was reminded this morning of the importance of the question posed above as VMware released a series of security advisories detailing ten vulnerabilities across many products, some of which are remotely exploitable. While security vulnerabilities in hypervisors are not new, it’s unclear to me how many heavily-virtualized enterprises or Cloud providers actually deal with what it means to patch this critical layer of infrastructure.

Once virtualized, we expect/assume that VM’s and the guest OS’s within them should operate with functional equivalence when compared to non-virtualized instances. We have, however, seen that this is not the case. It’s rare, but it happens that OS’s and applications, once virtualized, suffer from issues that cause faults to the underlying virtualization platform itself.

So here’s the $64,000 question – feel free to answer anonymously:

While virtualization is meant to effectively isolate the hardware from the resources atop it, the VMM/Hypervisor itself maintains a delicate position arbitrating this abstraction.  When the VMM/Hypervisor needs patching, how do you regression test the impact across all your VM images (across test/dev, production, etc.)?  More importantly, how are you assessing/measuring compound risk across shared/multi-tenant environments with respect to patching and its impact?

/Hoff

P.S. It occurs to me that after I wrote the blog last night on ‘high assurance (read: TPM-enabled)’ virtualization/cloud environments with respect to change control, the reference images for trust launch environments would be impacted by patches like this. How are we going to scale this from a management perspective?

Reblog this post [with Zemanta]

More On High Assurance (via TPM) Cloud Environments

April 11th, 2010 14 comments
North Bridge Intel G45
Image via Wikipedia

Back in September 2009 after presenting at the Intel Virtualization (and Cloud) Security Summit and urging Intel to lead by example by pushing the adoption and use of TPM in virtualization and cloud environments, I blogged a simple question (here) as to the following:

Does anyone know of any Public Cloud Provider (or Private for that matter) that utilizes Intel’s TXT?

Interestingly the replies were few; mostly they were along the lines of “we’re considering it,” “…it’s on our long radar,” or “…we’re unclear if there’s a valid (read: economically viable) use case.”

At this year’s RSA Security Conference, however, EMC/RSA, Intel and VMware made an announcement regarding a PoC of their “Trusted Cloud Infrastructure,” describing efforts to utilize technology across the three vendors’ portfolios to make use of the TPM:

The foundation for the new computing infrastructure is a hardware root of trust derived from Intel Trusted Execution Technology (TXT), which authenticates every step of the boot sequence, from verifying hardware configurations and initialising the BIOS to launching the hypervisor, the companies said.

Once launched, the VMware virtualisation environment collects data from both the hardware and virtual layers and feeds a continuous, raw data stream to the RSA enVision Security Information and Event Management platform. The RSA enVision is engineered to analyse events coming through the virtualisation layer to identify incidents and conditions affecting security and compliance.

The information is then contextualised within the Archer SmartSuite Framework, which is designed to present a unified, policy-based assessment of the organisation’s security and compliance posture through a central dashboard, RSA said.

It should be noted that in order to take advantage of said solution, the following components are required: a future release of RSA’s Archer GRC console, the upcoming Intel Westmere CPU and a soon-to-be-released version of VMware’s vSphere.  In other words, this isn’t available today and will require upgrades up and down the stack.

Sam Johnston today pointed me toward an announcement from Enomaly referencing the “High Assurance Edition” of ECP which laid claims of assurance using the TPM beyond the boundary of the VMM to include the guest OS and their management system:

Enomaly’s Trusted Cloud platform provides continuous security assurance by means of unique, hardware-assisted mechanisms. Enomaly ECP High Assurance Edition provides both initial and ongoing Full-Stack Integrity Verification to enable customers to receive cryptographic proof of the correct and secure operation of the cloud platform prior to running any application on the cloud.

  • Full-Stack Integrity Verification provides the customer with hardware-verified proof that the cloud stack (encompassing server hardware, hypervisor, guest OS, and even ECP itself) is intact and has not been tampered with. Specifically, the customer obtains cryptographically verifiable proof that the hardware, hypervisor, etc. are identical to reference versions that have been certified and approved in advance. The customer can therefore be assured, for example, that:
  • The hardware has not been modified to duplicate data to some storage medium of which the application is not aware
  • No unauthorized backdoors have been inserted into the cloud managment system
  • The hypervisor has not been modified (e.g. to copy memory state)
  • No hostile kernel modules have been injected into the guest OS
This capability therefore enables customers to deploy applications to public clouds with confidence that the confidentiality and integrity of their data will not be compromised.

Of particular interest was Enomaly’s enticement of service providers with the following claim:

…with Enomaly’s patented security functionality, can deliver a highly secure Cloud Computing service – commanding a higher price point than commodity public cloud providers.

I’m looking forward to exploring more regarding these two example solutions as they see the light of day (and how long this will take given the need for platform-specific upgrades up and down the stack) as well as whether or not customers are actually willing to pay — and providers can command — a higher price point for what these components may offer.  You can bet certain government agencies are interested.

There are potentially numerous benefits with the use of this technology including security, compliance, assurance, audit and attestation capabilities (I hope also to incorporate more of what this might mean into the CloudAudit/A6 effort) but I’m very interested as to the implications on (change) management and policy, especially across heterogeneous environments and the extension and use of TPM’s across mobile platforms.

Of course, researchers are interested in these things too…see Rutkowska, et. al and “Attacking Intel Trusted Execution Technology” as an example.

/Hoff

Related articles by Zemanta

Reblog this post [with Zemanta]

Slides from My Cloud Security Alliance Keynote: The Cloud Magic 8 Ball (Future Of Cloud)

March 7th, 2010 No comments

Here are the slides from my Cloud Security Alliance (CSA) keynote from the Cloud Security Summit at the 2010 RSA Security Conference.

The punchline is as follows:

All this iteration and debate on the future of the “back-end” of Cloud Computing — the provider side of the equation — is ultimately less interesting than how the applications and content served up will be consumed.

Cloud Computing provides for the mass re-centralization of applications and data in mega-datacenters while simultaneously incredibly powerful mobile computing platforms provide for the mass re-distribution of (in many cases the same) applications and data.  We’re fixated on the security of the former but ignoring that of the latter — at our peril.

People worry about how Cloud Computing puts their applications and data in other people’s hands. The reality is that mobile computing — and the clouds that are here already and will form because of them — already put, quite literally, those applications and data in other people’s hands.

If we want to “secure” the things that matter most, we must focus BACK on information centricity and building survivable systems if we are to be successful in our approach.  I’ve written about the topics above many times, but this post from 2009 is quite apropos: The Quandary Of the Cloud: Centralized Compute But Distributed Data You can find other posts on Information Centricity here.

Slideshare direct link here (embedded below.)

Reblog this post [with Zemanta]

Don’t Hassle the Hoff: Recent Press & Podcast Coverage & Upcoming Speaking Engagements

February 19th, 2010 No comments

Here is some of the recent coverage from the last couple of months or so on topics relevant to content on my blog, presentations and speaking engagements.  No particular order or priority and I haven’t kept a good record, unfortunately.

Important Stuff I’m Working On:

Press/Technology & Security eZines/Website/Blog Coverage/Meaningful Links:

Recent Speaking Engagements/Confirmed to  speak at the following upcoming events:

  • Govt Solutions Forum Feb 1-2 (panel |n DC)
  • Govt Solutions Forum Feb 24 D.C.
  • ESAF, San Francisco, March 1
  • Cloud Security Alliance Summit, San Francisco, March 1
  • RSA Security Conference March 1-5 San Francisco
  • Microsoft Bluehat Buenos Aires, Argentina – March 16-19th
  • ISSA General Assembly, Belgium
  • Infosec.be, Belgium
  • Codegate, South Korea, April 7-8
  • SOURCE Boston, April 21-23
  • Shot the Sherrif – Brazil – May 17th
  • Gluecon , Denver, May 26/27
  • FIRST, Miami, FL,  June 13-18
  • SANS DC – August 19th-20th

Conferences I am tentatively attending, trying to attend and/or working on logistics for speaking:

  • InterOp April 25-29 Vegas
  • Cisco Live – June 27th – July 1st Vegas
  • Blackhat 2010 – July 24-29 Vegas
  • Defcon
  • Notacon

Oh, let us not forget these top honors (buahahaha!)

  • Top 10 Sexy InfoSec Geeks (link)
  • The ThreatPost “All Decade Interview Team” (link)
  • ‘Cloud Hero’ and ‘Best Cloud Presentation’ – 2009 Cloudies Awards (link), and
  • 2010 RSA Social Security Bloggers Award nomination (link) 😉

[I often get a bunch of guff as to why I make these lists: ego, horn-tooting, self-aggrandizement. I wish I thought I were that important. 😉 The real reason is that it helps me keep track of useful stuff focused not only on my participation, but that of the rest of the blogosphere.]

/Hoff

Where Are the Network Virtual Appliances? Hobbled By the Virtual Network, That’s Where…

January 31st, 2010 15 comments

Allan Leinwand from GigaOm wrote a great article asking “Where are the network virtual appliances?” This was followed up by another excellent post by Rich Miller.

Allan sets up the discussion describing how we’ve typically plumbed disparate physical appliances into our network infrastructure to provide discrete network and security capabilities such as load balancers, VPNs, SSL termination, firewalls, etc.  He then goes on to describe the stunted evolution of virtual appliances:

To be sure, some networking devices and appliances are now available in virtual form.  Switches and routers have begun to move toward virtualization with VMware’s vSwitch, Cisco’s Nexus 1000v, the open source Open vSwitch and routers and firewalls running in various VMs from the company I helped found, Vyatta.  For load balancers, Citrix has released a version of its Netscaler VPX software that runs on top of its virtual machine, XenServer; and Zeus Systems has an application traffic controller that can be deployed as a virtual appliance on Amazon EC2, Joyent and other public clouds.

Ultimately I think it prudent for discussion’s sake to separate routing, switching and load balancing (connectivity) from functions such as DLP, firewalls, and IDS/IPS (security) as lumping them together actually abstracts the problem which is that the latter is completely dependent upon the capabilities and functionality of the former.  This is what Allan almost gets to when describing his lament with the virtual appliance ecosystem today:

Yet the fundamental problem remains: Most networking appliances are still stuck in physical hardware — hardware that may or may not be deployed where the applications need them, which means those applications and their associated VMs can be left with major gaps in their infrastructure needs. Without a full-featured and stateful firewall to protect an application, it’s susceptible to various Internet attacks.  A missing load balancer that operates at layers three through seven leaves a gap in the need to distribute load between multiple application servers. Meanwhile, the lack of an SSL accelerator to offload processing may lead to performance issues and without an IDS device present, malicious activities may occur.  Without some (or all) of these networking appliances available in a virtual environment, a VM may find itself constrained, unable to take full advantage of the possible economic benefits.

I’ve written about this many, many times. In fact almost three years ago I created a presentation called  “The Four Horsemen of the Virtualization Security Apocalypse” which described in excruciating detail how network virtual appliances were a big ball of fail and would be for some time. I further suggested that much of the “best-of-breed” products would ultimately become “good enough” features in virtualization vendor’s hypervisor platforms.

Why?  Because there are some very real problems with virtualization (and Cloud) as it relates to connectivity and security:

  1. Most of the virtual network appliances, especially those “ported” from the versions that usually run on dedicated physical hardware (COTS or proprietary) do not provide feature, performance, scale or high-availability parity; most are hobbled or require per-platform customization or re-engineering in order to function.
  2. The resilience and high availability options from today’s off-the-shelf virtual connectivity does not pair well with the mobility and dynamism of de-coupled virtual machines; VMs are ultimately temporal and networks don’t like topological instability due to key components moving or disappearing
  3. The performance and scale of virtual appliances still suffer when competing for I/O and resources on the same physical hosts as the guests they attempt to protect
  4. Virtual connectivity is a generally a function of the VMM (or a loadable module/domain therein.) The architecture of the VMM has dramatic impact upon the architecture of the software designed to provide the connectivity and vice versa.
  5. Security solutions are incredibly topology sensitive.  Given the scenario in #1 when a VM moves or is distributed across the pooled infrastructure, unless the security capabilities are already present on the physical host or the connectivity and security layers share a control plane (or at least can exchange telemetry,) things will simply break
  6. Many virtualization (and especially cloud) platforms do not support protocols or topologies that many connectivity and security virtual appliances require to function (such as multicast for load balancing)
  7. It’s very difficult to mimic the in-line path requirements in virtual networking environments that would otherwise force traffic passing through the connectivity layers (layers 2 through 7) up through various policy-driven security layers (virtual appliances)
  8. There is no common methodology to express what security requirements the connectivity fabrics should ensure are available prior to allowing a VM to spool up let alone move
  9. Virtualization vendors who provide solutions for the enterprise have rich networking capabilities natively as well as with third party connectivity partners, including VM and VMM introspection capabilities. As I wrote about here, mass-market Cloud providers such as Amazon Web Services or Rackspace Cloud have severely crippled networking.
  10. Virtualization and cloud vendors generally force many security vs. performance tradeoffs when implementing introspection capabilities in their platforms: third party code running in the kernel, scheduler prioritization issues, I/O limitations, etc.
  11. Much of the basic networking capabilities are being pushed lower into silicon (into the CPUs themselves) which makes virtual appliances even further removed from the guts that enable them
  12. Physical appliances (in the enterprise) exist en-mass.  Many of them provide highly scalable solutions to the specific functions that Alan refers to.  The need exists, given the limitations I describe above, to provide for integration/interaction between them, the VMM and any virtual appliances in order to offload certain functions as well as provide coverage between the physical and the logical.

What does this mean?  It means that ultimately to ensure their own survival, virtualization and cloud providers will depend less upon virtual appliances and add more of the basic connectivity AND security capabilities into the VMMs themselves as its the only way to guarantee performance, scalability, resilience and satisfy the security requirements of customers. There will be new generations of protocols, APIs and control planes that will emerge to provide for this capability, but this will drive the same old integration battles we’re supposed to be absolved from with virtualization and Cloud.

Connectivity and security vendors will offer virtual replicas of their physical appliances in order to gain a foothold in virtualized/cloud environments in order to intercept traffic (think basic traps/ACL’s) and then interact with higher-performing physical appliance security service overlays or embedded line cards in service chassis.  This is especially true in enterprises but poses many challenges in software-only, mass-market cloud environments where what you’ll continue to get is simply basic connectivity and security with limited networking functionality.  This implies more and more security will be pushed into the guest and application logic layers to deal with this disconnect.

This is exactly where we are today with Cloud providers like Amazon Web Services: basic ingress-only filtering with a very simplistic, limited and abstracted set of both connectivity and security capability.  See “Dear Public Cloud Providers: Please Make Your Networking Capabilities Suck Less. Kthxbye”  Will they add more functionality?  Perhaps. The question is whether they can afford to in order to limit the impact that connecitivity and security variability/instability can bring to an environment.

That said, it’s certainly achievable, if you are willing and able to do so, to construct a completely software-based networking environment, but these environments require a complete approach and stack re-write with an operational expertise that will be hard to support for those who have spent the last 20 years working in a different paradigm and that’s a huge piece of this problem.

The connectivity layer — however integrated into the virtualized and cloud environments they seem — continues to limit how and what the security layers can do and will for some time, thus limiting the uptake of virtual network and security appliances.

Situation normal.

/Hoff

Reblog this post [with Zemanta]

Hacking Exposed: Virtualization & Cloud Computing…Feedback Please

January 30th, 2010 26 comments

Craig Balding, Rich Mogull and I are working on a book due out later this year.

It’s the latest in the McGraw-Hill “Hacking Exposed” series.  We’re focusing on virtualization and cloud computing security.

We have a very interesting set of topics to discuss but we’d like to crowd/cloud-source ideas from all of you.

The table of contents reads like this:

Part I: Virtualization & Cloud Computing:  An Overview
Case Study: Expand the Attack Surface: Enterprise Virtualization & Cloud Adoption
Chapter 1: Virtualization Defined
Chapter 2: Cloud Computing Defined

Part II: Smash the Virtualized Stack
Case Study: Own the Virtualized Enterprise
Chapter 3: Subvert the CPU & Chipsets
Chapter 4: Harass the Host, Hypervisor, Virtual Networking & Storage
Chapter 5: Victimize the Virtual Machine
Chapter 6: Conquer the Control Plane & APIs

Part III: Compromise the Cloud
Case Study: Own the Cloud for Fun and Profit
Chapter 7: Undermine the Infrastructure
Chapter 8: Manipulate the Metastructure
Chapter 9: Assault the Infostructure

Part IV: Appendices

We’ll have a book-specific site up shortly, but if you’d like to see certain things covered (technology, operational, organizational, etc.) please let us know in the comments below.

Also, we’d like to solicit a few critical folks to provide feedback on the first couple of chapters. Email me/comment if interested.

Thanks!

/Hoff, Craig and Rich.

Reblog this post [with Zemanta]

To Achieve True Cloud (X/Z)en, One Must Leverage Introspection

January 6th, 2010 No comments

Back in October 2008, I wrote a post detailing efforts around the Xen community to create a standard security introspection API (Xen.Org Launches Community Project To Bring VM Introspection to Xen🙂

The Xen Introspection Project is a community effort within Xen.org to leverage the existing research presented above with other work not yet public to create a standard API specification and methodology for virtual machine introspection.

That blog was focused on introspection for virtualization proper but since many of the larger cloud providers utilize Xen virtualization as an underpinning of their service architecture and as an industry we’re suffering from a lack of visibility and deployable security capabilities, the relevance of VM and VMM introspection to cloud computing is quite relevant.

I thought I’d double around and see where we are.

It looks as though there’s been quite a bit of recent activity from the folks at Georgia Tech (XenAccess Project) and the University of Alaska at Fairbanks (Virtual Introspection for Xen) referenced in my previous blog.  The vCloud API proffered via the DMTF seems to also leverage (at least some of) the VMsafe API capabilities present in VMware‘s vSphere virtualization platform.

While details are, for obvious reasons sketchy, I am encouraged in speaking to representatives from a few cloud providers who are keenly interested in including these capabilities in their offerings.  Wouldn’t that be cool?

Adoption and inclusion of introspection capabilities will overcome some of the inherent security and visibility limitations we face in highly-virtualized multi-tenant environments due to networking constraints for integrating security functionality that I wrote about here.

I plan a follow-on blog in more detail once I finish some interviews.

/Hoff

Reblog this post [with Zemanta]

Great InformationWeek/Dark Reading/Black Hat Cloud & Virtualization Security Virtual Panel on 12/9

December 3rd, 2009 No comments

darkreading-blackhat

I wanted to let you know about about a cool virtual panel I’m moderating as part of the InformationWeek/Dark Reading/Black Hat virtual event titled “IT Security: The Next Decade” on December 9th.

There are numerous awesome speakers throughout the day, but the panel I’m moderating is especially interesting to me because I was able to get an amazing set of people to participate.  Here’s the rundown — check out the panelists:

Virtualization, Cloud Computing, And Next-Generation Security

The concept of cloud computing creates new challenges for security, because sensitive data may no longer reside on dedicated hardware.  How can enterprises protect their most sensitive data in the rapidly-evolving world of shared computing resources? In this panel, Black Hat researchers who have found vulnerabilities in the cloud and software-as-a-service models meet other experts on virtualization and cloud computing to discuss the question of cloud computing’s impact on security and the steps that will be required to protect data in cloud environments.

Panelists: Glenn Brunette, Distinguished Engineer and Chief Security Architect, Sun Microsystems; Edward Haletky, Virtualization Security Expert; Chris Wolf, Virtualization Analyst, Burton Group; Jon Oberheide, Security Researcher; Craig Balding, Cloud Security Expert, cloudsecurity.org

Moderator: Christofer Hoff, Contributing Editor, Black Hat

I wanted the perspective of architects/engineers, practitioners, researchers and analysts — and I couldn’t have asked for a better group.

Our session is 6:15pm – 7:00 EST.

Hope to “see” you there.

/Hoff

From the X-Files – The Cloud in Context: Evolution from Gadgetry to Popular Culture

November 27th, 2009 4 comments

apple1984

[This post was originally authored November 27, 2009.  I pushed it back to the top of the stack because I think it’s an interesting re-visitation of the benefits and challenges we are experiencing in Cloud today]

Below is an article I wrote many months ago prior to all the Nicholas Carr “electricity ain’t Cloud” discussions.  The piece was one from a collection that was distributed to “…the Intelligence Community, the DoD, and Congress” with the purpose of giving a high-level overview of Cloud security issues.

The Cloud in Context: Evolution from Gadgetry to Popular Culture

It is very likely that should one develop any interest in Cloud Computing (“Cloud”) and wish to investigate its provenance, one would be pointed to Nicholas Carr’s treatise “The Big Switch” for enlightenment. Carr offers a metaphoric genealogy of Cloud Computing, mapped to, and illustrated by, a keenly patterned set of observations from one of the most important catalysts of a critical inflection point in modern history: the generation and distribution of electricity.

Carr offers an uncannily prescient perspective on the evolution and adaptation of computing by way of this electric metaphor, describing how the scale of technology, socioeconomic, and cultural advances were all directly linked to the disruptive innovation of a shift from dedicated power generation in individual factories to a metered utility of interconnected generators powering distribution grids feeding all. He predicts a similar shift from insular, centralized, private single-function computational gadgetry to globally-networked, distributed, public service-centric collaborative fabrics of information interchange.

This phenomenon will not occur overnight nor has any other paradigm shift in computing occurred overnight; bursts of disruptive innovation have a long tail of adoption. Cloud is not the product or invocation of some singular technology, but rather an operational model that describes how computing will mature.

There is no box with blinking lights that can be simply pointed to as “Cloud” and yet it is clearly more than just timesharing with Internet connectivity. As corporations seek to drive down cost and gain efficiency force-multipliers, they have ruthlessly focused on divining what is core to their businesses, and expensive IT cost-centers are squarely in the crosshairs for rigorous valuation.

To that end, Carr wrote another piece on this very topic titled “IT Doesn’t matter” in which he argued that IT was no longer a strategic differentiator due to commoditization, standardization, and cost. This was followed by “The End of Corporate Computing” wherein he suggested that IT will simply subscribe to IT services as an outsourced function. Based upon these themes, Cloud seems a natural evolutionary outcome motivated primarily by economics as companies pare down their IT investment — outsourcing what they can and optimizing what is left.

Enter Cloud Computing

The emergence of Cloud as cult-status popular culture also has its muse anchored firmly in the little machines nestled in the hands of those who might not realize that they’ve helped create the IT revolution at all: the consumer. The consumer’s shift to an always-on, many-to-many communication model with unbridled collaboration and unfettered access to resources, sharply contrasts with traditional IT — constrained, siloed, well-demarcated, communication-restricted, and infrastructure-heavy.

Regardless of any value judgment on the fate of Man, we are evolving to a society dedicated to convenience, where we are not tied to the machine, but rather the machine is tied to us, and always on. Your applications and data are always there, consumed according to business and pricing models that are based upon what you use while the magic serving it up remains transparent.

This is Cloud in a nutshell; the computing equivalent to classical Greek theater’s Deus Ex Machina.

For the purpose of this paper, it is important that I point out that I refer mainly to so-called “Public Cloud” offerings; those services provided by parties external to the data owner who provides an “outsourced” service capability on behalf of the consumer.

This graceful surrender of control is the focus of my discussion. Private Clouds — those services that may operate on the corporation’s infrastructure or those of a provider but managed under said corporation’s control and policies, offers a different set of benefits and challenges but not to the degree of Public Cloud.

There are also hybrid and brokered models, but to keep focused, I shall not address these directly.

Cloud Reference Model

Cloud Reference Model

A service is generally considered to be “Cloud-based” should it meet the following characteristics and provide for:

  • The abstraction of infrastructure from the resources that deliver them
  • The democratization of those resources as an elastic pool to be consumed
  • Services-oriented, rather than infrastructure or application-centric
  • Enabling self-service, scale on-demand elasticity and dynamism
  • Employs a utility-like model of consumption and allocation

Cloud exacerbates the issues we have faced for years in the information security, assurance, and survivability spaces and introduces new challenges associated with extreme levels of abstraction, mobility, scale, dynamism and multi-tenancy. It is important that one contemplate the “big picture” of how Cloud impacts the IT landscape and how given this “service- centric” view, certain things change whilst others remain firmly status quo.

Cloud also provides numerous challenges to the way in which computing and resources are organized, operated, governed and secured, given the focus on:

  • Automated and autonomic resource provisioning and orchestration
  • Massively interconnected and mashed-up data sources, conduits and results
  • Virtualized layers of software-driven, service-centric capability rather than infrastructure or application- specific monoliths
  • Dynamic infrastructure that is aware of and adjusts to the information, applications and services (workloads) running over it, supporting dynamism and abstraction in terms of scale, policy, agility, security and mobility

As a matter of correctness, virtualization as a form of abstraction may exist in many forms and at many layers, but it is not required for Cloud. Many Cloud services do utilize virtualization to achieve scale and I make liberal use of this assumptive case in this paper. As we grapple with the tradeoffs between convenience, collaboration, and control, we find that existing products, solutions and services are quickly being re-branded and adapted as “Cloud” to the confusion of all.keep focused, I shall not address these directly.

Modeling the Cloud

There exist numerous deployment, service delivery models and use cases for Cloud, each offering a specific balance of integrated features, extensibility/ openness and security hinged on high levels of automation for workload distribution.

Three archetypal models generally describe cloud service delivery, popularly referred to as the “SPI Model,” where “SPI” refers to Software, Platform and Infrastructure (as a service) respectively.

NIST - Visual Cloud Model

NIST – Visual Cloud Model

Using the National Institute of Standards and Technology’s (NIST) draft working definition as the basis for the model:

Software as a Service (SaaS)

The capability provided to the consumer is to use the provider’s applications running on a cloud infrastructure and accessible from various client devices through a thin client interface such as a Web browser (e.g., web-based email).

The consumer does not manage or control the underlying cloud infrastructure, network, servers, operating systems, storage, or even individual application capabilities, with the possible exception of limited user-specific application configuration settings.

Platform as a Service (PaaS)

The capability provided to the consumer is to deploy onto the cloud infrastructure consumer-created applications using programming languages and tools supported by the provider (e.g., Java, Python, .Net). The consumer does not manage or control the underlying cloud infrastructure,

Infrastructure as a Service (IaaS)

The capability provided to the consumer is to rent processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software, which can include operating systems and applications. The consumer does not manage or control the underlying cloud infrastructure but has control over operating systems, storage, deployed applications, and possibly select networking components (e.g., firewalls, load balancers).

Understanding the relationship and dependencies between these models is critical. IaaS is the foundation of all Cloud services with PaaS building upon IaaS, and SaaS — in turn — building upon PaaS. We will cover this in more detail later in the document.

Peanut Butter & Jelly — Making the Perfect Cloud Sandwich

Infostructure/Metastructure/Infrastructure

Infostructure/Metastructure/Infrastructure

To understand how Cloud will affect security, visualize its functional structure in three layers:

  • The Infrastructure layer represents the traditional compute, network and storage hardware and operating systems familiar to us all. Virtualization platforms also exist at this layer and expose their capabilities northbound.
  • The Infostructure layer represents the programmatic components such as applications and service objects that produce, operate on or interact with the content, information and metadata.
  • Sitting in between Infrastructure and Infostructure is the Metastructure layer. This layer represents the underlying set of protocols and functions with layers such as DNS, BGP, and IP address management, which “glue” together and enable the applications and content at the Infostructure layer to in turn be delivered by the Infrastructure.

Certain areas of Cloud Computing’s technology underpinnings are making progress, but those things that will ultimately make Cloud the ubiquitous and transparent platform for our entire computing experience remain lacking.

Unsurprisingly, most of the deficient categories of technology or capabilities are those that need to be delivered from standards and consensus-driven action; things that have always posed challenges such as management, governance, provisioning, orchestration, automation, portability, interoperability and security. As security solutions specific to Cloud are generally slow in coming while fast innovating attackers are unconstrained by rules of engagement, it will come as no surprise that we are constantly playing catch up.

Cloud is a gradual adaptation rather than a wholesale re-tooling, and represents another cycle of investment which leaves us to consider where to invest our security dollars to most appropriately mitigate threat and vulnerability:

Typically, we react by cycling between investing in host-based controls > application controls > information controls > user controls > network controls and back again. While our security tools tend to be out of phase and less innovative than the tools of our opposition, virtualization and Cloud may act as much needed security forcing functions that get us beyond solving just the problem du jour.

The need to apply policy to workloads throughout their lifecycle, regardless of state, physical location, or infrastructure from which they are delivered, is paramount. Collapsing the atomic unit of the datacenter to the virtual machine boundary may allow for a simpler set of policy expressions that travel with the VM instance. At the same time, Cloud’s illusion of ubiquity and infinite scale means that we will not know where our data is stored, processed, or used.

Combine mobility, encryption, distributed resources with multiple providers, and a lack of open standards with economic cost pressure and even basic security capabilities seem daunting. Cloud simultaneously re-centralizes some resources while de-perimeterizing trust boundaries and distributing data. Understanding how the various layers map to traditional non-Cloud architecture is important, especially in relation to the Cloud deployment model used; there are significant trade-offs in integration, extensibility, cost, management, governance, compliance, and security.

Live by the Cloud, Die by the Cloud

Despite a tremendous amount of interest and momentum, Cloud is still very immature — pockets of innovation spread out across a long-tail of mostly-proprietary infrastructure-, platform-, and software-as-a-service offerings that do not provide for much in the way of or workload portability or interoperability.

Cloud is not limited to lower cost “server” functionality. With the fevered adoption of netbooks, virtualization, low-cost storage services, fixed/mobile convergence, the proliferation of “social networks,” and applications built to take advantage of all of this, Cloud becomes a single pane of glass for our combined computing experience. N.B., these powers are not inherently ours alone; the same upside can be used for wrongdoing.

In an attempt to whet the reader’s appetite in regards to how Cloud dramatically impacts the risk modeling, assumptions, and security postures of today, I will provide a reasonably crisp set of examples, chosen to bring pause:

Organizational and Operational Misalignment

The way in which most enterprise IT organizations are structured — in functional silos optimized to specialized, isolated functions — is diametrically opposed to the operational abstraction provided by Cloud.

The on-demand, elastic and self-service capabilities through simple interfaces and automated service layers abstract away core technology and support staff alike.

Few IT departments are prepared for what it means to apply controls, manage service levels, implement and manage security capabilities, and address compliance when the IT department is operationally irrelevant in that process. This leaves huge gaps in both identifying and managing risk, especially in outsourced models where ultimately the operational responsibility is “Cloudsourced” but the accountability is not.

The ability to apply specific security controls and measure compliance in mass-marketed Public Cloud services presents very real barriers to entry for enterprises who are heavily regulated, especially when balanced against the human capital (expertise) built-up by organizations.

Monoculture of Operating Systems, Virtualized Components, and Platforms

The standardization (de facto and de jure) on common interfaces to Cloud resources can expose uniform attack vectors that could affect one consumer, or, in the case of multi-tenant Public Cloud offerings, affect many. This is especially true in IaaS offerings where common sets of abstraction layers (such as hypervisors,) prototyped OS/application bundles (usually in the form of virtual machines) and common sets of management functions are used — and used to extend and connect the walled garden internal assets of enterprises to the public or semi-public Cloud environments of service providers operating infrastructure in proxy.

While most attack vectors target applications and information at the Infostructure layer or abuse operating systems and assorted hardware at the Infrastructure layer, the Metastructure layer is beginning to show signs of stress also. Recent attacks against key Metastructure elements such as BGP and DNS indicate that aging protocols do not fare well.

Segmentation and Isolation In Multi-tenant environments

Multi-tenancy in the Cloud (whether in the Public or Private Cloud contexts) brings new challenges to trust, privacy, resiliency and reliability model assertions by providers.  Many of these assertions are based upon the premise that that we should trust — without reliably provable models or evidence — that in the absence of relevant illustration, Cloud is simply trustworthy in all of these dimensions, despite its immaturity. Vendors claim “airtight” information, process, application, and service, but short of service level agreements, there is little to demonstrate or substantiate the claims that software-enabled Cloud Computing — however skinny the codebase may be — is any more (or less) secure than what we have today, especially with commercialized and proprietary implementations.

In multi-tenant Cloud offerings, exposures can affect millions, and placing some types of information in the care of others without effective compensating controls may erode the ROI valuation offered by Cloud in the first place, and especially so as the trust boundaries used to demarcate and segregate workloads of different consumers are provided by the same monoculture operating system and virtualization platforms described above.

Privacy of Data/Metadata, Exfiltration, and Leakage

With increased adoption of Cloud for sensitive workloads, we should expect innovative attacks against Cloud assets, providers, operators, and end users, especially around the outsourcing and storage of confidential information. The uptake is that solutions focused on encryption, at rest and in motion, will have the side effect of more and more tools (legitimate or otherwise) losing visibility into file systems, application/process execution, information and network traffic. Key management becomes remarkably relevant once again — on a massive scale.

Recent proof-of-concepts such as so-called side- channel attacks demonstrate how it is possible to determine where a specific virtual instance is likely to reside in a Public multi-tenant Cloud and allow an attacker to instantiate their own instance and cause it to be located such that it is co-resident with the target. This would potentially allow for sniffing and exfiltration of confidential data — or worse, potentially exploit vulnerabilities which would violate the sanctity of isolated workloads within the Cloud itself.

Further, given workload mobility — where the OS, applications and information are contained in an instance represented by a single atomic unit such as a virtual machine image — the potential for accidental or malicious leakage and exfiltration is real. Legal intercept, monitoring, forensics, and attack detection/incident response are heavily impacted, especially at the volume and levels of traffic envisioned by large Cloud providers, creating blind spots in ways we can’t fathom today.

Inability to Deploy Compensating or Detective Controls

The architecture of Cloud services — as abstract as they ought to be — means that in many cases the security of workloads up and down the stack are still dependent upon the underlying platform for enforcement. This is problematic inasmuch as the constructs representing compute, networking and storage resources — and security — are in many cases themselves virtualized.

Further we are faced with more stealthy and evasive malware that is able to potentially evade detection while co-opting (or rootkitting) not only software and hypervisors, but exploiting vulnerabilities in firmware and hardware such as CPU chipsets.

These sorts of attack vectors are extremely difficult to detect let alone defend against. Referring back to the monoculture issue above, a so-called blue- pilled hypervisor, uniform across tens of thousands of compute nodes providing multi-tenant Cloud services could be catastrophic. It is simply not yet feasible to provide parity in security capabilities between physical and Cloud environments; the maturity of solutions just isn’t there.

These are heady issues and should not be taken lightly when considering what workloads and services are candidates for various Cloud offerings.

What’s old is news again…

Perhaps it is worth adapting familiar attack taxonomies to Cloud.

Botnets that previously required massive malware- originated endpoint compromise in order to function can easily activate in standardized fashion, in apparently legitimate form, and in large numbers by criminals who wish to harness the organized capabilities of Bots without the effort. Simply use stolen credit cards to establish fake accounts using a provider’s Infrastructure-as-a-Service and hundreds or thousands of distributed images could be activated in a very short timeframe.

Existing security threats such as DoS/DDoS attacks, SPAM and phishing will continue to be a prime set of tools for the criminal ecosystem to leverage the distributed and well-connected Cloud as well as targeted attacks against telecommuters using both corporate and consumerized versions of Cloud services.

Consider a new take on an old problem based on ecommerce: Click-fraud. I frame this new embodiment as something called EDoS — economic denial of sustainability. Distributed Denial of Service (DDoS) attacks are blunt force trauma. The goal, regardless of motive, is to overwhelm infrastructure and remove from service a networked target by employing a distributed number of attackers. An example of DDoS is where a traditional botnet is activated to swarm/overwhelm an Internet connected website using an asynchronous attack which makes the site unavailable due to an exhaustion of resources (compute, network, or storage.)

EDoS attacks, however, are death by a thousand cuts. EDoS can also utilize distributed attack sources as well as single entities, but works by making legitimate web requests at volumes that may appear to be “normal” but are done so to drive compute, network, and storage utility billings in a cloud model abnormally high.

An example of EDoS as a variant of click fraud is where a botnet is activated to visit a website whose income results from ecommerce purchases. The requests are all legitimate but purchases are never made. The vendor has to pay the cloud provider for increased elastic use of resources but revenue is never recognized to offset them.

We have anti-DDoS capabilities today with tools that are quite mature. DDoS is generally easy to spot given huge increases in traffic. EDoS attacks are not necessarily easy to detect, because the instrumentation and business logic is not present in most applications or stacks of applications and infrastructure to provide the correlation between “requests” and “ successful transactions.” In theexample above, increased requests may look like normal activity. Many customers do not invest in this sort of integration and Cloud providers generally will not have visibility into applications that they do not own.

Ultimately the most serious Cloud concern is presented by way of the “stacked turtles” analogy: layer upon layer of complex interdependencies at the Infastructure, Metastructure and Infostructure layers, predicated upon fragile trust models framed upon nothing more than politeness. Without re-engineering these models, strengthening the notion of (id)entity management, authentication and implementing secure protocols, we run the risk of Cloud simply obfuscating the fragility of the supporting layers until something catastrophic occurs.

Combined with where and how our data is created, processed, accessed, stored, and backed up — and by whom and using whose infrastructure — Cloud yields significant concerns related to on-going security, privacy, compliance and resiliency.

Moving Forward – Critical Areas of Focus

The Cloud Security Alliance (http://www. cloudsecurityalliance.org) issued its “Guidance for Critical Areas of Focus” related to Cloud Computing Security and defined fifteen domains of concern:

  • Cloud Architecture
  • Information lifecycle management
  • Governance and Enterprise Risk Management
  • Compliance & Audit
  • General Legal
  • eDiscovery
  • Encryption and Key Management
  • Identity and Access Management
  • Storage
  • Virtualization
  • Application Security
  • Portability & Interoperability
  • Data Center Operations Management
  • Incident Response, Notification, Remediation
  • “Traditional” Security impact (business continuity, disaster recovery, physical security)

The sheer complexity of the interdependencies between the Infrastructure, Metastructure and Infostructure layers makes it almost impossible to recommend focusing on only a select subset of these items since all are relevant and important.

Nevertheless, those items in boldface most deserve initial focus just to retain existing levels of security, resilience, and compliance while information and applications are moved from the walled gardens of the private enterprise into the care of others.

Attempting to retain existing levels of security will consume the majority of Cloud transition effort.  Until we see an expansion of available solutions to bridge the gaps between “traditional” IT and dynamic infrastructure 2.0 capabilities, any company can only focus on the traditional security elements of sound design, encryption, identity, storage, virtualization and application security. Similarly, until a standardized set of methods allow well-defined interaction between the Infrastructure, Metastructure and Infostructure layers, companies will be at the mercy of industry for instrumenting, much less auditing,

Cloud elements — yet, as was already stated, the very sameness of standardization creates shared risk. As with any change of this magnitude, the potential of Cloud lies between its trade-offs. In security terms, this “big switch” surrenders visibility and control so as to gain agility and efficiency. The question is, how to achieve a net positive result?

Well-established enterprise security teams who optimize their security spend on managing risk versus purely threat, should not be surprised by Cloud. To these organizations, adapting their security programs to the challenges and opportunities provided by Cloud is business as usual. For organizations unprepared for Cloud, the maturity of security programs they can buy will quickly be outmoded.

Summary

The benefits of Cloud are many. The challenges are substantial. How we deal with these challenges and their organizational, operational, architectural, and technical impacts will fundamentally change the way in which we think about assessing and assuring the security of our assets.

Cloud Providers and Security “Edge” Services – Where’s The Beef?

September 30th, 2009 16 comments

usbhamburgerPreviously I wrote a post titled “Oh Great Security Spirit In the Cloud: Have You Seen My WAF, IPS, IDS, Firewall…” in which I described the challenges for enterprises moving applications and services to the Cloud while trying to ensure parity in compensating controls, some of which are either not available or suffer from the “virtual appliance” conundrum (see the Four Horsemen presentation on issues surrounding virtual appliances.)

Yesterday I had a lively discussion with Lori MacVittie about the notion of what she described as “edge” service placement of network-based WebApp firewalls in Cloud deployments.  I was curious about the notion of where the “edge” is in Cloud, but assuming it’s at the provider’s connection to the Internet as was suggested by Lori, this brought up the arguments in the post
above: how does one roll out compensating controls in Cloud?

The level of difficulty and need to integrate controls (or any “infrastructure” enhancement) definitely depends upon the Cloud delivery model (SaaS, PaaS, and IaaS) chosen and the business problem trying to be solved; SaaS offers the least amount of extensibility from the perspective of deploying controls (you don’t generally have any access to do so) whilst IaaS allows a lot of freedom at the guest level.  PaaS is somewhere in the middle.  None of the models are especially friendly to integrating network-based controls not otherwise supplied by the provider due to what should be pretty obvious reasons — the network is abstracted.

So here’s the rub, if MSSP’s/ISP’s/ASP’s-cum-Cloud operators want to woo mature enterprise customers to use their services, they are leaving money on the table and not fulfilling customer needs by failing to roll out complimentary security capabilities which lessen the compliance and security burdens of their prospective customers.

While many provide commoditized solutions such as anti-spam and anti-virus capabilities, more complex (but profoundly important) security services such as DLP (data loss/leakage prevention,) WAF, Intrusion Detection and Prevention (IDP,) XML Security, Application Delivery Controllers, VPN’s, etc. should also be considered for roadmaps by these suppliers.

Think about it, if the chief concern in Cloud environments is security around multi-tenancy and isolation, giving customers more comfort besides “trust us” has to be a good thing.  If I knew where and by whom my data is being accessed or used, I would feel more comfortable.

Yes, it’s difficult to do properly and in many cases means the Cloud provider has to make a substantial investment in delivery platforms and management/support integration to get there.  This is why niche players who target specific verticals (especially those heavily regulated) will ultimately have the upper hand in some of these scenarios – it’s not socialist security where “good enough” is spread around evenly.  Services like these need to be configurable (SELF-SERVICE!) by the consumer.

An example? How about Google: where’s DLP integrated into the messaging/apps platforms?  Amazon AWS: where’s IDP integrated into the VMM for introspection?

I wrote a couple of interesting posts about this (that may show up in the automated related posts lists below):

My customers in the Fortune 500 complain constantly that the biggest providers they are being pressured to consider for Cloud services aren’t listening to these requests — or aren’t in a position to respond.

That’s bad for everyone.

So how about it? Are services like DLP, IDP, WAF integrated into your Cloud providers’ offerings something you’d like to see rather than having to add additional providers as brokers and add complexity and cost back into Cloud?

/Hoff