Rational Survivability

I'll be updating my speaking itinerary shortly, but I wanted to let folks know I'm working on three major VirtSec/CloudSec presentations for 2009:
 


The Frogs Who Desired a King

The Frogs Who Desired a King is based on the topical reference to one of Aesop's fable about a discontented population of frogs who appealed to Zeus for a king.

Ultimately, through a comedy of errors, the frogs finally got their new king — a stork — which promptly ate them.

We, as a discontented legion of frogs, decry our dark overlords' choices (or lack thereof) of security in virtualized and cloud computing environments and long for security solutions to magically solve all our problems. 

Just like the frogs, we better be careful what we wish for, as our prayers might just be answered, consuming us all. This is the sequel to my "Four Horsemen of the Virtualization Security Apocalypse" series.

Cloudifornication: Indiscriminate Information Intercourse Involving Internet Infrastructure

What was in is now out. 

This metaphor holds true not only as an accurate analysis of adoption trends of disruptive technology and innovation, but also parallels the amazing velocity of how our datacenters are being re-perimiterized and quite literally turned inside out.

One of the really scary things happening with the massive convergence of cloud computing is its effect on security models and the information they seek to protect.

Where and how our data is created, processed, accessed, stored, backed up in what is sure to become massively overlaid cloud-based services — and by whom and whose infrastructure — yields significant concerns related to security, privacy, compliance and survivability. 

This "infrastructure intercourse" makes it very interesting to try and secure your assets when you don't own the infrastructure and in many cases cannot provide the same levels of functionality we can today.

Mozart's "The Marriage of Figaro": Complexity & Insecurity Of the Cloud

Mozart's sequel to the Barber of Seville was lauded as one of the most profound works of its time. 

Its staggering complexity, inviting overtures, rich textures and variety of orchestration were perceived by many as unapproachable, unfathomable and in some cases unintelligible. 

Yet so remarkable and unique was the composition that people flocked to its performances although in many cases were blinded by the simplicity of its underlying complexity.

Such are the parallels with the deeply profound cacophony surrounding the issues of securing Cloud Computing and the tonal miscues hidden amongst its various acts.

This presentation will review the most pressing security, privacy, sustainability and resiliency
issues surrounding the marriage of convenience, economics and computing.

Here's a theme I've been banging around for quite some time as it relates to virtualization, cloud computing and security.  I've never really sat down and written about it, however.

As we trend towards consolidating and (re)centralizing our computing platforms — both endpoints and servers — using virtualization and cloud computing as enablers to do so, we're also simultaneously dealing with the decentralization and distributed data sets that come with technologies such as Web2.0, mobility and exposure of APIs from cloud platforms.*

So here we are all frothed up as virtualization and cloud computing have, in a sense, led us back to the resource-based consolidation of the mainframe model with all it's centralized splendor and client virtualization/thin clients/compartmentalized remote access is doing the same thing for endpoints. 

But the interesting thing is that with Moore's Law, the endpoints are also getting more and more powerful even though we're dumbing them down and trying to make their exposure more limited despite the fact that they can still efficiently process and store data locally.

These models, one could argue, are diametrically opposed when describing how to secure the platforms versus the information that resides on or is utilized by them.  As the cyclic waffling between centralized versus distributed continues, the timing of how and where we adapt to securing them always lags behind.  Which do we focus on securing and where?  The host, centralized server, network.

The unfortunate answer is always "yes."

Remember this (simplified) model of how/where we secure things?


If you juxtapose the image above mentally with how I represent the centralized <–> distributed trends in IT below, it's no wonder we're always behind the curve.  The computing model technology changes much more quickly than the security technology and processes do, thus the disconnect:

I need to update the diagram above to split out the "computing" layer
into client and server as well as extend the data layer to reference
storage modalities also, but it gets the job done.

At any rate, it's probably obvious and common sense, but when explaining to people why I spend my time pointing out gaps with security in virtualization and cloud models, I found this useful.

/Hoff

* It's important to note that while I refer to/group cloud computing models as centralized, I understand they have a distributed element to them, also.  I would ask you to think about the multiple cloud overlays as centralized resources, regardless of how intrinsically "distributed" in processing/load balancing they may be.

P.S. I just saw an awesome post titled "The Rise of the Stupid Endpoint" on the vinternals blog that shares many of the same points, although much more eloquently.  Check it out here.  Awesome!

In honor of my friend @quine on Twitter who today complained thusly:


In case you're reading this with Lynx (you web pimp, you!,) Zach was lamenting the fact that vendors who don't support customer operating systems of choice are simply sloughing off development efforts and support by suggesting that customers should simply run it as a VM instead.

Ah, it used to be called "software," but now it's a "virtual appliance!"  Silly rabbit, tricks are for kids.

One might suggest this is a perfectly reasonable use of virtualization technology — neigh one of the very purposes behind its genesis.  I'd agree, to a point.  However, I've noticed an alarming uptake recently by product managers who are simply short-cutting roadmap/development paths by taking the "lazy" way out.

Hey, it cuts down support, testing, regression and troubleshooting…for the vendor.  But in my favorite commentary, it's simply a "squeezing the balloon problem" because it surfaces a whole host of other issues such as performance, scale, and in some cases support for various virtualization platforms.

What say you?  Do you see this happening more in your enterprise?  Do you care?  Is it a good thing?

/Hoff

This post contains nothing particularly insightful other than a pronounced giant sucking sound that's left a vacuum in terms of forward motion regarding security and virtualization.

From the desk of Cisco's "Virtualization Wow!" Campaign: When is a switch with a server not a virtualization platform?  When it's a server with a switch as a virtualization platform, of course! 😉

I can't help but smile at the announcement that Cisco is bringing to market a blade-based chassis which bundles together Intel's Nehalem-based server processors, the Nexus 5000 switch, and VMware's virtualization and management platform.  From InformationWeek:

This news was actually broken back in the beginning of December by virtualization.info but I shamefully missed it.  It looked like a bunch of others did, too.

This totally makes sense as virtualization has driven convergence across the compute, network and storage realms and has highlighted the fact that the provisioning, automation, and governance — up and down the stack — demands a unified approach for management and support.

For me, this is the natural come-about of what I wrote about in July of 2007 in a blog post titled "Cisco & VMware – The Revolution Will Be…Virtualized?":

This [convergence of network, compute and virtualization, Ed.] is interesting for sure and if you look at the way in which the
demand for flexibility of software combined with generally-available
COTS compute stacks and specific network processing where required, the
notion that Cisco might partner with VMWare or a similar vendor such as
SWSoft looks compelling.  Of course with functionality like KVM in the Linux kernel, there's no reason they have to buy or ally…

Certainly there are already elements of virtualization within
Cisco's routing, switching and security infrastructure, but many might
argue that it requires a refresh in order to meet the requirements of
their customers.  It seems that their CEO does.

When I last blogged about Cisco's partnership with VMware and (what is now called) the Nexus 1000v/VN-Link, I made reference to the fact that I foresaw the extraction of the VM's from the servers and suggested that we would see VM's running in the Nexus switches themselves.  Cisco representatives ultimately put a stake in the sand and said this would never happen in the comments of that blog post.

Now we know what they meant and it makes even more sense.

So the bundling of the Nexus 5000* (with the initiator,) the upcoming protocol for VM-Flow affinity tagging, the integrated/converged compute and storage capabilities, and Intel's SR-IOV/MR-IOV/IOMMU technologies in the Nehalem, all supported by the advances with vNetworking/VN-Link makes this solution a force to be reckoned with.

Other vendors, especially those rooted in servers and networking such as HP and IBM, are likely to introduce their own solutions, but given the tight coupling of the partnership, investment and technology development between Intel, VMware and Cisco, this combo will be hard to beat. 

Folks will likely suggest that Cisco has no core competency in building, selling or supporting "servers," but given the channel and partnership with VMware — with virtualization abstracting that hardware-centric view in the first place — I'm not sure this really matters.  We'll have to see how accurate I am on this call.

Regardeless of the semantic differences of where the CPU execution lives (based on my initial prediction,) all the things I've been talking about that seemed tangentially-related but destined to come together seem to have.  Further, here we see the resurgence (or at least redefinition of Big Iron, all over again…)

Remember the Four Horsemen slides and the blog post (the network is the computer, is the network, is the…) where I dared you to figure out where "the network" wasn't in the stack?  This is an even more relevant question today. 

It's going to be a very interesting operational and organizational challenge from a security perspective when your server, storage, networking and virtualization platform all come from a single source.

California Dreamin'…

/Hoff

* Not that the 1000v ain't cool, but that little slide that only appeared once at VMworld about the 5000v and the initiator was just too subtely delicious not to be the real juice in the squeeze. The 1000v obviously has its place and will be a fantastic solution, but for folks who are looking for a one-stop shop for their datacenter blueprint designs heavily leveraging virtualization, this makes nothing but sense.

I keep hearing about the impending doom of (specifically) rogue VM sprawl — our infrastructure overrun with the unchecked proliferation of virtual machines running amok across our enterprises.  Oh the horror!

Most of the examples use the consolidation of server VM's onto hosts as delivered by virtualization as their example.

I have to ask you though, given what it takes to spin up a VM on a platform such as VMware, how can you have a "rogue" VM sprawling its way across your enterprise!? 

Someone — an authorized administrator — had to have loaded it into inventory, configured its placement on a virtual switch, and spun it up via VirtualCenter or some facsimile thereof depending upon platform.

That's the definition of a rogue?  I can see where this may be a definitional issue, but the marketeers are getting frothed up over this very issue, whispering in your ear constantly about the impending demise of your infrastructure…and undetectable hypervisor rootkits, too.  🙂

It may be that the ease of which a VM *can* be spun up legitimately can lead to the overly-exhuberant deployment of VM's without understanding the impact this might have on the infrastructure, but can we please stop grouping stupidity and poor capacity planning/impact analysis with rogueness?  They're two different things.

If administrators are firing off VMs that are unauthorized, unhardened, and unaccounted for, you have bigger problems than that of virtualization and you ought to consider firing them off. 

The inventory of active VMs is a reasonably easy thing to keep track of; if it's running, I can see it. 

I know "where" it is and I can turn it off.  To me, the bigger problem is represented by the offline VMs which can live outside that inventory window, just waiting to be reactivated from their hypervisorial hibernation.

But does that represent "rogue?"

You want an example of a threat which represents truly rogue VM "sprawl" that people ought to be afraid of?  OK, here's one, and it happened to me.  I talk about it all the time and people usually say "Oh, man, I never thought of that…"  usually because we're focused on server virtualization and not the client side.

Now, that's a rogue VM.  An accidental rogue VM.  Imagine if it were an attacker.  Perhaps he/she was a legitimate user but disgruntled.  Perhaps he/she decided to use wireless instead of wired.  How much fun would that be?

Stop with the "rogue (server) VM" FUD, wouldya?

Kthxbye.

/Hoff

Here's a nod to Rich Miller who pointed over (route the node, not the packet) to a blog entry from Andreas Antonopoulos titled "Virtual Routing – The anti-matter of network routing."

The premise, as brought up by Doug Gourlay from Cisco at the C-Scape conference was seemingly innocuous but quite cool:

"How about using netflow information to re-balance servers in a data center"

Routing: Controlling the flow of network traffic to an optimal path between two nodes

Virtual-Routing or Anti-Routing: VMotioning nodes (servers) to optimize the flow of traffic on the network.

Using netflow information, identify those nodes (virtual servers)
that have the highest traffic "affinity" from a volume perspective (or
some other desired metric, like desired latency etc) and move (VMotion,
XenMotion) the nodes around to re-balance the network. For example,
bring the virtual servers exchanging the most traffic to hosts on the
same switch or even to the same host to minimize traffic crossing
multiple switches. Create a whole-data-center mapping of traffic flows,
solve for least switch hops per flow and re-map all the servers in the
data center to optimize network traffic.

My first reaction was, yup, that makes a lot of sense from a network point of view, and given who made the comment, it does make sense. Then I choked on my own tongue as the security weenie in me started in on the throttling process, reminding me that while this is fantastic from an autonomics perspective, it's missing some serious input variables.

Latency of the "network" and VM spin-up aside, the dirty little secret is that what's being described here is a realistic and necessary component of real time (or adaptive) infrastructure.  We need to get ultimately to the point where within context, we have the ability to do this, but I want to remind folks that availability is only one leg of the stool.  We've got the other nasty bits to concern ourselves with, too.

Let's look at this from two perspectives: the network plumber and the security wonk

From the network plumbers' purview, this sounds like an awesome idea; do what is difficult in non-virtualized environments and dynamically adjust and reallocate the "location" of an asset (and thus flows to/from it) in the network based upon traffic patterns and arbitrary metrics.  Basically, optimize the network for the lowest latency and best performance or availability by moving VM's around and re-allocating them across the virtual switch fabric (nee DVS) rather than adjusting how the traffic gets to the static nodes.

It's a role reversal: the nodes become dynamic and the network becomes more static and compartmentalized.  Funny, huh?

—

The security wonk is unavailable for comment.  He's just suffered a coronary event.  Segmented network architecture based upon business policy, security, compliance and risk tolerances make it very difficult to perform this level of automation via service governors today, especially in segmented network architecture based upon asset criticality, role or function as expressed as a function of (gulp) compliance, let's say. 

Again, the concept works great in a flat network where asset grouping is, for the most part, irrelevant (hopefully governed by a policy asserting such) where what you're talking about is balancing the compute with network and storage, but the moment you introduce security, compliance and risk management as factors into the decision fabric, things get very, very difficult.

Now, if you're Cisco and VMware, the
models for how the security engines that apply policy consistently
across these fluid virtualized networks is starting to take shape, but what we're
missing are the set of compacts or contracts that consistently define
and enforce these policies no matter where they move (and control *if* they can move) and how they factor these requirements into
the governance layer.

The standardization of governance approaches — even at the network layer — is lacking. 
There are lots of discrete tools available but the level of integration
and the input streams and output telemetry are not complete.

If you take a look, as an example, at CIRBA's exceptional transformational analytics and capacity management solution, replete with their multi-dimensional array of business process, technical infrastructure and resource mapping, they have no input for risk assessment data, compliance or "security" as variables.

When you look at the utility brought forward by the dynamic, agile and flexible capabilities of virtualized infrastructure, it's hard not to extrapolate all the fantastic things we could do. 

Unfortunately, the crushing weight of what happens when we introduce security, compliance and risk management to the dance means we have a more sobering discussion about those realities.

Here's an example reduced to the ridiculous: we have an interesting time architecting networks to maximize throughput, reduce latency and maximize resilience in the face of what can happen with convergence issues and flapping when we have a "routing" problem.

Can you imagine what might happen when you start bouncing VM's around the network in response to maximizing efficiency while simultaneously making unavailable the very resources we seek to maximize the availability of based upon disassociated security policy violations?  Fun, eh?

While we're witnessing a phase shift in how we design and model our networks to support more dynamic resources and more templated networks, we can't continue to mention the benefits and simply assume we'll catch up on the magical policy side later.

So for me, Virtual Routing is the anti-matter of network SECURITY, not network routing…or maybe more succinctly, perhaps security doesn't matter at all?

/Hoff

I just read the latest blog of Alistair Croll from GigaOm titled "Cloud Security: The Sky Is Falling!" in which he suggests that we pillow-hugging security wonks ought to loosen our death grips on our data because not only are we flapping our worry feathers for nothing, but security in "the Cloud" will result in better security than we have today. 

As "the Cloud" provider adds customers, the capability to secure the infrastructure and the data transiting it, ultimately becomes an issue of scale, too. The more automation that is added, the more false positives show up, especially in light of the fact that the service provider has little or no context of the information, business processes or business impact that their monitoring tools observe.  You can get rid of the low-hanging fruit, but when it comes down to impacting the business, some human gets involved.

Read more…

I'm working on the sequel to my Four Horsemen of the Virtualization Security Apocalypse presentation.

It's called "The Frogs Who Desperately Wanted a King: An Information Security Fable of Virtualization, RTI and Cloud Computing Security." (Okay, it also has the words "interpretive dance" in it, but that's for another time…)

Many of the interesting issues from the Four Horsemen regarding the deficiencies of security solutions and models in virtualized environments carries over directly to operationalizing security in the Cloud. 

As a caveat, let's focus on a cost-burdened "large" enterprise who's involved in moving from physical to virtual to cloud-based services.

I'm not trying to make a habit of picking on Amazon AWS, but it's just such a fantastic example for my point, which is quite simply:

Why?  Let me paint the picture…

In non-virtualized environments, we generally use dedicated appliances or integrated solutions that provide one of more discrete security functions. 

These solutions are built generally on hardened OS's, sometimes using custom hardware and sometimes COTS boxes which are tarted up.  They are plumbed in between discretely segregated (physical or logical) zones to provide security boundaries defined by arbitrary policies based upon asset classification, role, function, access, etc.  We've been doing this for decades.  It's the devil we know.

In virtualized environments, we currently experience some real pain when it comes to replicating the same levels of security, performance, resiliency, and scale using software-based virtual appliances to take the place of the hardware versions in our physical networks when we virtualize the interconnects within these zones.

There are lots of reasons for this, but the important part is realizing that many of the same hardware solutions are simply not available as virtual appliances and even when they are, they are often not 1:1 in terms of functionality or capability.  Again, I've covered this extensively in the Four Horsemen.

So if we abstract this to its cloudy logical conclusion, and use AWS as the "platform" example, we start to face a real problem for an enterprise that has a decade(s) of security solutions, processes and talent that is focused on globalizing, standardizing and optimizing their existing security infrastructure and are now being forced to re-evaluate not only the technology selection but the overall architecture and operational model to support it.

Now, it's really important that you know, dear reader, that I accept that one can definitely deploy security controls instantiated as both network and host-based instances in AWS.  There are loads of options, including the "firewall" provided by AWS.

However, the problem is that in the case of building an AMI for AWS supporting a particular function (firewall, WAF, IPS, IDS, etc.) you may not have the same solutions available to you given the lack of support for a particular distro, lack of "port" to a VA/VM, or issues surrounding custome kernels, communication protocols, hardware, etc…  You may be limited in many cases to having to rely on open source solutions.

In fact, when one looks at most of the examples given when describing securing AWS instances, they almost always reference OSS solutions such as Snort, OSSEC, etc.  There's absolutely NOTHING wrong with that, but it's only one dimension.

That's going to have a profound effect across many dimensions.  In many cases, enterprises have standardized on a particular solution not because it's special from a "security" perspective, but because it's the easiest to manage when you have lots of them and they are supportable 24/7 by vendors with global reach and SLA's that represent the criticality of their function.

That is NOT to say that OSS solutions are not easy to manage or supportable in this fashion, but I believe it's a valid representation of the state of things.

(Why am I getting so defensive about OSS? 😉

Taking it further, and using my favorite PCI in the Cloud argument, what if the web application firewall that you've spent hundreds of thousands of dollars purchasing, tuning and deploying in support of PCI DSS throughout the corporate enterprise simply isn't available as a software module installable in an AMI in the cloud?  Or the firewall?  Or the IPS? 

In the short term this is a real problem for customers.  In the long term, it's another potential life preserver for security ISV's and an opportunity for emerging startups to think about new ways of solving this problem.

/Hoff

I read James Urquhart's first blog post written under the Cisco banner today titled "The network: the final frontier for cloud computing" in which he describes the evolving role of "the network" in virtualized and cloud computing environments.

The gist of his post, which he backs up with examples from Greg Ness' series on Infrastructure 2.0, is that in order to harness the benefits of virtualization and cloud computing, we must automate; from the endpoint to the underlying platforms — including the network — manual processes need to be replaced by automated capabilities:

When was the last time you thought “network” when you heard
“cloud computing”? How often have you found yourself working out
exactly how you can best utilize network resources in your cloud
applications?  Probably never, as to date the network hasn’t registered
on most peoples’ cloud radars.

This is understandable, of course, as the early cloud efforts try to
push the entire concept of the network into a simple “bandwidth”
bucket.  However, is it right? Should the network just play dumb and
let all of the intelligence originate at the endpoints?
…
The writing is on the wall. The next frontier to get explored in
depth in the cloud world will be the network, and what the network can
do to make cloud computing and virtualization easier for you and your
organization

If you walked away from James' blog as I did initially, you might be left with the impression that this isn't really about "the network" gaining additional functionality or innovative capabilities, but rather just tarting up the ability to integrate with virtualization platforms and automate it all.

Doesn't really sound all that sexy, does it.  Well, it's really not, which is why even today in non-virtualized environments we don't have very good automation and most processes still come down to Bob at the helpdesk. Virtualization and cloud are simply giving IT a swift kick in the ass to ensure we get a move on to extract as much efficiency and remove as much cost from IT as possible.

Don't be fooled by the simplicity of James' post, however, because there's a huge moose lurking under the table instead of on top of it and it goes toward the fundamental crux of the battle brewing between all those parties interested in becoming your next "datacenter OS" provider.

There exists one catalytic element that produces very divergent perspectives in IT around what, where, why and who automates things and how, and that's the very definition of "the network" in virtualized and cloud models.

How someone might describe "the network" as either just a "bandwidth bucket" of feeds and speeds or an "intelligent, aware, sentient platform for service delivery" depends upon whether you're really talking about "the network" as a subset or a superset of "the infrastructure" at large.

Greg argues that core network services such as IP adddress management, DNS, DHCP, etc. are part of the infrastructure and I agree, but given what we see today, I would say that they are part-in-parcel NOT a component of "the network" — they're generally separate and run atop the plumbing.  There's interaction, for sure, but one generally relies upon these third party service functions to deliver service.  In fact, that's exactly the sort of thing that Greg's company, Infoblox, sells.

This contributes to part of this definitional quandary.

Now we have this new virtualization layer injected between the network and the rest of the infrastructure which provides a true lever and frictionless capability for some of this automation but further confuses the definition of "the network" since so much of the movement and delivery of information is now done at this layer and it's not integrated with the traditional hardware-based network.*

See what I mean in this post titled "The Network Is the Computer…(Is the Network, Is the Computer…)"

This is exactly why you see Cisco's investment in bringing technologies such as VN-Link and the Nexus-1000v virtual switch to virtualized environments; it homogenizes "the network." It claws back the access layer so they can allow the network teams to manage the network again (and "automate" it) while also getting their hooks deeper into the virtualization layer itself. 

And that's where this gets interesting to me because in order to truly automate virtualized and cloud computing environments, this means one of three things as it relates to where core/critical infrastructure services live:

1. They  will continue to be separate as stand-alone applications/appliances or bundled atop an OS
2. They become absorbed by "the (traditional) network" and extend into the virtualization layer
3. They get delivered as part of the virtualization layer

So if you're like most folks and run Microsoft-based "core network services" for things (at least internally) like DNS, DHCP, etc., what does this mean to you?  Well, either you continue as-is via option #1, you transition to integrated services in "the network" via option #2 or you end up with option #3 by the very virtue that you'll upgrade to Windows Server 2008 and Hyper-V anyway.

SO, this means that the level of integration between, say, Cisco and Microsoft will have to become as strong as it is with VMware in order to support the integration of these services as a "network" function, else they'll continue — in those environments at least — as being a "bandwidth bucket" that provides an environment that isn't really automated.

In order to hit the sweet spot here, Cisco (and other network providers) need to then start offering core network services as part of "the network."  This means wrestling it away from the integrated OS solutions or simply buying their way in by acquiring and then integrating these services ($10 Cisco buys Infoblox…)

We also see emerging vendors such as Arista Networks who are entering the grid/utility/cloud computing network market with high density, high-throughput, lower cost "cloud networking" switches that are more about (at least initially) bandwidth bucketing and high-speed interconnects rather than integrated and virtualized core services.  We'll see how the extensibility of Arista's EOS affects this strategy in the long term.

There *is* another option and that's where third party automation, provisioning, and governance suites come in that hope to tame this integration wild west by knitting together this patchwork of solutions. 

What's old is new again.

/Hoff

*It should be noted, however, that not all things can or should be
virtualized, so physical non-virtualized components pose another
interesting challenge because automating 99% of a complex process isn't
a win if the last 1% is a gating function that requires human
interaction…you haven't solved the problem, you've just made it less
steps that still requires Bob at the helpdesk..