VMware’s vShield – Why It’s Such A Pain In the Security Ecosystem’s *aaS…
I’ve become…comfortably numb…
Whilst attending VMworld 2011 last week, I attended a number of VMware presentations, hands-on labs and engaged in quite a few discussions related to VMware’s vShield and overall security strategy.
I spent a ton of time discussing vShield with customers — some who love it, some who don’t — and thought long and hard about writing this blog. I also spent some time on SiliconAngle’s The Cube discussing such, here.
I have dedicated quite a lot of time discussing the benefits of VMware’s security initiatives, so it’s important that you understand that I’m not trying to be overtly negative, nor am I simply pointing fingers as an uneducated, uninterested or uninvolved security blogger intent on poking the bear. I live this stuff…every day, and like many, it’s starting to become messy. (Ed: I’ve highlighted this because many seem to have missed this point. See here for example.)
It’s fair to say that I have enjoyed “up-to-the-neck” status with VMware’s various security adventures since the first marketing inception almost 4 years ago with the introduction of the VMsafe APIs. I’ve implemented products and helped deliver some of the ecosystem’s security offerings. My previous job at Cisco was to provide the engineering interface between the two companies, specifically around the existing and next generation security offerings, and I now enjoy a role at Juniper which also includes this featured partnership.
I’m also personal friends with many of the folks at VMware on the product and engineering teams, so I like to think I have some perspective. Maybe it’s skewed, but I don’t think so.
There are lots of things I cannot and will not say out of respect for obvious reasons pertaining to privileged communications and NDAs, but there are general issues that need to be aired.
Geez, enough with the CYA…get on with it then…
As I stated on The Cube interview, I totally understand VMware’s need to stand-alone and provide security capacities atop their platform; they simply cannot expect to move forward and be successful if they are to depend solely on synchronizing the roadmaps of dozens of security companies with theirs.
However, the continued fumbles and mis-management of the security ecosystem and their partnerships as well as the continued competitive nature of their evolving security suite makes this difficult. Listening to VMware espouse that they are in the business of “security ecosystem enablement” when there are so few actually successful ecosystem partners involved beyond antimalware is disingenuous…or at best, a hopeful prediction of some future state.
Here’s something I wrote on the topic back in 2009: The Cart Before the Virtual Horse: VMware’s vShield/Zones vs. VMsafe APIs that I think illustrates some of the issues related to the perceived “strategy by bumping around in the dark.”
A big point of confusion is that vShield simultaneously describes both an ecosystem program and a set of products that is actually more than just anti-malware capabilities which is where the bulk of integration today is placed.
Analysts and journalists continue to miss the fact that “vShield” is actually made up of 4 components (not counting the VMsafe APIs):
- vShield Edge
- vShield App
- vShield Endpoint
- vShield Manager
What most people often mean when they refer to “vShield” are the last two components, completely missing the point that the first two products — which are now monetized and marketed/sold as core products for vSphere and vCloud Director — basically make it very difficult for the ecosystem to partner effectively since it’s becoming more difficult to exchange vShield solutions for someone else’s.
An important reason for this is that VMware’s sales force is incentivized (and compensated) on selling VMware security products, not the ecosystem’s — unless of course it is in the way of a big deal that only a partnership can overcome. This is the interesting juxtaposition of VMware’s “good enough” versus incumbent security vendors “best-of-breed” product positioning.
VMware is not a security or networking company and ignoring the fact that big companies with decades of security and networking products are not simply going to fade away is silly. This is true of networking as it is security (see software-defined networking as an example.)
Technically, vShield Edge is becoming more and more a critical piece of the overall architecture for VMware’s products — it acts as the perimeter demarcation and multi-tenant boundary in their Cloud offerings and continues to become the technology integration point for acquisitions as well as networking elements such as VXLAN.
As a third party tries to “integrate” a product which is functionally competitive with vShield Edge, the problems start to become much more visible and the partnerships more and more clumsy, especially in the eyes of the most important party privy to this scenario: the customer.
Jon Oltsik wrote a story recently in which he described the state of VMware’s security efforts: “vShield, Cloud Computing, and the Security Industry”
So why aren’t more security vendors jumping on the bandwagon? Many of them look at vShield as a potentially competitive security product, not just a set of APIs.
In a recent Network World interview, Allwyn Sequeira, VMware’s chief technology officer of security and vice president of security and network solutions, admitted that the vShield program in many respects “does represent a challenge to the status quo” … (and) vShield does provide its own security services (firewall, application layer controls, etc.)
Why aren’t more vendors on-board? It’s because this positioning of VMware’s own security products which enjoy privileged and unobstructed access to the platform that ISV’s in the ecosystem do not have. You can’t move beyond the status quo when there’s not a clear plan for doing so and the past and present are littered with the wreckage of prior attempts.
VMware has its own agenda: tightly integrate security services into vSphere and vCloud to continue to advance these platforms. Nevertheless, VMware’s role in virtualization/cloud and its massive market share can’t be ignored. So here’s a compromise I propose:
- Security vendors should become active VMware/vShield partners, integrate their security solutions, and work with VMware to continue to bolster cloud security. Since there is plenty of non-VMware business out there, the best heterogeneous platforms will likely win.
- VMware must make clear distinctions among APIs, platform planning, and its own security products. For example, if a large VMware shop wants to implement vShield for virtual security services but has already decided on Symantec (Vontu) or McAfee DLP, it should have the option for interoperability with no penalties (i.e., loss of functionality, pricing/support premiums, etc.).
Item #1 Sounds easy enough, right? Except it’s not. If the way in which the architecture is designed effectively locks out the ecosystem from equal access to the platform except perhaps for a privileged few, “integrating” security solutions in a manner that makes those solutions competitive and not platform-specific is a tall order. It also limits innovation in the marketplace.
Look how few startups still exist who orbit VMware as a platform. You can count them on less fingers that exist on a single hand. As an interesting side-note, Catbird — a company who used to produce their own security enforcement capabilities along with their strong management and compliance suite — has OEM’d VMware’s vShield App product instead of bothering to compete with it.
Now, item #2 above is right on the money. That’s exactly what should happen; the customer should match his/her requirements against the available options, balance the performance, efficacy, functionality and costs and ultimately be free to choose. However, as they say in Maine…”you can’t get there from here…” at least not unless item #1 gets solved.
In a complimentary piece to Jon’s, Ellen Messmer writes in “VMware strives to expand security partner ecosystem“:
Along with technical issues, there are political implications to the vShield approach for security vendors with a large installed base of customers as the vShield program asks for considerable investment in time and money to develop what are new types of security products under VMware’s oversight, plus sharing of threat-detection information with vShield Manager in a middleware approach.
…and…
The pressure to make vShield and its APIs a success is on VMware in some respects because VMware’s earlier security API , the VMsafe APIs, weren’t that successful. Sequiera candidly acknowledges that, saying, “we got the APIs wrong the first time,” adding that “the major security vendors have found it hard to integrate with VMsafe.”
Once bitten, twice shy…
So where’s the confidence that guarantees it will be easier this time? Basically, besides anti-malware functionality provided by integration with vShield endpoint, there’s not really a well-defined ecosystem-wide option for integration beyond that with VMware now. Even VMware’s own roadmaps for integration are confusing. In the case of vCloud Director, while vShield Edge is available as a bundled (and critical) component, vShield App is not!
Also, forcing integration with security products now to directly integrate with vShield Manager makes for even more challenges.
There are a handful of security products besides anti-malware in the market based on the VMsafe APIs, which are expected to be phased out eventually. VMware is reluctant to pin down an exact date, though some vendors anticipate end of next year.
That’s rather disturbing news for those companies who have invested in the roadmap and certification that VMware has put forth, isn’t it? I can name at least one such company for whom this is a concern. 🙁
Because VMware has so far reserved the role of software-based firewalls and data-loss prevention under vShield to its own products, that has also contributed to unease among security vendors. But Sequiera says VMware is in discussions with Cisco on a firewall role in vShield. And there could be many other changes that could perk vendor interest. VMware insists its vShield APIs are open but in the early days of vShield has taken the approach of working very closely with a few selected vendors.
Firstly, that’s not entirely accurate regarding firewall options. Cisco and Juniper both have VMware-specific “firewalls” on the market for some time; albeit they use different delivery vehicles. Cisco uses the tightly co-engineered effort with the Nexus 1000v to provide access to their VSG offering and Juniper uses the VMsafe APIs for the vGW (nee’ Altor) firewall. The issue is now one of VMware’s architecture for integrating moving forward.
Cisco has announced their forthcoming vASA (virtual ASA) product which will work with the existing Cisco VSG atop the Nexus 1000v, but this isn’t something that is “open” to the ecosystem as a whole, either. To suggest that the existing APIs are “open” is inaccurate and without an API-based capability available to anyone who has the wherewithal to participate, we’ll see more native “integration” in private deals the likes of which we’re already witnessing with the inclusion of RSA’s DLP functionality in vShield/vSphere 5.
Not being able to replace vShield Edge with an ecosystem partner’s “edge” solution is really a problem.
In general, the potential for building a new generation of security products specifically designed for VMware’s virtualization software may be just beginning…
Well, it’s a pretty important step and I’d say that “beginning” still isn’t completely realized!
It’s important to note that these same vendors who have been patiently navigating VMware’s constant changes are also looking to emerging competitive platforms to hedge their bets. Many have already been burned by their experience thus far and see competitive platform offerings from vendors who do not compete with their own security solutions as much more attractive, regardless of how much marketshare they currently enjoy. This includes community and open source initiatives.
Given their druthers, with a stable, open and well-rounded program, those in the security ecosystem would love to continue to produce top-notch solutions for their customers on what is today the dominant enterprise virtualization and cloud platform, but it’s getting more frustrating and difficult to do so.
It’s even worse at the service provider level where the architectural implications make the enterprise use cases described above look like cake.
It doesn’t have to be this way, however.
Jon finished up his piece by describing how the VMware/ecosystem partnership ought to work in a truly cooperative manner:
This seems like a worthwhile “win-win,” as that old tired business cliche goes. Heck, customers would win too as they already have non-VMware security tools in place. VMware will still sell loads of vShield product and the security industry becomes an active champion instead of a suspicious player in another idiotic industry concept, “coopitition.” The sooner that VMware and the security industry pass the peace pipe around, the better for everyone.
The only thing I disagree with is how this seems to paint the security industry as the obstructionist in this arms race. It’s more than a peace pipe that’s needed.
Puff, puff, pass…it’s time for more than blowing smoke.
/Hoff
Related articles
- VMware’s (New) vShield: The (Almost) Bottom Line (rationalsurvivability.com)
- VMware strives to expand security partner ecosystem (infoworld.com)
- VMware preparing data loss prevention features for vShield (infoworld.com)
- Highlights from VMworld 2011 (datacenterknowledge.com)
- Clouds, WAFs, Messaging Buses and API Security… (rationalsurvivability.com)
- AWS’ New Networking Capabilities – Sucking Less
(rationalsurvivability.com) - Sourcefire Enables Application Control Within Virtual Environments (it-sideways.com)
- Using The Cloud To Manage The Cloud (informationweek.com)
- Virtualizing Your Appliance Is Not Cloud Security (securecloudreview.com)
- OpenFlow & SDN – Looking forward to SDNS: Software Defined Network Security(rationalsurvivability.com)
I don’t disagree that web applications/web services are complex. I further don’t disagree that protecting the web services and messaging buses that make up the majority of the exposed interfaces in vCloud Director don’t require sophisticated protection.
It is so tempting to write about 
My wife is in the midst of an extended multi-phasic, multi-day delivery process of our fourth child. In between bouts of her moaning, breathing and ultimately sleeping, I’m left to taunt people on Twitter and think about Cloud.
The investment and skillsets needed to rectify two often diametrically-opposed operational models doesn’t maximize returns, it bifurcates and diminishes efficiencies and blurs cost allocation models making both internal IT and public cloud look grotesquely inaccurate.
Ultimately, the reason I agree so strongly with this is because of the architectural, operational and compliance complexity associated with all the mechanics one needs to allow for interoperable, scaleable, secure and manageable workloads between an internal enterprise’s operational domain (cloud or otherwise) and the public cloud.
If your Tier-1 workloads can run in a public cloud and satisfy all your requirements, THAT’S where they should run in the first place! You maximize your investment internally by scaling down and ruthlessly squeezing efficiency out of what you have as quickly as possible — writing those investments off the books.
…to which I responded that clearly there is room for improvement in L3+ and security. I expect we’ll see some 😉
The feasibility for this vector was deftly demonstrated shortly afterward by the guys at SensePost (
/Hoff
[Like a good w[h]ine, this post goes especially well with a couple of other courses such as
Recent Comments