How Would You Architect A New Security Monitoring Product?Cloud, appliance, software? If you were planning on developing a security monitoring platform, which architecture would you use?
I came up with a simple survey question earlier this month -- "If you were to develop a new Security Information and Event Management (SIEM) or Log Management product, how would you deliver it?" -- with the hopes that my survey would be short enough to entice people to answer. FluidSurveys was used as the survey platform for this exercise because it was easy to use and, most importantly, free.
I presented several options, including vendor-supplied appliance, software-only, prepackaged virtualized instance, cloud with multitenant SaaS port, and a hybrid option with some combination of appliance, software, virtualized instance, or cloud. At the time of this blog post, I've received 57 votes and 37 responses to justify the individual's selection. Using Twitter, I was able to reach the four corners of the globe and saw respondents originating from the United States, Canada, Mexico, Russia, China, India, Pakistan, Germany, France, and Spain.
As you can see from the resulting graph, the answers were really all over the board.
(click image for larger view)
The majority of respondents (46 percent) stated that if they were to develop a new SIEM or log management product, they'd design it as a hybrid platform that spanned multiple distribution and consumption methods. One respondent stated that the hybrid model "allows for different operating environments of the various customers which you will be coming across. It also allows for a more scalable customer base, ranging from SMB to large corporations."
The argument of supporting multiple and varying use cases also popped up in several comments. One comment that perhaps explains the hybrid model best was the following: "Locking yourself into one area will limit its usefulness. Sometimes I want a VM, sometimes I want an appliance, other times I want a software product. (I'm not personally into the whole cloud thing, but that's me)."
Some respondents (19 percent) still cling to vendor-supplied appliances, citing "more control as a vendor, less risk during deployment, maintenance, and support," and the "dependability, control of components" that come with shipping a box to the customer. A software model on user-selected hardware came in third-place with 18 percent of the votes. Respondents had several comments about this model, ranging from distrust of the cloud to cost savings on shipping and hardware costs to flexibility. A handful of respondents (12 percent) said they would lead with a VM-only packaging model, claiming that "an image is trivial to deploy and doesn't have the lock in or infrastructure requirements of an appliance," "flexibility" to adapt to a wide variety of environments, and "ease of deployment."
Perhaps what surprised me the most is that only 5 percent of respondents said they would develop their monitoring products using a multitenant SaaS model. With the proliferation of cloud platforms, in addition to how the cloud enables anyone in their garage to develop a product, I would have thought that the cloud-enabling low barrier to entry would have been a more popular selection. The one comment reported for this option was that the individual would want "Quick access to the most potential customers; lowest installation and setup costs; our ability to mitigate security concerns by encrypting everything in transit and at rest."
Certainly lots to think about. I guess the conclusion that I could draw as a result of this survey is that there is still no right answer to which security monitoring architecture is best -- and that would appeal to everyone.
Andrew Hay is senior analyst with 451 Research's Enterprise Security Practice (ESP) and is an author of three network security books. Follow him on Twitter: @andrewsmhay