Dark Reading is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Security Management

11/8/2017
01:30 PM
Simon Marshall
Simon Marshall
Simon Marshall
50%
50%

SlashNext Seeks an End to the Sandbox

Sandboxing isn't effective in the modern world, according to startup SlashNext. But what, precisely, comes next?

The first two phases of threat detection development have been highly successful in their time. First, signatures of common attacks were used to close the door. Then sandbox identified threats in a virtual environment and locked them out. But those two technologies are now past their consume-by date.

According to Jack Miller, chief information security officer at SlashNext, a Pleasanton, Calif.-based firm, they're flawed. Few people would argue with that, but the risk seems to lie in continued use of them as security challenges evolve. So, will those technologies be superseded any time soon?

"Yes, [but] while our current network approach will not eliminate the need for an endpoint agent or an email solution, eventually those solutions will be built using the same detection methods that we are using on the network."

By early 2014, the firm claims, both signatures and sandboxes were consistently failing to detect polymorphic malware (which morphs several times during its lifetime to evade detection) and exploits. Apparently, because these technologies were not really designed to detect social engineering, phishing and callback attacks, targets were left wide-open. Social engineering and phishing attacks represent about 45% of all Internet access threats, nearly twice the number for malware and viruses, according to the regular Verizon Data Breach Digest.

"We realized that while tools will sometimes miss attacks, the top researchers would always find them eventually," said Miller. "So, we built a system that mimicked how teams of the top security researchers perform their analysis."

This system plays into the current trend of human judgement playing an equal role to security technologies which remove time-consuming crunch work. SlashNext has just launched a new approach which uses "reasoning engines" -- not machine learning -- which it claims are the equivalent of virtual analyst teams, operating in a specific forensic domain. The engines learn like humans, automating the detection process and providing its carbon-based counterparts with enough evidence to make a final decision about whether a suspected threat is likely real or not.

In a typically human activity, team members get together to vote on the evidence, and interestingly, are asked to make a black or white yes/no conclusion. "Because we have so many virtual teams looking at so many different forensic aspects of the session, we are able to increase each false negative rate for each domain, effectively lowering our false positive rate knowing that even if some of the domains miss the attack, others will still catch it," explains Miller.

He says that because analysis is performed across initial and subsequent sessions, even if an attack is missed first time, teams will be able to catch it in a later lifecycle. Analysts learn and adapt to new threats, while the number of false positives is reduced.

The concept uses an omnichannel tactic, specifically tackling attacks initiated by phishing, rather than malware itself. The idea is to cut the effectiveness of a single threat that multiplies its chance of success by initiating through multiple devices. If person A is induced to offer up bank login details through a false web page on a PC, SlashNext detects it and provides blocking here, and also to person B who is using their mobile device.

SlashNext is pitting this technology directly against current sandbox virtual machines which it finds unsuitable for several reasons; examples are that sandboxes are difficult to maintain a single software profile that matches the configuration of all deployed devices in an enterprise. Malware objects delivered by an encrypted network stream can't be extracted and examined. They're limited because as they observe what an object does, they're only useful for files such as .EXE, .PDF or .DOC.

Malware has also developed to the point where it can identify that it is in a sandbox environment, and therefore 'decides' not to execute, evading the sandbox which believes it is therefore a benign file.

SlashNext netted an A Series round in April of $9 million after selling the technology on a limited basis two years ago.

Related posts:

— Simon Marshall, Technology Journalist, special to Security Now

Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
COVID-19: Latest Security News & Commentary
Dark Reading Staff 9/25/2020
Hacking Yourself: Marie Moe and Pacemaker Security
Gary McGraw Ph.D., Co-founder Berryville Institute of Machine Learning,  9/21/2020
Startup Aims to Map and Track All the IT and Security Things
Kelly Jackson Higgins, Executive Editor at Dark Reading,  9/22/2020
Register for Dark Reading Newsletters
White Papers
Video
Cartoon
Current Issue
Special Report: Computing's New Normal
This special report examines how IT security organizations have adapted to the "new normal" of computing and what the long-term effects will be. Read it and get a unique set of perspectives on issues ranging from new threats & vulnerabilities as a result of remote working to how enterprise security strategy will be affected long term.
Flash Poll
How IT Security Organizations are Attacking the Cybersecurity Problem
How IT Security Organizations are Attacking the Cybersecurity Problem
The COVID-19 pandemic turned the world -- and enterprise computing -- on end. Here's a look at how cybersecurity teams are retrenching their defense strategies, rebuilding their teams, and selecting new technologies to stop the oncoming rise of online attacks.
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2020-15208
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, when determining the common dimension size of two tensors, TFLite uses a `DCHECK` which is no-op outside of debug compilation modes. Since the function always returns the dimension of the first tensor, malicious attackers can ...
CVE-2020-15209
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, a crafted TFLite model can force a node to have as input a tensor backed by a `nullptr` buffer. This can be achieved by changing a buffer index in the flatbuffer serialization to convert a read-only tensor to a read-write one....
CVE-2020-15210
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, if a TFLite saved model uses the same tensor as both input and output of an operator, then, depending on the operator, we can observe a segmentation fault or just memory corruption. We have patched the issue in d58c96946b and ...
CVE-2020-15211
PUBLISHED: 2020-09-25
In TensorFlow Lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, saved models in the flatbuffer format use a double indexing scheme: a model has a set of subgraphs, each subgraph has a set of operators and each operator has a set of input/output tensors. The flatbuffer format uses indices f...
CVE-2020-15212
PUBLISHED: 2020-09-25
In TensorFlow Lite before versions 2.2.1 and 2.3.1, models using segment sum can trigger writes outside of bounds of heap allocated buffers by inserting negative elements in the segment ids tensor. Users having access to `segment_ids_data` can alter `output_index` and then write to outside of `outpu...