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.

Partner Perspectives  Connecting marketers to our tech communities.
10:30 AM
Partner Perspectives
Partner Perspectives
Partner Perspectives

The Performance Penalties of Bloatware-Based Next-Gen Firewalls

Why some organizations turn off critical security features to maintain network QoS.

At last month’s Oracle OpenWorld 2014, Intel president Renee James spoke of the need to eliminate the “performance penalties” of today’s most urgently needed applications and services. Oftentimes, such penalties result from the practice of bolting new hardware or software onto legacy solutions to address new issues. A “bloatware” effect emerges, where the new unwieldy solutions take their toll on an enterprise in the form of management complexity and performance degradation.

Two recent studies suggest that in the critical area of next-generation firewalls, some firewall designs inflict performance penalties so pronounced that organizations resort to turning key security features off. The features are the incremental next-generation capabilities that allow firewall product-marketing departments to keep up with the latest craze, and supposedly move firewalls past the basics of deny/allow. Essentially, those next-gen firewall features are now relegated back to what they were originally -- or even worse, shelfware. Sadly, this also makes an organization look silly: If all the features need to be turned off to maintain performance, why was the investment made in the first place?

The first study, a July 2014 Spiceworks survey of 504 IT professionals, found that more than one-third of respondents admit to turning off advanced firewall features found to be slowing their networks’ performance. The report showed network administrators are most likely to disable deep packet inspection (31%), anti-spam (29%), VPN (28%), data filtering (28%), and antivirus (28%) features to address impeded performance. 

An October report from product-testing firm Miercom provides a clue as to why IT professionals are turning off their security protection. Miercom found a 35% industry average performance loss when deep packet inspection (DPI) is turned on, and a 75% industry average performance loss when DPI, antivirus, and application control is turned on. 

A key component of today’s next-generation firewalls, DPI plays the critical role of detecting malicious activity within regular network traffic and preventing intrusions by blocking offending traffic automatically, before damage occurs. Disarming DPI removes a critical component of an organization’s network security. 

Needless to say, these findings are unsettling, given the good intentions and budgets IT teams put into next-generation firewalls as cornerstone products of their security strategies.

Why do people hobble their next-gen firewalls?

The Spiceworks survey specifically cites the inability of next-generation firewalls to properly cluster and share network traffic loads. We attribute this failure to the fact that many of today’s next-gen firewalls have been built on legacy firewall software designs, which lack the ability to cluster in the way organizations require for today’s enterprise needs.

These firewalls have been updated for next-gen functionality, but the layer upon layer of services and features have bloated their presence on the network to the point that they weigh down the performance of the networks they support. 

There is hope. Not all products use an old-fashioned design or exhibit the same performance penalties. Implementations experiencing lower levels of degradation tend to be developed based on newer product designs built specifically for today’s robust and demanding network environments. Rather than bolting on new components to older solutions, these best performers are built from the ground up to optimize hardware and software to support both security and performance. 

Such designs nurture higher firewall throughputs through the ability to cluster multiple units. The additional nodes these clusters provide can dramatically increase a system’s ability to handle high volumes of traffic while inspecting and investigating it, without affecting the user experience. 

Upgrading existing legacy systems is, of course, a cost-effective practice that can yield value. But when performance penalties are as profound as those uncovered by Miercom, and the forced choice between security and performance so inescapable as evidenced by the Spiceworks survey, it’s difficult to find a silver lining when the end result is making your enterprise more vulnerable. 

Theory vs. practice

For next-generation firewalls to truly deliver on their network security potential, companies need to look beyond data-sheet theory to practical run-time performance. Most firewalls can give you similar features. Challenge yourself and your vendors to show you performance of security features in practice. As security leaders, we need to build our network security defenses on solutions optimized for performance in the enterprise environments of today, not the artificial and theoretical demands of a vendor’s lab. Failing to do so puts the necessities of app functionality and network performance at odds with the necessity of protecting the product of that functionality -- network traffic: a situation no organization should have to face.

Comment  | 
Print  | 
More Insights
Newest First  |  Oldest First  |  Threaded View
User Rank: Apprentice
11/4/2014 | 6:37:34 PM
Re: Smaller Footprint, Bigger Thinking

There are two concepts here – one is core product design, another is deployment optimization. Bloatware is traditionally used to mean OS and applications that grow fatter and more sluggish with each FCS. As part of bloatware deployment, optimization can be a useful tool for tailoring a product to the specific policies, traffic patterns, and risk posture of your organization and network topology. However, it is also often a crutch used to work around poor designs, finessing features and options into the optimal set of tradeoffs. Yet they are still tradeoffs -- something is lost.

Instead, we think you have the right to buy a product that works efficiently out of the box, without heavy optimization. If you have the time, budget, and desire to optimize, you can still choose to do so, but I prefer to see products that perform well at the default settings – with all the features you've invested in turned on.

User Rank: Ninja
10/31/2014 | 12:59:35 PM
Smaller Footprint, Bigger Thinking
When I first walked into a past project - create a new OS template VM for all future SLES VM rollouts - I knew I was going to take the existing installation footprint and cut it down from almost 10 GB to roughly 1 GB.  When a second template came up for our virtual firewall "appliances" I took that 1 GB and made it 600 MB.  My point is, from strictly a size factor, I understood how to review from a configuration/release management perspective all the applications and libraries contained in the VM and whittle the package down to only the most essential packages.

Now, in terms of performance, it's a similar thing.  You identify exactly what you need in your firewall, what you can replace with intermittant services rather than persistant, and also what can be replaced with pure staff power; re-design and roll out with only what you need; the understanding being that you do not turn anything off later without replacing it with another form of that service, or a modified version.  But also, performance can be affected by the software itself - not all algorithms are equal and before buying deep analytics that are going to sit on your network, you should understand the code and whether there are better options out there; even bringing an in-house team to write the code so the footprint is small and the drag on performance is as minimal as possible may be worth the money in the long run.

Trim all the digital fat, and make sure only the best designed code is in place.

P.S.  I suspect that of those admins interviewed, many of them didn't fully understand the options available in their software.  Even bloatware can often be configurable to minimize features and for performance fine-tuning.  Know your apps backward and forward, spend time with the application tech support to identify areas for improvement, and if that isn't working - buy another product.

Inside the Ransomware Campaigns Targeting Exchange Servers
Kelly Sheridan, Staff Editor, Dark Reading,  4/2/2021
Beyond MITRE ATT&CK: The Case for a New Cyber Kill Chain
Rik Turner, Principal Analyst, Infrastructure Solutions, Omdia,  3/30/2021
Register for Dark Reading Newsletters
White Papers
Current Issue
2021 Top Enterprise IT Trends
We've identified the key trends that are poised to impact the IT landscape in 2021. Find out why they're important and how they will affect you today!
Flash Poll
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
PUBLISHED: 2021-04-12
Synapse is a Matrix reference homeserver written in python (pypi package matrix-synapse). Matrix is an ecosystem for open federated Instant Messaging and VoIP. In Synapse before version 1.28.0 requests to user provided domains were not restricted to external IP addresses when transitional IPv6 addre...
PUBLISHED: 2021-04-12
Synapse is a Matrix reference homeserver written in python (pypi package matrix-synapse). Matrix is an ecosystem for open federated Instant Messaging and VoIP. In Synapse before version 1.28.0 Synapse is missing input validation of some parameters on the endpoints used to confirm third-party identif...
PUBLISHED: 2021-04-12
In Gradle before version 7.0, files created with open permissions in the system temporary directory can allow an attacker to access information downloaded by Gradle. Some builds could be vulnerable to a local information disclosure. Remote files accessed through TextResourceFactory are downloaded in...
PUBLISHED: 2021-04-12
Synapse is a Matrix reference homeserver written in python (pypi package matrix-synapse). Matrix is an ecosystem for open federated Instant Messaging and VoIP. In Synapse before version 1.28.0 Synapse is missing input validation of some parameters on the endpoints used to confirm third-party identif...
PUBLISHED: 2021-04-12
Advanced Authentication versions prior to 6.3 SP4 have a potential broken authentication due to improper session management issue.