Slashdot Mirror
Anti-Botnet Market is Black Eye for AV Industry
alternative coup writes "eWEEK is running a story on the emergence of an anti-botnet market to fill a perceived need for software to deal with botnet-related malware (Trojans, keyloggers, rootkits, etc.). The article characterizes this as 'another black eye' for the existing anti-virus industry — asking consumers to pay twice for protection from things that anti-malware suites are missing. Venture capital money is flowing to these anti-bot products, an implicit statement that the AV giants are not doing their jobs. 'For companies such as Symantec, which sells the Sana-powered Norton AntiBot and anti-malware subscriptions, it's a nickel-and-dime situation. Symantec officials say Norton AntiBot is for a specialized, technical market segment looking for high-end tools to deal with botnets, but [Andrew Jaquith, an analyst with The Yankee Group] said it's a case of anti-malware companies double-dipping.'"
Symantec has already lost me as a customer. I began shifting my clients away from it as soon as the new spybot 1.5 released. It has a modicum of registry protection and it generally isnt a crapshoot as to whether or not its going to brick the computer its installed on...brick may be a strong term, but Norton/Symantec's footprint is way too much for a client machine...and now they want to add more.
Yeah...ditch these people now. AV on the client is a scam. Effective management and AV at the chokepoints can often provide enough protection I've found.
...has infuriated me for some time. This idea that some things are 'viruses' and others, 'spyware'. Last year, I tried to nail down Sophos on this very thing. If I'm protected against viruses, shouldn't I also, by default, be protected against spyware since that's how it usually gets on there in the first place?
'Oh no', they tell me. 'That's different...' Yeah. I see that. Now we got this going on.
People want their computers to be protected against any form of intrusion - from within or without - regardless of how it's classified. The reality is, that there are now forms of malware out there that are either undetectable or incurable once you have them. I use a gateway to help protect our computers, but every once in a while it still happens.
"...Well, there's egg and bacon; egg sausage and bacon; egg and spam; egg bacon and spam; egg bacon sausage and spam..."
IANAAVE (I am not an anti-virus expert), but it seems to me that much of the bloat comes from the ever increasing virus signature database these engines have to keep in memory (especially for on-access real time scanning). Considering that there seems to be no end in site for these signature files and the high rate of virus mutation, virus signature tables seem to be an extremely antiquated and inefficient model for detection.
Of course, heuristics won't be a silver bullet as it brings its own set of problems (ie: false positives), but I think we'll see more of this used as time goes on. IANAB (I am not a biologist), but is seems that our body's immune system operates more on heuristics than some exhaustive chemical look up table. Considering the millions (billions?) of years nature has invested in our immune system I think we would do well to take a page from mother nature on this one.
Faith is a willingness to accept something w/o complete proof and to act on it. Reason allows you to correct that faith.
How can an OS add on fix a fundamental problem of the security of an operating system and the applications that are running on top of it?
It is my firm belief that AV software can never fix the real problem: broken OS security model and application bugs. For the AV software vendors this is always a game of catch up, the virus/trojan/worm/bot etc. creators have a huge advantage: numbers. They have more people figuring out ways to infect your computers, brake through your buggy and exposed application interfaces, send out executables with backdoors and viruses.... there are probably thousand times as many people working on the ways to take over PCs than there are people who are in 'business' of preventing this from happening.
And really, it is not that complex of a problem: run OS administration applications in one security level, run user applications in another security level, use hardware infrastructure to prevent these levels from intersecting and taking over each other, but of-course allow the highes level administration applications to take precedence over any user application and at least kill it. Do not allow execution of applications that are not authorized by the user. There are more good ideas than that, but basically do not allow a user application to hijack the system by pretending to be an OS administration application, do not allow user applications to change their access levels, do not allow them to hide their processes from observers. Designate protected data storage on disks, and allow that data only to be modified by certain applications that are assigned by the user.
However this is not a job for some ad on AV software.
You can't handle the truth.
Anti-virus, anti-spyware, firewall -- all of these protections should be built in to the operating system.
We shouldn't have to add third-party tools to make an OS secure. It should be secure (or at least, secure-able) out of the box.
Charging more for a suite of software that all does the same thing sounds like a last-gasp attempt to deliver some profits before architectural changes force these companies out of business.
the trick to understanding it was to know that "AV" stands for "antivirus", not "adult video"
what does the adult video industry have to do with botnets? and nevermind the black eyes, that's a kind of adult video i'm not into
live and learn
intellectual property law is philosophically incoherent. it is your moral duty to ignore it or sabotage it
... the best protection against botnets is never install Windows?
That will only hold true as long as the market share for the non-Windows operating systems remains at its current levels. Whether Mac or Linux is intrinsically more secure than Windows is a subject for another (lengthy and heated) discussion, but the fact remains that practically, an OS is only as secure as the user running it lets it be. Linux users are much more secure from threats than Windows users for two reasons. One: since Linux accounts for such an infinitesimally small percentage of market share, malware coders don't waste their time coding for Linux. Two: since most Linux users are enthusiasts who generally know what they are doing, they can harden their installs to a greater degree than your average Joe-Sixpack Windows user.
A large upsurge in Linux use, especially by the 'typical' user that clicks on anything and everything, and runs their console session as root, would be irresistible to the malware coders, and you'd see the same situation you're seeing with Windows now.
____
~ |rip/\/\aster /\/\onkey
Except what you'll see is 50 million computer users running Linux as root all the time because an OEM configured it that way rather then be annoyed with support calls asking how to install some new program. Those 50 million people then get an email about free XXX videos, run an attachment that installs various kinds of malware, and we're right back where we started.
Clueless users given the ability to become administrators (which they can if they own the machine) will defeat any OS security.
-- "So they told me that using the download page to download something was not something they anticipated." - Bill Gates
why don't ISPs just set up honey pots and use them as test beds to determine what traffic is being generated by a bot, and kill the traffic as it leaves the costumer's computer
That doesn't solve the problem - it just moves it. Onto the vendors of networking hardware.
Core routers are "dumb as rocks" and can be relatively low reliability. The idea there is to treat each packet as a hot potato and move it on with as little "thought" about it as possible - so limited processing power can handle large numbers of packets. If the box goes down the others can find a way around it. But not thinking about each packet means these boxes are gullible.
Edge routers (the last router before the customer, or sometimes the one between two competing ISPs) are smarter and more robust: In the core there are multiple connections, but at the (customer) edge there is usually only one line to only one box, so it has to be as reliable as a phone switch. (If the ISP hasn't routed ALL traffic to/from the user through an extra box at the Network Op Center) it has to act as a "reverse firewall" to protect the gullible network routers from the users and keep the user from using resources he hasn't paid for. It's also the only box on the carrier side where all the customers' packets come together. So if the carrier is to provide comprehensive anti-malware service, that's where it ends up.
Edge routers have a lot of brains and a significant amount of memory. But for their main jobs they only have to look at headers and keep a small amount of state per customer. Add "deep packet inspection" for anti-malware on the current model and you explode the resources required. Now they have to look at the whole content of every packet and apply thousands of tests to it, exploding processor requirements. Worse they have to keep the state for every flow rather than just every customer - and a single tool-generated web page may be hundreds or thousands of separate flows, running in parallel due to browser optimization. And the state for each of the flows is enormous, including the state of the processing of each of the signatures being tested. Finally, they may actually have to hold the packets themselves, to reorder and/or defragment them for the analysis. So the storage requirements explode. And this resource requirement increases their susceptability to DOS attacks.
Further, smartening up the edge routers still further and giving them massive storage upgrades and inbound firewall duties makes them, not the users' machines, the primary target for malware vendors. They'd now have to spoof or subvert this machine to get their stuff to the users. But what a prize! Once it's subverted they get access to ALL the users and their traffic, regardless of the users' OS or anti-malware tools. (The zero-day window becomes "pwnership" of ALL the customers' data - no race between the infection spreading and the AV companies working out and deploying a signature.) Once in control, tapping should be a snap: The routers already have a government-mandated "lawful intercept" capability in place - just reconfigure it to send to the malware operation rather than the authorities. And talk about monocultures: The number of edge router vendors can be expressed with a single digit, likely with (at least at first) only one deep-packet-inspection product each. And they'll no doubt ally with the current anti-malware vendors to obtain their algorithms and signature updates.
So going to ISP-based filtering transfers the computational load of defense from a distributed web of end-users' machines to a small set of ISP boxes, increases the "software monoculture" vulnerability, provides an upstream target that the end user can't defend with a limited number of instances, makes it as vulnerable as the current worst-of-breed approach (microsoft OS and tools plus signature-based active immunity), gives access to ALL users on EVERY success, and raises the cost of the network boxes (and thus your networking bill).
Lowered security at a higher price doesn't seem like a good approach to me.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way