17.1 C
London
Saturday, September 23, 2017
Home Tags Banner

Tag: Banner

The power of machine learning comes at a price. Once you have the skills, the toolkit, the hardware, and the data, there is still the complexity involved in creating and fine-tuning a machine learning model.But if the whole point of machine learning is to automate tasks that previously required a human being at the helm, wouldnrsquo;t it be possible to use machine learning to take some of the drudgework out of machine learning itself?[ Review: TensorFlow, Spark MLlib, Scikit-learn, MXNet, Microsoft Cognitive Toolkit, and Caffe machine learning and deep learning frameworks. | Roundup: 13 frameworks for mastering machine learning. | Cut to the key news and issues in cutting-edge enterprise technology with the InfoWorld Daily newsletter. ]Short answer: a qualified yes.

A collection of techniques, under the general banner of “automated machine learning,” or AML, can reduce the work needed to prepare a model and refine it incrementally to improve its accuracy.To read this article in full or to leave a comment, please click here
In Q2 2017, the average share of spam in global email traffic amounted to 56.97%, which was only 1.07 p.p. more than in the previous quarter. One of the most notable events of this quarter – the WannaCry epidemic – did not go unnoticed by spammers: numerous mass mailings contained offers of assistance in combating the ransomware.
The Open Container Initiative, a consortium founded to develop open standards around Docker-style containers across platforms, has delivered 1.0 milestones for two crucial specifications under its banner.The new standards aren't likely to affect the way developers work with containers.

The real impact is likely to be felt by commercial producers of container-related products, especially if they are angling to have OCI certification applied to what they produce.[ 9 lies programmers tell themselves. | 9 bad programming habits we secretly love. ]OCI's newly finalized standards cover two key components of the container ecosystem -- the image format for containers, and the runtime specification.

The OCI Image Format, as the first is formally called, is easy enough to grasp.
It describes the way a container image is laid out internally and what its various components are.To read this article in full or to leave a comment, please click here
Although the beginning of Q1 2017 was marked by a decline in the amount of spam in overall global email traffic, in March the situation became more stable, and the average share of spam for the quarter amounted to 55.9%.

The US (18.75%) remained the biggest source of spam, followed by Vietnam (7.86%) and China (7.77%).
We can all rest easy.

Ajit Pai is going to restore the free and open internet we’ve been pining for, lo these past two years.In his speech at a FreedomWorks event this week, the FCC chief lamented the lost golden age of broadband, which we lived in before onerous net neutrality regulations were passed that mandated ISPs treat all internet traffic equally and forbid them from blocking or throttling users’ access to content.[ Read ‘em and weep: 5 ways your ISP is screwing you. | 5 more ways your ISP is screwing you. | Cut to the key news in technology trends and IT breakthroughs with the InfoWorld Daily newsletter, our summary of the top tech happenings. ]In the two years since the “serious mistake” of Title II classification was foisted upon the telecommunications industry, the country has been plagued by a decline in infrastructure investment, according to Pai.

The consequences are dire: Fewer Americans will have access to high-speed internet, there will be fewer jobs, less competition, and declining test scores—no wait, he failed to mention that last one. Regardless, net neutrality is the culprit.To read this article in full or to leave a comment, please click here
Egham, 18 April 2017 – OKI Europe Ltd says that large format printers such as its newly-launched, award winning, ColorPainter E-64s, offer sign makers, exhibition services providers, screen printers and others the opportunity to move away from low-margin vinyl and banner printing into more profitable applications, giving their business the boost needed to drive future growth and success.The E-64s and other larger models in its ColorPainter range overcome the environmental, quality and media flexibility barriers... Source: RealWire
Gaze-based technology "can also track whether the users have viewed them."
A technological boom in medicine both encouraged medical institutions to use exclusively information systems in processing data and led to the emergence of new types of technological equipment and personal devices that can be used to interact with traditional systems and networks.

This means that the threats that are relevant for them can also be relevant for medical systems.
Includes "net neutrality" and "EpiPen"; still on the sidelines about how to say "GIF."

When choosing a security suite, you probably look for familiar company names rather than trusting your security to an unknown. Germany-based G Data may not have huge mindshare in the United States, but it's big in Europe. G Data Internet Security includes all the features you'd expect in a suite, including an antivirus, a firewall, parental controls, and a spam filter. Unfortunately, the quality of the components spans quite a range, from very good to very poor.

Bitdefender, Kaspersky, and ESET Internet Security 10 are among the suites that cost roughly $80 for three licenses. There's another group around $60 that includes Webroot, Trustport, and Avast. G Data falls in between, with a $64.95 subscription price for three licenses. If you need just one installation, you can cut $10 from that price.

This product's main window features the familiar bold G Data color scheme, with a red banner holding a row of icons at top. Some security vendors use precisely the same component layout throughout the product line, showing unavailable features as disabled. Not G Data. The home screen shows a detailed security status, with links to important components, but there are more components displayed in the suites banner than that of the standalone antivirus. To the three top-row icons found in the antivirus, the suite adds icons for its backup, firewall, and parental control features.

Shared with Antivirus

The antivirus protection in this suite is precisely what you get in G Data Antivirus 2017. I'll summarize my findings here, but if you want full details you should read my review of the antivirus.

Four of the five antivirus labs that I follow include G Data in their tests and reports. It earned an above-average rating in the RAP (Reactive and Proactive) test from Virus bulletin, but didn't do quite as well in the three-part testing performed by AV-Test Institute. G Data earned the maximum six points for protection against malware, and six more for low false positives, but a drag on performance dropped its score to 4.5 in that category. A total of 16.5 points is good, but Kaspersky Internet Security took a perfect 18 points in this test. Bitdefender and Trend Micro were close behind, with 17.5 points.

In the real-world attack simulation tests by SE Labs, G Data took AA certification, the second-highest of five possible levels. Emsisoft, Kaspersky, Norton, and Trend Micro managed an AAA rating. Like most tested products, G Data failed the pass/fail banking Trojans test performed by MRG-Effitas. Its aggregate score of 8.7 points is good, but Kaspersky leads with 9.8 of 10 possible points, and Norton got 9.7 points.

Like Webroot, Comodo Antivirus 10, and PC Matic, G Data detected 100 percent of the samples in my malware collection. Not-quite-perfect blocking of a few samples results in an overall score of 9.8 points. That's very good, but the other three I mentioned managed a perfect 10. G Data wasn't fooled at all by my hand-tweaked samples; it blocked them all. Comodo, by contrast, missed 30 percent of the modified versions.

For a different look at malware blocking, I use a feed of recently discovered malware-hosting URLs supplied by MRG-Effitas. G Data blocked 78 percent of the samples in this test, almost all by completely blocking access to the URL. Norton tops this test, with 98 percent protection.

The same Web-based protection component should also serve to steer the hapless user away from fraudulent sites that try to steal login credentials. However, G Data fared poorly in my antiphishing test, with a detection rate 44 percent lower than Norton's. While most products lag Norton in this test, more than half of them did better than G Data. Only Bitdefender, Kaspersky, and Webroot SecureAnywhere Internet Security Plus have eked out a better score than Norton.

Other Shared Features

Exploit protection is usually associated with the firewall component, but G Data offers it in the standalone antivirus. In testing, it didn't block exploits at the network level, but wiped out the executable payload for 50 percent of the samples. That's quite good. Champion in this test is Symantec Norton Security Deluxe, which stopped 63 percent of the attacks at the network level.

My hands-on testing confirmed that G Data's keylogger protection and ransomware protection are effective. For those tests, I had to turn off all other protective layers.

Similar to the SafePay feature in Bitdefender Internet Security 2017, BankGuard protects your browsers from man-in-the-middle attacks and other data-stealing attacks. The AutoStart manager lets you reversibly disable programs from launching at startup, or set them to launch after a delay.

See How We Test Security Software

Basic Firewall

Every firewall needs to at least match the abilities of the built-in Windows Firewall that it replaces. Specifically, it must block outside attacks and put the system's ports in stealth mode, so they're not visible from the Internet. G Data's firewall fended off my port scans and other Web-based attacks, and popped up a notification that it had done so. So far, so good!

The settings page for G Data's firewall is pleasantly simple. A large slider lets you choose one of five preset security levels: Maximum, High, Standard, Low, and Disabled. Three other pages of settings are deliberately unavailable, with their configuration changed automatically as you switch security levels. True firewall experts can choose custom settings, thereby enabling access to those pages. But most should leave the firewall set to its default Standard level.

Most firewall components also keep track of how programs are using your network connection. Advanced firewalls like Norton's automatically define permissions for millions of known programs and carefully watch how unknowns behave, smacking them down if they show signs of misusing the network. Less advanced firewalls rely on the user to determine whether unknown programs should be allowed to access the network, which sometimes results in a deluge of popup queries.

G Data's firewall runs by default in autopilot mode, meaning you won't see any queries. It's not entirely clear just what it does in this mode, but as far as I can tell, it allows all outbound connections and rejects unsolicited inbound connections. That's not doing a lot.

To see the program control component in action, I turned off autopilot. Cleverly, the program offers to temporarily turn autopilot back on if it detects you're launching a full-screen application.

When I tried launching a guaranteed-unknown program (a small browser I coded myself), G Data popped up asking whether to allow or block access, once or always. That's exactly what should have happened. I tried a few leak test utilities, programs that try to gain access to the Internet without triggering the firewall's program control. G Data caught some, but not all, of these.

Unfortunately, it also popped up repeatedly for some Windows internal components. Note, too, that firewall popups appear for any user account, including non-Administrator accounts. While your toddler is playing games online, she may accidentally tell G Data to always block access by some Windows component. In that case, you'll need to open the Application Radar window from the Firewall status screen to unblock that application.

A firewall isn't much use if a malicious program can reach in and flip the off switch. I couldn't find a way to disable G Data by manipulating the Registry, though it didn't protect its Registry data against change the way Bitdefender, McAfee Internet Security, and others do. The last time I tested G Data, I found that I could terminate some of its processes using Task Manager. This time around, all 11 processes received protection.

Alas, G Data's essential Windows services are still vulnerable to a simple attack that could be carried out programmatically. I set the Startup Type for each of six services to disabled and then rebooted the computer. That effectively eliminated G Data's protection. In a similar situation, Comodo Firewall 10 Firewall seemed to succumb, but recovered on reboot.

This firewall component handles the basic tasks of protecting against outside attack and preventing programs from misusing your Internet connection, but that's about all. And the vast majority of competing products manage to harden their Windows services against tampering more thoroughly than G Data does.

Cloud Storage Backup

When you first click the backup icon, you just get a big, empty page. A bit of investigation reveals the New Task button. Clicking it brings up a disclaimer pointing out that the subscription you have offers online backup only. If you want advanced features like making local backups or burning backups to optical media, you must upgrade to G Data Total Security. You can check a box to suppress this disclaimer in the future.

To start designing a backup job, you select files and folders for backup. You do this using a folder/file tree. Checking or unchecking a folder selects or deselects all its contained folders and files. If you simply check the tree item with your username, representing all your user data, that may be enough.

The selection tree exhibits a strange redundancy that might cause trouble. For example, after the entry with your name is an entry called Libraries. If you check your username entry, the corresponding entries under Libraries (Music, Videos, Documents, and Pictures) do not get checked. But if after that you check Libraries and then uncheck it, those four entries under your username lose their checkmarks. This is just one of several redundancies in the tree, so you should carefully review your selections before proceeding.

The next step is target selection, but your only choice is cloud backup. Well, there's also an option to copy the archived data to an FTP server, but not many users are equipped to perform the necessary configuration. When I tried to continue at this point, the program admonished me, "Cloud has been selected as target, but no login has been entered." Guessing at this point, I clicked a button for network login—no joy. I finally thought to click the cloud icon. This triggered a menu titled New Account, which in turn asked me to select Dropbox or Google Drive. That could be clearer.

Kaspersky Total Security also offers to store backups on Dropbox, but this is just one of its many options. With Kaspersky, you can also back up your files to any local, removable, or network drive, or to an FTP server.

You can optionally create a schedule, separately for a full backup of all data and for a partial backup containing only changed data. Do you know what the difference between a differential backup and an incremental backup is? If not, just leave it set at the default. For each type of backup you can choose one-off, daily, weekly, or monthly backup, or just run the backup manually when you think of it.

Now you can review the dozens of options on the final page of settings. Some are disabled, most are set to the best configuration, but there's one you might want to tweak. By default, G Data opts for fast compression, making the backup process as speedy as possible. If you're short on cloud space, consider setting it to emphasize good compression, instead.

You can create as many backup jobs as you like. You might choose redundancy, backing up to both Dropbox and Google Drive. These jobs appear in the previously blank main backup window.

As for restoring backed-up files, it's a snap. Choose the backup, choose to restore all files or just some of them, and choose whether to restore to the original location or a new location.

Norton gives you 25GB of hosted online backup storage, and makes setting up a backup job very easy. Webroot completes also offers 25GB of storage, and serves as a full file-syncing tool. The backup system in G Data does the job, but it requires that you use third-party cloud storage, and it could be much, much simpler for users.

Porous Parental Control

This suite's parental control system is minimal, consisting of content filtering and time scheduling for Internet or computer use.

The content filter can block websites matching five categories: Drugs, Hackers, Violence, Extremist, and Pornography. There's also an option to block all HTTPS sites, but it's a ridiculous option. Yes, it would prevent access to secure anonymizing proxies, but it would also block any site that sensibly uses a secure connect, including Google, Unicef, and Wikipedia.

Parents can limit time on the computer, the Internet, or both. When enabled, the default in each case is 1.5 hours per day, 10.5 hours per week, and 45 hours per month. These times line up nicely. For example, 1.5 hours on each of seven days equals 10.5 hours. You can also define a weekly schedule, in one-hour increments, for when the child can use the Internet, or the computer. This feature uses a handy grid that makes it easy to set allowed and blocked times.

When I put G Data's scheduler to the test, I found that time-scheduler relies on the system clock. Resetting the clock to an allowed time defeats it. Admittedly, I couldn't find a similar way to defeat the daily cap.

Content filtering is keyword based, and it's both too lax and too strict. Photo-based pornographic sites with no banned words in the URL or page text flew right past the filter, while perfectly innocent sites triggered the over-zealous filter. For example, it blocks any page on blogspot.com because the filter found "pot" in the URL. Pages on the American Kennel Club site that used the word bitch (perfectly valid in this context) got the axe. And so on.

You'd think the Hackers category would block secure anonymizing proxy websites, but it doesn't. By connecting to one, I completely eluded the filter—don't think your teenager won't figure this out.

G Data does report which websites it blocked for each user, along with a date/time stamp and explanation. The explanation helped me confirm that, for example the app did indeed block a blogspot.com page due to the embedded word pot.

This is just not a useful system. If you need parental control in your security suite, look elsewhere. The parental control component in Norton is an Editors' Choice as a standalone. ZoneAlarm's is based on ContentWatch Net Nanny 7, another Editors' Choice. And Kaspersky Total comes with the excellent Kaspersky Safe Kids.

Simple Spam Filter

The need for local spam filtering gets smaller and smaller as more people use services that filter spam at the server level. If you're one of the few who don't get spam skimmed out of your email feed before it arrives, it's nice to have spam filtering handled by your security suite.

G Data analyzes incoming POP3 and IMAP email messages, flagging suspected spam messages, messages with a high spam probability, and messages with a very high spam probability. It prefixes [suspected spam] to the subject line for the first category, [spam] for the other two. You can change these tags, if you like, but most users will surely leave them at their default values.

This spam filter integrates with Microsoft Outlook, automatically diverting marked messages into the spam folder. Those using a different email client must create email rules based on the subject tags, not a terribly challenging task.

G Data uses quite a few different criteria to develop a spam score for each message. It checks the message text for certain keywords, and the message subject for a different set of keywords. You can edit either keyword list. It also includes a self-learning content filter system that's meant to improve accuracy over time.

The spam filter can also check spam messages against real-time blacklists. This process tends to slow the email download, so by default it only uses those blacklists for suspicious messages. Digging deeper, you can configure the spam filter to reject messages written in languages you don't speak. But really, most users can just leave the spam filter settings alone.

You can put specific addresses or domains on the whitelist, to ensure that the spam filter never blocks them. Conversely, you can blacklist addresses or domains to ensure they always get filtered. There's no option to import the content of your address book, or automatically whitelist addresses to which you send mail, like you get with ESET, Trend Micro Internet Security, and others.

If you do need local spam filtering, and want your security suite to handle it, G Data is as good as any. It doesn't offer the comprehensive feature collection that Check Point ZoneAlarm Extreme Security 2017 does, but on the flip side, it doesn't require any attention from you.

File Shredder

On a seriously icon-infested desktop, you not notice the appearance of a new icon titled G Data Shredder. This is a secure deletion utility, for use when you want to delete a file beyond the possibility of forensic recovery. Many encryption utilities come with a shredder, for thoroughly wiping out the originals of files that have been encrypted.

Simply deleting a file sends it to the Recycle Bin, and bypassing the Recycle Bin leaves the file's data still on disk, just marked as space that can be reused. Overwriting that data just once is enough to defeat software-based recovery. Recovery experts use hardware systems to peel back the layers and find previously stored data, but those techniques run into the limitations of physics at about seven overwrites. Why G Data lets you choose up to 99 overwrites I do not know. Three should be plenty for normal use.

Once you've configured the shredder, you drag files and folders onto its icon for secure deletion. You'll also find a Shred choice on the right-click menu.

Minor Performance Impact

While testing G Data, I occasionally felt the system might be running a little slow, but then, my virtual machines necessarily don't have a lot of resources. Running my hands-on performance tests revealed only minor impacts on system performance.

The biggest hit (not big, but biggest) came in my boot time test. Averaging many runs before installation of the suite and many more after, I found that the boot process took 26 percent longer with G Data loading at boot time. Given that most people reboot only when forced to, that's not a big deal.

To check whether a security suite affects everyday file manipulation activities, I time a script that moves and copies an eclectic collection of files between drives. Averaging multiple runs with no suite and with G Data installed, I found the script took 18 percent longer. That's not bad; the average for this test among current products is 23 percent. And there was no measurable slowdown for my zip/unzip test, which compresses and decompresses that same file collection repeatedly.

While G Data didn't put much of a drag on performance, some competing products had even less impact. Webroot, in particular, didn't show measurable impact in any of the three tests.

Component Quality Varies

G Data Internet Security 2017 includes all of the expected security suite components and even offers a backup system. The antivirus performed well in testing, but the parental control system is both limited and ineffective, and the basic firewall could be disabled by a determined hacker. You're better off with a suite in which all of the components do a good job.

For the purpose of defining Editors' Choice products, I distinguish basic suites like G Data, feature-packed mega-suites, and cross-platform multi-device suites. In the basic suites arena, Bitdefender Internet Security and Kaspersky Internet Security are my Editors' Choice products. Both cost a bit more than G Data, but they also offer much better security.

Sub-Ratings:
Note: These sub-ratings contribute to a product's overall star rating, as do other factors, including ease of use in real-world testing, bonus features, and overall integration of features.
Firewall:
Antivirus:
Performance:
Privacy:
Parental Control:

Back to top

PCMag may earn affiliate commissions from the shopping links included on this page. These commissions do not affect how we test, rate or review products. To find out more, read our complete terms of use.

G Data Antivirus 2017

Some antivirus companies that are big in Europe don't get as much mindshare here in the US.

G Data is one such security software maker.

According to the G Data website, G Data developed the very first antivirus in 1985; while some dispute that claim, the company has clearly been around for a while.

G Data Antivirus 2017 is the company's latest, and it does a good bit more than the basics of antivirus protection.

At $39.95 per year for a single license, G Data is in good company price-wise.

Bitdefender, Kaspersky Anti-Virus, Norton, and Webroot are among the numerous products at that price point.

For another $10, you can install G Data on up to three PCs.
If you go for a multi-PC license, you create an account for the first installation, then log in to that account for the rest.

G Data's main window features a bold red banner across the top. Not red for danger, or for stop—it's just red.

The rest of the main window displays the status of the product's numerous protection features, in several groups.

A green checkmark icon indicates that the feature is fully active.

For a partially disabled component, the icon changes to a yellow exclamation point; a fully disabled feature gets a grey dash icon. Naturally, you want to see green across the board.

Lab Results

G Data participates in testing with three of the five independent testing labs that I follow.
In Virus Bulletin's RAP (Reactive And Proactive) test, it scored 85.19 percent.

The average score for products I follow is 81.99 percent, so G Data comes in above average. PC Pitstop PC Matic scored highest in the latest test, with 94.75 percent, but failed overall due to many false positives.

Testers at AV-Test Institute look at antivirus products from three different perspectives, assigning up to six points for each of the criteria.

G Data earned 6 points in the all-important protection category, and by avoiding false positives (detection of valid programs as malicious) it managed another six points for usability.

A small impact on performance dragged its score in that category down to five points, however.

The overall score of 17 points wasn't quite enough to earn it a Top Product rating, but it's good.

In that same test, Kaspersky scored a perfect 18 points.

Bitdefender, Quick Heal, and Trend Micro Antivirus+ Security got 17.5 points.

These four earned the designation Top Product.

Most of the lab tests I follow report a range of results. MRG-Effitas takes a different tack.

To pass the banking Trojans test, a product must protect against every sample used; anything less is failure. Over 70 percent of tested products fail, G Data among them.

Due to the binary pass/fail nature of this test, I give it less weight when calculating an aggregate lab score.

G Data's three lab results worked out to an aggregate score of 8.7 points, which better than most companies manage. However, based on tests from all five labs, Kaspersky took 9.8 of 10 available points, the best aggregates score.

Avira Antivirus and Norton managed 9.7 points, each tested by three of the five labs.

Effective Malware Blocking

Your antivirus utility has many opportunities to save your PC from malware attack.
It can block access to the malware-hosting website, eliminate the threat on download, detect and delete known malware based on its signature, and even detect unknown malware based on behavior alone.

G Data includes all of these layers of protection, and my hands-on testing showed them in action.

In addition to scanning files on access, G Data scans your computer any time it's idle.

Between real-time protection and idle-time scanning, there isn't a screaming need for a full scan of your whole computer.
If you want a full scan, you click the Idle Time Scan link on the main window and choose Check Computer.

A full scan of my standard test system took an hour and 40 minutes, over twice the current average of about 45 minutes.

But once again, unless you actively suspect an infestation you should be able to just rely on the idle-time scan.

When I opened the folder containing my current collection of malware samples, G Data started examining them.

The process was slower than with many competing products, but clearly very thorough.
In most cases, it offered to quarantine the item as its default action; for a few, it advised simply blocking the file from execution.

By the time it finished, 97 percent of the samples were either quarantined or deactivated.

I keep a second set of samples on hand; these are modified versions of the originals.

To create each modified sample, I change the filename, append nulls to change the file size, and overwrite some non-executable bytes.

G Data detected all of the same samples, even in their tweaked form.
In addition, it detected all the remaining samples after execution, for a 100 percent detection rate. Webroot SecureAnywhere AntiVirus, F-Secure, and Ashampoo Anti-Virus 2016 also detected 100 percent of the samples. PC Matic also blocked 100 percent of the samples, but then, it blocks any unknown program.

Webroot managed a perfect 10 points in this test.

G Data, like F-Secure Anti-Virus, allowed a few executable traces to hit the test system, but the 9.8 points both of them earned is still very respectable.

For another view of each product's ability to protect against malware, I use a feed of current malware-hosting URLs supplied by MRG-Effitas.
I launch each URL in turn, discarding any that are defective, and noting whether the antivirus blocks access to the URL, wipes out the malware download, or fails to respond at all.
I keep at it until I've accumulated data for 100 malicious URLs.

G Data earned a 78 percent detection rate in this test, in most cases by blocking access to the malware-hosting URL.

That's just a middling score.
Symantec Norton AntiVirus Basic and PC Pitstop managed 98 percent protection, with Avira close behind at 75 percent.

I didn't see G Data's behavior monitoring kick in during these tests, because other protection layers beat it to the punch.
In any case, behavior monitoring in some antivirus products bombards the user with dire warnings about good and bad programs alike.

For a sanity check, I installed about 20 old PCMag utilities, programs that tie into the operating system in ways that malware might also do.

G Data didn't flag any of the PCMag utilities, but it did give the stink-eye to two of my hand-written test programs.
It popped up a clear warning that the test program might be malicious, with a detailed list of its reasons, and its reasons made total sense.

A program that launches Internet Explorer and manipulates it to download malware? That's suspicious! I'm pleased to see that behavior monitoring kicks in for a pattern of suspicious behavior, not for every little potential problem.

So-So Phishing Protection

Writing a data-stealing Trojan and getting it somehow installed on victim PCs can be a tough job.
Simply tricking users into giving away their passwords and other personal data can be quite a bit easier. Phishing websites masquerade as financial sites, Web-based email services, even online games.
If you enter your username and password on the fraudulent site, you've given the fraudsters full access to your account.

If the website looks just like PayPal but the URL is something goofy like armor-recycling.ru, at least some users will detect the fraud.

But sometimes the URL is so close to the real thing that only those with sharp eyes will spot it as a fake.

Antivirus programs that have a Web protection component usually attempt to protect users against phishing as well, and G Data is no exception.

To test the efficacy of a product's antiphishing component, I first scour the Web for extremely new phishing URLs, preferably URLs that were reported as fraudulent but that haven't yet been analyzed and blacklisted.
I launch each simultaneously in one browser protected by the product under test and another protected by long-time fraud fighter Norton.
I also launch each URL in instances of Chrome, Firefox, and Internet Explorer, relying on the browser's built-in phishing detection.

Because the collection of fraudulent sites differs every time, I report results in relative terms rather than absolute detection rate.

Very few products do better than Norton in this test, but many come closer than G Data did.

G Data's detection rate came in 45 percentage points below Norton's, which a is poor result.
Internet Explorer and Chrome both did a better job than G Data. Yes, G Data beat Firefox, but Firefox hasn't been doing very well lately.

The lesson here? Don't turn off your browser's built-in phishing protection.

Bonus Features

Along with the expected antivirus features, G Data gives you several features that you'd expect to see in a security suite.
I tested its exploit protection by hitting the test system with about 30 exploits generated by the CORE Impact penetration tool.
It identified 30 percent of the exploits by name and blocked another 20 percent using more generic detection.

That 50 percent detection total is as good as what Kaspersky Internet Security managed in this test. Norton leads this test, with 63 percent protection.

Like Safepay in Bitdefender Antivirus Plus 2017 and Kaspersky's Safe Money, G Data's BankGuard feature aims to protect your financial transactions.

Bitdefender uses a whole separate desktop to run Safepay, and Kaspersky puts a glowing green border around the browser protected by Safe Money.

By contrast, BankGuard works invisibly to protect all your browsers.

The only way to see it in action is to encounter a Trojan that attempts a man-in-the-browser attack or other data-stealing technique.

The related keylogger protection feature was easier to test than BankGuard.
I installed a popular free keylogger, typed some data into Notepad, typed into my browsers, and then typed in Notepad again. When I brought up the keylogger's keystroke capture report, it showed no keystrokes between the two uses of Notepad.

To test G Data's ransomware protection component, I first turned off every other feature related to real-time malware protection. When I launched a ransomware sample, it quickly popped up a warning about suspicious behavior that suggests encrypting ransomware, with the caveat that if you are actively running an encryption utility yourself, you can ignore the warning. My G Data contact noted that in most cases, some other layer of protection will block the ransomware before it gets to this point.

G Data has long featured the ability to manage the programs that launch automatically when your system boots.
Its Autostart Manager can delay launch of any such program for from one to 10 minutes, or set it to never launch at startup. You can also configure it to launch the program when the system's startup activity has died down.

This is a more fine-grained control than you get with the similar feature in Norton.

A Mature Product

G Data has been around longer than almost any of its competitors, and G Data Antivirus 2017 is a mature product.
Since my last review, it has added components specifically designed to protect against exploits, keyloggers, banking Trojans, and ransomware.
It earned a great score in my hands-on malware-blocking test, and took decent scores from the independent testing labs. However, it proved less effective at blocking access to malicious and fraudulent URLs.

Bitdefender Antivirus Plus and Kaspersky Anti-Virus earn top scores from the independent labs.
Symantec Norton AntiVirus Basic scored high in all of my hands-on tests, and includes an impressive set of bonus features. Webroot SecureAnywhere Antivirus goes even farther with behavior-based detection, making it the tiniest antivirus around.

And a single license for McAfee AntiVirus Plus lets you install protection on every device in your household. Out of the huge range of antivirus products, these five have earned the title Editors' Choice.

Back to top

PCMag may earn affiliate commissions from the shopping links included on this page.

These commissions do not affect how we test, rate or review products.

To find out more, read our complete terms of use.

reader comments 6 Share this story Many of the biggest legal disputes in technology relate to "intellectual property," a broad term used for laws relating to everything from copyrights to patents, trademarks to trade secrets.

This year saw significant changes in the copyright and patent landscapes. "Patent trolls" who sue technologists for fun and profit got smacked down by courts more often—and harder—than ever before.

At the same time, universities were filing patent lawsuits at an increased rate, and often winning.In the copyright realm, the Oracle v.

Google
trial dominated the spring.

A jury was left to decide the murky rules about when using an API could be "fair use." That legal uncertainty led to the two tech giants clashing over the ethics of each others' business practices and the history of the smartphone industry. In two very different cases in 2016, copyright issues led to criminal charges being filed. US authorities are seeking to extradite and put on trial a man named Artem Vaulin, who they say made $16 million annually by running a massive online storehouse of pirated films and songs.

And more than three years after they were condemned by a federal judge, lawyers behind a vast array of copyright lawsuits, a firm known as Prenda Law, were arrested and accused of fraud. Here's a look back at 2016's most dramatic IP cases. Graphiq CEO Kevin O'Connor and former director of operations Danny Seigle.

Graphiq (formerly FindTheBest) became the first company to win attorneys' fees in a patent case under the Supreme Court's new Octane Fitness standard.

An appeals court approved the fee award in January 2016. Patent trolls continued to face stiff fines throughout 2016. eDekka, the most litigious patent company just a year ago, collapsed and dropped its appeal after being hit with fees in East Texas. Carnegie Mellon University ended a prolonged patent battle with Marvell Technology in February, with Marvell agreeing to pay a $750 million settlement—the largest payout ever for a patent related to computer science. Pictured here is CMU Professor José Moura, inventor on the two patents in the case. An image explaining one of two patents owned by Carnegie Mellon University, which describe a method of reducing noise when reading data from hard disks.

The patents were used by CMU to sue Marvell Technology. Universities have increasingly been willing to become plaintiffs in high-stakes patent lawsuits, and are sometimes partnering with professional patent enforcement companies to do so.

The Electronic Frontier Foundation launched a "reclaim invention" campaign in June 2016, seeking to pressure universities not to partner with such "patent trolls." Since the US Supreme Court's 2014 Alice v.

CLS Bank
decision, it's been easier to get software patents thrown out of court. Until this year, the US Court of Appeals for the Federal Circuit had only upheld software patents in one post-Alice case.

But in 2016, the Federal Circuit gave approval to software patents in three more cases.

The image above is pulled from the McRo v.

Bandai Namco Games
opinion.

A Federal Circuit panel said McRo's digital animation patents could survive, rejecting arguments from public interest groups like EFF that McRo was being allowed to essentially patent mathematics. In May, a second jury trial between Oracle and Google over whether the Android operating system violated Java copyrights ended with a second resounding win for Google.

The testimony of Jonathan Schwartz, former president of Sun MicroSystems, loomed large in the case.
Schwartz testified that he had no problem with Android, since Google had followed the rules around Java intellectual property that Sun had established. Noah Berger/Bloomberg via Getty Images Oracle attorneys tried to sway the jury by painting former Sun Microsystems president Jonathan Schwartz as a hypocrite, who praised Google in public but privately decried its licensing practices.
It didn't work.

Above is a slide from Oracle's closing argument. In June, a Los Angeles federal jury considered whether or not Led Zeppelin's "Stairway to Heaven" was ripped off from a song by psychedelic rock band Spirit.

The jury found in Led Zeppelin's favor, quelling some fears that the music industry may continue to be plagued with copyright lawsuits over similar-sounding songs.

The case followed a high-profile 2015 trial in which a jury found that the hit song "Blurred Lines" infringed the copyright of Marvin Gaye's "Got to Give it Up." In July, US prosecutors charged Artem Vaulin, a 30-year-old Ukrainian man, with criminal copyright infringement for running the popular website KickAssTorrents.
Vaulin was arrested and is being held in Poland awaiting extradition.
It's the highest profile criminal copyright case since the US charged Kim Dotcom—who's still living in New Zealand, where he's desperately hoping to avoid extradition.

Above is a screenshot of the now-shuttered torrent website. On July 21, the Electronic Frontier Foundation filed a lawsuit that's been a long time coming.

EFF claims that the DMCA's ban on circumventing digital locks violates the First Amendment.

Digital locks may need to be sidestepped "in order to create a running critical commentary on... a political debate, sporting event, or movie," all legitimate activities that should be protected by fair use, EFF argues.

The government has asked for the case to be dismissed, and the matter is awaiting a judge's decision. Pictured above is EFF client Andrew "bunnie" Huang, who wants to market a product for editing HD television signals, but is hampered by copyright limitations he believes are unconstitutional. Record label EMI sued MP3tunes, an early music locker service, in 2007, along with its founder Michael Robertson, pictured above in a 2006 photo.

The litigation caused MP3tunes to go bankrupt in 2012, but Robertson kept fighting his battle in court.
In October 2016, the 2nd Circuit appeals court upheld and even expanded EMI's court win—a disastrous result for Robertson and MP3tunes.

Today, cloud music services are thriving.

But the MP3tunes precedent shows that innovators who cross the music industry still must risk paying a heavy price. In an opinion published December 6, the US Supreme Court stopped Apple from collecting $399 million in patent infringement damages from Samsung over iPhone-related design patents.

The high court held that the lower court erred when it allowed Apple to automatically collect "lost profits" damages based on the entire value of a phone.
It was the first time in more than a century that the Supreme Court took a case involving design patents. Pictured above is one of the infringed patents, D618,677, describing a black rectangle with rounded corners. The lawyers behind Prenda Law were denounced in 2013 by a federal judge who called them a "porno-trolling collective" that had abused federal courts.
In December 2016, two of those lawyers, John Steele and Paul Hansmeier, were arrested and charged with fraud and perjury. Pictured above is John Steele's banner advertisement from his old firm, which practiced family law. Two band members of 60's rock band The Turtles, pictured above, have turned the once-obscure issue of pre-1972 songs into a hot copyright issue.

The Turtles sued Sirius XM and Pandora, demanding royalties for their old sound recordings, which are not protected by federal law.
Sirius and Pandora lost key legal battles in 2015, and Sirius paid out a $210 million settlement to record labels.

But the Turtles case went on, and on Dec. 21, 2016 the New York Court of Appeals handed a big victory to Sirius, saying that the state's common law offered no copyright protection for pre-1972 recordings.

The decision may be influential in other states. Nokia and Apple fought each other over smartphone patents between 2009 and 2011, but settled their case. Nokia has backed out out of the smartphone business, but is still licensing its patents, so the two companies are back at war. Nokia has sued Apple over patents in 11 different countries. Meanwhile, Apple has filed an antitrust lawsuit against Nokia, accusing the Finnish firm of working together with "patent-assertion entities"—a.k.a. patent trolls—to "maximize the royalties that can be extracted from product companies."