How to increase Wi-Fi coverage throughout your whole home or office

 

 

 Now that Wi-Fi is becoming an integral part of just about every new piece of consumer electronics, and even many appliances, it is more important than ever to have a usable wireless network throughout your home and office — and possibly your yard. Unfortunately, the typical ISP’s simplistic model of providing a single “all-in-one” modem+router+wireless gateway for your site quickly breaks down when faced with the real-world challenges of serving larger, or sprawling, homes or apartments. This is especially true for buildings with lots of wiring in the walls, lath and plaster walls, or metallic heating or air conditioning infrastructure in the floor or ceiling. The good news is that there are several ways to solve this problem, with varying costs and complexities. We’ll walk you through the available options and provide some perspective on which may be best for you.

Wi-Fi Range Extenders: Cheap, easy, but you get what you pay for

The very simplest approach to extending your Wi-Fi network is to add one or more range extenders. These small units plug into an electrical outlet and then can be configured to repeat the signal from your current Wi-Fi source. They are a wonderful solution if you don’t need much bandwidth, and don’t want to take the time, or spend the money, to deal with a wired device.

For older or slower networks, particularly those with limited broadband speeds from their ISP, one or more wireless extenders may be perfect. However, if you’re getting broadband speeds upwards of 50 Mbps, you probably won’t be able to take advantage of all of it in the areas where you use an extender (many extenders have very large “faceplate” maximum bandwidth ratings, but in real world tests their performance is typically substantially less). I’ve used them mostly to fill in coverage gaps while sorting out a better long-term solution.

Hard-wired Wi-Fi Access Point options

By far the best way to get high-performance networking throughout a larger home or office is to deploy wireless Access Points throughout. Depending on your needs and budget, APs can be anything from an old router or inexpensive consumer devices to a well-organized mesh of purpose-built APs. We’ll walk you through some of the pros and cons of each option.

Using routers as Access Points

Many of us have older routers laying around. Most of them can be used as a wireless Access Point by simply only connecting the LAN port to your main router. Some routers, like Apple’s Airport routers, have a built-in bridge mode, in which case you’d connect the one that will be used as an Access Point through its WAN port to your main Apple router. Re-using a router used to be especially valuable when standalone APs were much more expensive than consumer-grade routers. More than once I’ve bought a router specifically to use as an inexpensive AP. However, now there are a number of reasonably-priced dedicated AP options, so unless you already own a router you can re-use, it probably isn’t worth buying a new one for this purpose.

One other nice feature of using a router as an AP is that routers typically come with a number of LAN ports — allowing them to double as hubs. Dedicated APs usually only have a single Ethernet port, so if you have some wired devices to connect at the same location, you’ll also need a standalone hub.

Using Powerline to get Ethernet to your access points

Often you’ll want to put an Access Point someplace where you don’t have wired Ethernet. There are a variety of products than can supply Ethernet over your existing in-home electrical wiring — a technology called Powerline. They are typically sold in pairs, so one can be used at the transmitting end, and one at the receiving end, and are sold by most networking vendors, including Netgear, Linksys, and TP-Link.

I’ve used several Powerline products, and for the most part they work fairly well, with two caveats. First, if you have a complex electrical system, you may need to make sure that your transmitter and receiver devices are on the same leg of the service, or not isolated from each other by some other components. They also typically can’t be plugged into a power strip. Some include a passthrough socket, so at least they don’t consume an outlet, but many don’t. Since you’ll also be plugging an access point in, you may run out of outlets. Second, they can sometimes drop connectivity and need to be reset. That’s a simple matter of unplugging and re-plugging them. But if they are in hard-to-reach locations, or you are relying on them for un-attended operation, that may be a drawback.

The good news is that Powerline adapters, when coupled with a reasonable access point, will often provide better performance than a simple Wi-Fi range extender. However, like range extenders, none of them actually measure up to their claimed performance. For example, the highest-performance model tested by our sister site PCMag.com, the D-Link Powerline AV2 2000, is rated at over a Gigabit, but benchmarked at just over 90Mbps.

Mesh of Access Points

If you’re willing to invest the time and money, and have wired Ethernet throughout your home or office, then a mesh-like solution of APs is the way to go. Mesh solutions can be administered from a single location, and provide the best options for optimum channel management to avoid interference between units. Until recently, mesh AP solutions were expensive, and a bit tricky to administer — because they have traditionally been marketed to larger enterprises. Now though, tech-savvy users can get excellent price-performance from products like Ubiquti’s Unifi wired Access Points.
Unifi Access Points come in a range of prices and performance levels, and can be powered over their Ethernet connection (using PoE). Unifi offers a powerful, centralized, management console, but it requires a little effort to understand how to use it properly if you’re not an experienced network administrator. The units do a great job of managing radio channels and handing off clients between each other — something that most of us have had to do manually until now. My only gripe is that they don’t offer any additional LAN ports, so if you need both to connect both wired and wireless devices in a particular location you’ll need to add a small hub.

For novice users, or for those who are setting up a new installation and also need routing and a firewall, startups Luma and Eero are launching mesh router/APs designed to be very easy to configure from your smartphone. You’ll pay a little more for the added convenience, and you won’t get the wide range of options you do with Unifi, but a 3-pack of either model is enough for most homes that are large enough to need a multiple Access Point solution.

How much performance do you need?

One lesson I re-learn every few years is no matter what performance level you buy into, it eventually becomes obsolete, or at least superseded. We’ve moved through 802.11b to g to n to dual-band to 802.11ac. 802.11n is more than enough performance for most current devices (with stated maximum data rates between 300Mbps and 450Mbps depending on the number of device antennas — although actual data transfer rates will be lower). However, if you’re buying new equipment and have a reasonable budget, getting 802.11ac support probably makes sense. Most 802.11ac devices claim support for 867Mbps or more, although as usual, real world data rates are lower.

Upgrading to Gigabit (wiring & routers)

It won’t do you much good to upgrade your Wi-Fi to some outrageously great speed if your wired infrastructure throttles it back to 100Mbits. So you’ll want to check and make sure your router can at least switch your internal traffic at 1Gbps. Or you can supplement it with a small Gigabit switch or hub. If your ISP is providing you with more than 100Mbps, then of course you’ll have one more reason to make sure your router supports Gigabit Ethernet both on the LAN side and the WAN side.

Handy tip for upgraders: If you have in-wall Ethernet and find you are only getting 100Mbps over it, check to see how many wires are connected to each jack. 100Mbps Ethernet only required 2 pairs of wires to be connected, but 1Gbps requires all four pairs. If you have extra pairs available, upgrading your wiring to 1Gbps may be as simple as connecting them. That worked in our house even with 20+ year old in-wall Ethernet cables. For patch cables, Cat 5 or better is recommended, although in my experience many Cat 4 cables will also work with Gigabit Ethernet.

Picking the solution that’s right for you

The key here is to decide which connectivity solution will work best for you — simply relaying Wi-Fi with no wired connection, relaying Wi-Fi using Ethernet over your power lines, or using APs wired into your Ethernet. If you need performance and have the budget, and access to wired Ethernet, wired APs are the way to go. If you want as much performance as you can get, but can’t run Ethernet cables to your remote locations, then a Power line solution may be right for you. And if you just want the simplest way to get up and running, wireless range extenders can be installed and configured in a few minutes.

Read More

Vulkan API reaches 1.0, claims broad API support, cross-OS compatibility

 

 

 It’s been some 18 months since Khronos announced the next-generation of OpenGL, and the final version of the spec, Vulkan, is finally ready for deployment. As of today, everything related to Vulkan — drivers, SDKs, and early software support is ready for launch. This is a change from Khronos’ usual practice, which is to announce a new version of the OpenGL API with vendor support following at a later time.


Vulkan is a direct descendent of AMD’s Mantle and it shares the same underlying philosophy as both its predecessor and DirectX 12. What sets Vulkan apart from DirectX 12 is that it supports a much wider range of operating systems, from Windows 7 and 8.1 up through Windows 10, Linux, SteamOS, Android, and Tizen. That’s a significant theoretical advantage for the new API, though Microsoft is doing everything it can to push Windows 7 or 8.1 gamers to adopt Windows 10.

Like DX12, Vulkan is designed to minimize driver overhead, scale across multiple CPUs, and offer developers greater control over how workloads are executed across the GPU. Mobile GPU companies like Qualcomm and Imagination Technologies have both pledged support for the API and intend to support it in their own development tools. In theory, this broad industry base could make it more likely that PC games will cross over to Android devices, Steam Machines, and Linux as a whole. Even Vivante has pledged to support the API, though its products are typically used in lower-end mobile hardware.

The Khronos announcement contains quotes from game developers, device manufacturers, GPU companies, and Intel. There’s one company that’s made conspicuous by its absence, however: Apple.
Last year, we theorized that Apple’s embrace of Metal as its low-overhead API in both OS X and iOS meant that the company wouldn’t be supporting Vulkan. That now seems to be the case. Apple could, of course, add Vulkan support at a later date. AMD, Intel, Nvidia, and Imagination Technologies have all announced support for Vulkan, which means it’s up to Apple whether or not to include the option. For now, however, Apple seems to be content with Metal, which means developers will still need at least two APIs to target the entire computing market.



Google has pledged to treat Vulkan as a first-class API, and the new project will implement a “feature set” concept similar to the feature levels that DirectX currently uses. Vulkan won’t be split by an embedded versus desktop distinction the way OpenGL ES split from OpenGL. Instead, Khronos will define feature sets for Linux and Windows, while Google will define them for Android. These feature sets will be based on developer feedback and have not yet been determined.

At least one game, The Talos Principle, is definitely being ported to Vulkan, and we expect the Source 2 engine to support the new API broadly as well. If Valve throws significant weight behind it, Vulkan could become the preferred option for games written for that engine, which would give its chances of competing head-to-head with DirectX 12 a significant lift.

Read More

Samsung announces new 14nm, octa-core SoC: Exynos 7 Octa 7870

 

 Mobile World Congress is right around the corner, and that means the various SoC vendors are gearing up to talk about their next-generation hardware. Yesterday, Samsung announced a new Exynos processor built on its 14nm technology and aimed at midrange devices — the Exynos 7 Octa 7870.
Samsung’s verbiage implies that this new chip will tap the company’s second-generation 14nm technology (14nm LPP), rather than the 14nm LPE process that powered both the Galaxy 6’s Exynos 7 7420 and Apple’s iPhone 6. The new midrange SoC will feature eight Cortex-A53 cores clocked at up to 1.6GHz; the Korean manufacturer claims that these chips consume 30% less power than its 28nm HKMG technology while offering identical performance. This appears to be Samsung’s first foray into eight-core Cortex-A53 processors — other companies, like MediaTek and Qualcomm, have launched a number of chips in this family, but Samsung has held off.

It’s not clear how overall performance will compare against other midrange Samsung SoCs. The company’s previous 28nm chips were typically Cortex-A15 + Cortex-A7 cores in a 4+4 big.Little configuration. Data from Anandtech suggests that the Cortex-A53 is as much as 49% faster, clock-for-clock, than the older Cortex-A7 (a direct comparison against the Cortex-A15). Compare the SPEC and Geekbench results for the Cortex-A15 against the A53, and it’s clear that the 64-bit low-power CPU doesn’t compare well against the 32-bit Core-A15 in terms of absolute performance.

Samsung is already expected to launch the Exynos 7 7870 into an updated version of the Galaxy J7 (2015 version pictured here).

While I don’t expect an octa-core A53 to be a high performer, there’s a positive side to this: Samsung is clocking the Cortex-A53 at just 1.6GHz. That’s significantly lower than we saw from MediaTek at 28nm, where its CPUs often hit 1.9 – 2.1GHz — and that means these new cores should be quite power-friendly.

I’m not going to make any promises before we know if smartphones bound for the United States will use this SoC, but Samsung’s 1.6GHz octa-core might make for a great midrange smartphone with modest top-end performance but excellent battery life.

It may face some steep competition from Qualcomm when it launches. The upcoming Snapdragon 625 is also an eight-core Cortex-A53 design built on 14nm LPP at Samsung, but with a top speed of 2GHz. Expect to see both devices duking it out for market share between the two companies, though they may be confined to Asian markets, which have traditionally put more emphasis on CPU core counts.

FCC data suggests that the upcoming Samsung Galaxy J7 will pack a 3300 mAh battery, substantially more than the 2,550 mAh battery used in devices like the Galaxy S6. Combined with the power improvements at 14nm, these new phones should have excellent battery life.  Samsung’s Galaxy S7 is expected to debut the company’s first custom CPU core, but Samsung is leaning on ARM for its midrange products and designs.

P.S.: Samsung, please do something about your product nomenclature before next year rolls around and we’re all forced to say “Exynos 8 Octa” or something equally silly.

Read More

Google, Microsoft back Apple on encryption; Trump blasts Apple

 

 Earlier this week, a federal judge ordered Apple to provide the FBI with a custom version of its operating system that can bypass the encryption Apple built into iOS 9 and prevent the phone from deleting its own data after 10 failed attempts to unlock the hardware. Apple’s CEO, Tim Cook, released a strongly-worded letter in which he vowed to fight the order in order to protect Apple customers’ right to privacy.
Now, other major voices in the tech industry are publicly backing his statements, while politicians decry Apple’s actions. Google’s CEO, Sundar Pichai, released a statement of his own, saying: “We build secure products to keep your information safe and we give law enforcement access to data based on valid legal orders,” Pichai tweeted. “But that’s wholly different than requiring companies to enable hacking of customer devices and data. Could be a troubling precedent.”

Google CEO Sundar Pichai

The Reform Government Surveillance organization, of which Microsoft is a member, released a statement that indirectly backed Apple, even if the group didn’t mention the company by name. It recognizes the need for legitimate law enforcement activities, but states: “[T]echnology companies should not be required to build in backdoors to the technologies that keep their users’ information secure. RGS companies remain committed to providing law enforcement with the help it needs while protecting the security of their customers and their customers’ information.”

Donald Trump weighs in

None of this sat well with Donald Trump. During an interview on Fox and Friends, the presidential candidate blasted Cook for his stance on encryption. “To think that Apple won’t allow us to get into her cell phone, who do they think they are? No, we have to open it up.” (The phone in question actually belonged to Farook, the man involved in the San Bernardino shooting, not his wife, Tashfeen Malik.)

“I agree 100% with the courts,” Trump said. “In that case, we should open it up. I think security overall — we have to open it up.”
I suspect Trump’s statements will resonate with many people, if only because the San Bernardino shootings, which killed 14 people, were the worst terrorist attack within the United States since 9/11. The FBI wants Apple to unlock a terrorist’s property so it can conduct counter-terrorism operations. Why would Apple stand in the way of that?

Why Apple won’t back down

It’s important to understand the precedent the government is attempting to create with this case. Apple has been ordered to create a modified iPhone software file that does three things:
1). Bypass or disable the auto-erase feature that deletes all phone data after 10 incorrect login attempts.
2). Enable the FBI to automatically submit mass codes in an attempt to crack the device without being required to type them on-screen.
3). The software must not introduce a delay between entering passcodes other than that required by the underlying hardware.
The court order mandates Apple must customize this software in a way that only allows it to run on the target device. This is clearly meant to assuage concerns that the software might be deployed indiscriminately — but it’s not clear that Apple can actually implement that restriction in a way that absolutely guarantees the file couldn’t be modified to run on other devices.
Apple, in other words, is being ordered to create software that deliberately compromises its own product’s security. While the judge clearly believes that the software can be kept from leaking to the Internet, history tends to suggest otherwise.

Magic backdoors that only open for the right key one day of the year belong in stories about elves and dragons

Once this precedent is set, it will be used. Again, history is an effective guide. Law enforcement officials in the United States have justified extensive surveillance programs and data sweeps because the 1979 case Smith v. Maryland allowed investigators to install and use pen registers. A pen register is a device, typically installed at a telephone company’s central offices, that records the outgoing phone numbers dialed by a single user (but not the conversations).

Nothing in the text of the Smith v. Maryland decision suggests that the justices envisioned a future in which their ruling would be used to justify mass surveillance of the entire Internet. Yet here we are.
Apple isn’t fighting this order because it wants to shield terrorists or prevent the FBI from investigating this specific case. It’s fighting the order because it knows that if the order stands, it will be flooded with demands for device unlocks both in the United States and abroad. The software will eventually leak, it will be abused by law enforcement, and the end result will fatally compromise iDevice security. Google, Microsoft, and every other smartphone vendor will be roped in, in very short order. The government would doubtlessly order Intel, ARM, and every other security vendor to implement similar procedures in their own products (think TPM on Intel systems and ARM’s TrustZone).

Device security is challenging enough already. Forcing companies to compromise their own encryption standards in the name of safety will not improve the situation.

Read More

New iOS 9.2.1 update un-bricks iPhones, fixes ‘Error 53′

 

 Earlier this month we reported on Apple’s “Error 53″ message and how it bricked perfectly functional iPhones without warning, often after the devices were upgraded to iOS 9. Apple eventually clarified that the problem, which was limited to devices equipped with Touch ID, occurred if a third-party performed repairs to the device and replaced either the home button or the cable that connected the Touch ID sensor to the smartphone’s logic board.

Now, Apple has released a new build of iOS 9.2.1 that will “un-brick” devices previously crippled by Error 53. This new build of the operating system doesn’t include any other new features or capabilities, and it’s strictly meant for customers who currently have a $600 rectangular paperweight sitting somewhere around the house.

Initially, Apple defended bricking customer devices by claiming it did so to ensure Touch ID and Apple Pay were kept secure. This never made much sense, given the length of time that could pass between when a customer repaired a device and when the device failed. A security system that doesn’t lock down an insecure device for eight to 12 months isn’t particularly secure.

Error 53

Now, Apple has backed away from this explanation altogether. In a statement to TechCrunch, Apple explained the problem thusly:

This reports as an Error 53 in iTunes and appears when a device fails a security test. This test was designed to check whether Touch ID works properly before the device leaves the factory.
Today, Apple released a software update that allows customers who have encountered this error message to successfully restore their device using iTunes on a Mac or PC.
We apologize for any inconvenience, this was designed to be a factory test and was not intended to affect customers. Customers who paid for an out-of-warranty replacement of their device based on this issue should contact AppleCare about a reimbursement.

There is one caveat to this update: Touch ID will no longer function on an iPhone that’s been repaired by a third party. First, kudos to Apple for doing the right thing on this issue. Disabling Touch ID if the component doesn’t authenticate properly is the sane way to deal with the problem, particularly when the alternative was bricking a device with zero warning or explanation. Given that Apple doesn’t license any third-party stores to provide repairs in the United States, this entire situation looked like an attempt to force people to pay whatever “official” rate Apple wanted to charge for device repair, while freezing small competitors out of business.

The Touch ID sensor. Image by iFixit

Giving users the option to decide if they want to pay for Apple to fix their Touch ID system is a much better way to deal with the problem. We’re also glad to see the company acknowledging users who were forced to buy new devices and offering them reimbursements for the expense. Presumably they won’t have to return the devices they purchased, since the original models were likely thrown away.

Read More