Welcome to the TeraStor

For years I’ve been slotting 2-terabyte drives into a droboPro to store my media collection and backups. While it worked well, the droboPro never entirely fit the bill:

  • I could never get Ethernet to work reliably, and tended to leave it connected via USB.
  • The drobo is a “dumb storage device” and couldn’t run apps.
  • It took up a lot of desk space.
  • The drobo’s fans tended to run loud, especially with seven drives installed.

As I pondered filling the last of the droboPro’s slots with another drive, I was also considering other options; in the end, I made the plunge and built a full custom NAS system. Doing so was an excellent challenge and learning process… so I decided to write about it here.

What’s a NAS?

Network Attached Storage is usually a purpose-built computer operating as a home server: in this case, a server used to store and share media files. A NAS typically runs “headless”, i.e. without a monitor, keyboard, or other peripherals, running on your network 24/7.

MotherboardC2750D4I motherboard, shown at an angle

While there are an increasing number of all-in-one pre-built NAS solutions available, I decided to build mine from scratch. This is not for the faint of heart: you should have some familiarity with computer assembly and the command line before attempting a NAS build.

Brian Moses’ 2015 DIY NAS was a big inspiration for the build. I used a ASRock C2750D4I motherboard rather than Brian’s C2550D4I due to the C2750D4I being the same size with double the number of cores (useful for transpiling video and running background apps). I chose FreeNAS as the server, as the Debian OS was comfortable and familiar.

The AsRock mini-ITX board packs a lot of features into a diminutive package:

  • 12 SATA ports (4 × SATA2, 8 × SATA3)
  • Dual Onboard Intel Gigabit Ethernet with teaming
  • Max 64GB RAM

My only issue with the board was the fact that it was limited to just three USB ports (two at the rear + 1 header on the board), and that they were USB 2, not 3. This made the FreeNAS installation slightly tricky (see below), but I could always add a USB 3 card via the single PCI-E slot if I needed to.


Three hard drives, shown overlapping

I wanted to start big in storage to contain the files from the drobo and leave the new device alone for as long as possible. I invested in three Western Digital 6TB Caviar Red drives, designed specifically for NAS applications; the 2 ~ 4TB drives out of the droboPro would be used for physical offsite backup.

FreeNAS can be installed and run from CD or USB. Not needing an optical drive I would use only once, I opted for the latter. That meant buying a small, dedicated USB stick to store and load the OS (I used a SanDisk Cruzer 16GB low-profile USB stick)


FreeNAS needs 8GB of RAM to operate, with 12GB total usually recommended for plugins or “jails” (FreeNAS apps). 12GB of RAM provides coverage for up to 32TB of raw storage; as hard drive pool size increases, so too does the need for RAM. A very rough rule of thumb for FreeNAS recommends adding 1GB of RAM for every terabyte of added storage.

This was complicated by my discovery that ASRock boards are somewhat particular when it comes to RAM. Looking at the Qualified Vendor list for the board, I found that Memphis Memory made 16GB DDR3L ECC SODIMMs that were recommended; I bought two, leaving two slots free on the board, which should provide me with enough RAM for 64TB of storage.

CaseAluminium case, shown at an angle

For the case I went with Silverstone’s tiny DS308, which offers room for up to 12 drives, eight of them hot-swappable 3.5″ devices, and sits neatly on my desk.


Since the NAS is usually running 24/7, the build needed a cool, efficient power supply. I settled on a Silverstone 80-plus certfied Gold 450W ST45SF-G.


I’ll go into more detail about how I set up the NAS in the next article; for now, I thought I’d mention some oddities to the setup, just in case others follow this build:

  • The AsRock board takes a surprisingly long time to POST: about 30 seconds. Be patient.
  • The board also assumes the first time it is run that it is doing so remotely: it won’t start without a cable being plugged into its IPMI port. (After the machine is configured, you can detach the cable from there and move it to one of the standard Ethernet ports).
  • Deciding to install FreeNAS from USB presented a quandary: both rear USB ports were occupied by the final USB storage for the OS and a keyboard. I had to purchase a USB header for the second USB stick, attach it directly to the board, and tell the machine to boot from that in order to install.

It’s been two months with the build now, and I couldn’t be happier: media streams to every computer in the house, seamlessly. There are some improvements to make, however… I’ll leave the details of those, together with more information about FreeNAS and Plex, for the next article.


In Which I Spend Far Too Much On Ethereal Objects

I have a sickness: I fund far too many Kickstarter and IndieGoGo projects. But I am trying to change: in the past I’ve funded projects solely because I wanted to see them succeed, whereas now I’m paying money to ventures for products that I actually intend to use. A quick survey:

  • My support for Star Citizen continues: my collection of (currently unflyable) digital spaceships has grown to include an Aegis Retaliator.
  • skylockI have a Skylock on order: a solar powered, iPhone-activated bike lock with theft attempt alerts, and my deposit for a Lily drone camera is still in place.
  • I’ve backed probably the nerdiest thing I’ve ever seen: a reissue of the original 1975 “worm” NASA graphics standards manual.
  • I’ve been a big fan of ScotteVest multifunction garments for a long time, and I’m always looking for more, so it was natural for me to back the blazer cut of “The World’s Best Travel Jacket
  • As part of my web design and development teaching, I’ve funded Richard Rutter’s handbook of Web Typography.
  • Finally, I’ve backed the construction of thDome Lighte Dome Light 2.0, a steampunk-esque LED Edison light.

I’ve been very tempted to buy an Electric Object, but the current
physical limits of the hardware (particularly an ugly, thick colored power cord, making wall hanging unreasonable) and the current exchange rate places it just out of reach at the moment.

Man’s Best Friend: A Tag-Along Drone

For some time I’ve wanted a portable, throw-in-the-air-and-forget tag-along drone to take remote video and photographs from above. Lily looks like it could finally fulfill that role.

For some time I’ve wanted a portable, throw-in-the-air-and-forget tag-along drone to take remote video and photographs from above. Lily looks like it could finally fulfill that role.

Lily is an offshoot of an interesting split that’s happening right now in prosumer drones. Like most technologies, UAVs started in the military and slowly made their way into the hands of the public. The civilian enthusiasts that used them were often pioneers in the application of drone technology, and very good fliers; they’re also the ones most likely to look in disdain at the cheap RC knockoffs solid in big box stores, sold to people that use them for a weekend before storing them in a closet (or worse, lose them due to foolish behaviour).

Right now, there are several companies competing to make the next big switch: to become the “Apple” of the drone consumer market, with products that are well designed, low maintenance, and easy to operate. DJI is probably the best known of these, but Lily is another.

Lily takes the “set and forget” aspect of such drones even further: the user wears a wristwatch-sized location device, and simply throws the man-portable Lily into the air. For 20 minutes, the drone follows the operator with a high-resolution camera from five to 50 feet above their head, and up to 200 feet away. Waterproof, the drone can be used by surfers and skiers, in all weather conditions, operating against a 20 mph wind.

While it will never replace a drone flown by an experienced operator, I can think of dozens of shots where Lily could be used. It’s ideal for lone photographers like myself, especially for taking action sequences without a crew or camera operator.

On a far larger scale, there’s a strong debate to be had as to whether the society goal of “a drone in everyone’s pocket” is a worthwhile one to persue. I can absolutely see a time when people being surrounded by a small fleet of nano-drones will not be seen as unusual in the slightest. Conjecture on just what that society will be like I shall leave for a future article.

Lily is being pre-sold for $499 US until June 15, with delivery expected in February 2016; full retail price is expected to be $999.

An Alternative Future For the Canadian Air Force

Today being Canada Day, with the Snowbirds flying in formation over North Bay, seems an appropriate opportunity to discuss the future of the Canadian Air Force. Previously I have discussed why the F35 is not an appropriate aircraft for Canada’s 21st century military. In brief: the airframe is extremely expensive and unsuited to Canadian conditions, the country lacks the infrastructure to support the aircraft during sustained combat operations, and the fuel limitations of the F35 means that it cannot reach the Northwestern Passage, an area of increasing strategic interest for Canada and the most likely region to see future conflict.

My alternative for the Canadian Air Force is a radical departure from traditional military organizations, but one that I believe has far greater benefit to the nation over the long term. I would propose that the Canadian Air Force concentrates its human assets on rescue, interdiction and troop support, and transition the major portion of every other aspect of its mission, including surveillance and combat, to drones.

This proposal runs sharply against the heroic image that the Air Force (particularly the fighter wing) has nurtured since WWI. But a modern military does not win battles on romantic heroism. It might recruit with such ideals, but a military force faces its practical test on the battlefield; and encouraging a solution derived from a romantic “knights of the air” mystique against 21st century threats exhibits the same mindset, and long-term effectiveness, as the Polish army mounting cavalry charges against German invaders in 1938. The military succeeds in its mission because it uses the best tools, intelligence, and application of force available: and Canada’s mission, especially in the Arctic, is most effectively prosecuted with drones.

It may surprise most to learn that Canada has a long experience with drones: the CL-89 “Midge”, developed by Canadair in the 1960’s, was one of the first drone systems successfully deployed in a war zone, and was purchased by several countries. This was followed by the radically designed contra-rotating CL-227 Sentinel.

Starved of a sustained program of government funding and support, Canadian innovation in UAV technology has lapsed since in the 1980s. In Afghanistan, Canadian forces are dependent on outsourced technology:  French SAGEM CU-161 Sperwer UAVs have played a surveillance role; more recently, Heron models, purchased from Israel, have participated in offensive action.

This leaves Canada at the mercy of foreign sources for UAV technology: the cancellation of the Sperwer program left the Canadian military with no drone support in Afghanistan until the Heron was brought in. The JUSTAS program, now in Phase II, seeks to determine the next classes of UAV that Canada will purchase, and has included transcontinental overflights of Northrop Gruman Global Hawk drones. All the drones currently in competition for JUSTAS are foreign, primarily of US and Israeli manufacture.

Let’s look at some of the advantages UAVs provide:

Photograph of two pilots in a UAV control roomReduced or eliminated Canadian causalities during conflict
Because they are piloted remotely, the most dangerous part of a drone crewman’s day is driving home after a shift.
The skill set to pilot a drone is already built into the vast majority of Air Force applicants.
It takes years to train a good fighter pilot, an investment of millions of dollars. But anyone who has played Microsoft Flight Simulator can run a drone after just a few weeks of training.
Photograph of an ultralight drone in flightDrones have far greater range and can remain on station for far longer periods
Due to the fact that they are not man-rated, drones can be “long-loft” devices, remaining in flight for days at a time, depending on the propulsion system used. Global Hawk drones have crossed Canada during service trials, with no need for mid-air refuel.
Because they can be built far smaller, drones have a greatly reduced radar and visual signature.
Adding stealth technologies to a UAV is far more effective and better understood. Depending on how they are constructed and powered, UAVs can fly higher, lower and longer than fighter jets, and faster or slower (the latter being an advantage for surveillance and ground support). For example, the Zephyr UAV has been aloft for more than 60 continuous hours in trials.
Alternative fuels are an option
Napoleon Bonaparte claimed that an army marches on its stomach, but a modern military runs on petroleum. Drones do not require jet fuel: prop-driven UAVs can run on biofuel or solar. Intelligent systems onboard drones can use thermals and rotary wind shear to gain lift; modern aerodynamic design can provide extremely high glide ratios and operational ranges that do not require power of any kind.
Lower risk, lower loss, lower cost.
The most expensive UAV system available, the Global Hawk, is less than ⅕th the price of an F-35. Most systems are far cheaper: Canada could order hundreds, even thousands, for the price of one fighter jet.The physical destruction of a drone due to an accident or enemy action is easily compensated for; the loss of a $300 million airframe and a fighter pilot with years of training, from a fleet of only 50 aircraft, not so much.
Drones can, in theory, outperform any piloted aircraft.
The performance envelope of a drone is determined by material strength and thrust. In a traditional fighter, the limiting factor is the physical endurance of the human pilot.
A greater return on investment
Drones can be built, completed and maintained almost entirely within Canada, using local tools and talent. Rather than a few Canadian companies being sub-contractors to the F-35 program, the nation can build the drones in-house and benefit directly from the technological innovation.
Drones are far more adaptive to changing battlefield conditions and scope of conflict
A peacekeeping mission is far better (and faster) complemented with a dozen drones and a maintenance crew of a few personnel than the massive commitment required to support a single F-35. Drones can easily be ruggedized for Arctic missions.
UAVs are far more suited for 5th generation warfare
The fourth generation of warfare, which we are currently experiencing, is characterized by non-traditional asymmetrical forces on a fluid battlefield without defined fronts or differentiation between military, civilian, and government. To this, fifth generation warfare will add “swarm” attacks, rather than the force-on-force engagements that have characterized warfare for the last several thousand years.“Swarm” units are loosely coupled, self-arranging, and attack on all vectors: physical, electronic and psychological. Some swarm units will be human, others remotely controlled, and some entirely artificial and autonomous. Fifth generation warfare overwhelms, confuses and demoralizes traditional military forces, demonstrated perhaps most notoriously and effectively during 9/11 and in General Paul K Van Riper’s defeat of an entire carrier group using only small craft during the Millennium Challenge 2002 war game exercise.

UAV technology is advancing rapidly; by no longer supporting a drone program directly, Canada misses out on the secondary and tertiary effects of research and innovation. Provided with solid funding and long-term commitment, Canadian designers and researchers – artificial intelligence programmers, aeronautical engineers, and scientists – could develop drones in a 10-year timeframe that were optimized for native conditions, and cover a wide range of purposes.

In my opinion, Canada should invest in the absolute minimum number of fighter airframes required to replace the CF-18 fleet and fulfill our commitment to NATO – I would suggest the tried, proven, and far cheaper F/A-18E/F Super Hornet – while adding a few A-10s for ground support, with the rest of the Air Force budget devoted to developing UAVs for combat and surveillance and a C3 support system.

The F35 Joint Strike Fighter: The Wrong Tool For Canada’s Air Force

Now that the Tories have been elected to a majority in the 2011 Canadian government, the controversial proposal to purchase 65 American F-35 fighter jets for the Canadian Air Force is likely to move forward.

The F-35 program has been deeply troubled since its inception. Aside from the projected cost for Canadian taxpayers – almost $30 billion over the 30-year lifespan of the airframe (far larger than the initial estimate of $9 – $16 billion) – the Canadian government’s very participation in the program is a fait accompli: no competitive bidding process to replace Canada’s aging CF-18 fleet was ever entered into (ignoring Ottawa’s own contracting regulations), and no alternatives were ever seriously considered. If Canada stays in the program, the nation pays whatever price the F35 is delivered at.

Most media coverage has focused on the ballooning costs of the program and the lack of a competitive open bidding process. But purchase of the F-35 is wrong for Canada for far more fundamental reasons.


The F-35 is a high-thrust interceptor designed primarily for air-to-air engagements with targets beyond visual range and limited targeting of ground installations after most detection infrastructure has been destroyed, neither function ever used by the Canadian military. That does not mean that the airframe cannot be adapted into other roles, but its suitability for domestic service must be seriously questioned.

Canada’s Future Strategic Concerns

Global warming is pushing Canada’s interests northwards: rising temperatures are opening up the NorthWest Passage to shipping during the spring and summer. Increasingly ice-free areas also allow growing access to resources in the region, including oil, gas, and fishing.

Sovereignty over most of the strait is disputed; Canada considers the area part of her internal waters, meaning that the country could restrict access, inspect and levy cargo, fine polluters, and limit research. The US and many European counties consider the Northwestern passages an international strait. The area will be the primary strategic area for Canada in the 21st century.


If the F35 did enter Canadian service, its most appropriate armament for engagement in northern waters would be the Joint Strike Missile, which would have to be purchased from Norway (Canada is not a development partner for the weapons platform). The range of the JSM is 240 kilometres, and the F35 is only able to carry two of them (more could be carried externally, compromising the F35’s stealth capability).


Effective range of F-35 with external tanks from Canadian Force Bases with airfields
Effective range of F-35 with external tanks from Canadian Force Bases with airfields




The effective radius of the F-35 is 590 nautical miles, or 728 nm (1348 kilometres), with external tanks. That is at cruising speed, not full combat thrust. Adding external tanks also degrades the stealth capability of the airframe.

The F-35 can be refuelled while in flight, but not from Canada’s two CC-150 Polaris tankers, at least not in any way the F-35 is currently configured. The tankers are principally stationed at 8 Wing Trenton in Ontario, and have a fully-loaded range of 4,630 kilometres. And if an F-35 does successfully mate with a CC-130 it loses its entire stealthed profile.

Canada’s Air Force Wings

There are Air Force bases across Canada, but in regards to the Arctic only three are relevant: 5 Wing Goose Bay, 14 Wing Greenwood and CFB Gander. (The planned military training centre in Resolute Bay supports only Army personnel; the deep-water port being built in Nanisivik, on the northern tip of Baffin Island, only Navy).

There are four very small Forward Operating Locations north of the 60th parallel, but they remain deactivated except for brief visits. The FOL’s are in Rankin Inlet, Nunavut, Inuvik, Northwest Territories, and Iqaluit, Nunavut.  440 “Vampire” Transport Squadron, operating out of Yellowknife, is the only permanently based Air Force asset in the north.

Adding F35’s To Canada’s Air Force Means No Effective Force Projection Into The Northwest Passage

As you can see from the map above, the operating radius of the F-35 from Goosebay, Greenwood and Gander, even with external tanks, is severely limited.

Effective range of F-35 on internal tanks only from Forward Operating Locations
Effective range of F-35 on internal tanks only from Forward Operating Locations

Deploying the aircraft from the Forward Operating Locations – of which Yellowknife is the only one that is permanently staffed, and all of which would require major upgrades in order to house and maintain F-35’s – still does not provide complete coverage of the Northwest Passage for surveillance or combat. Pushed to their limit and given an exact target location, the aircraft would be able to dip below cloud cover, take one shot, and then be forced to immediately return to base. Scrambled from the FOLs under full combat thrust, they would never make it to the strait at all.

Arctic Night and a Single Engine Is Not a Good Combination

As the Canadian Air Force is pushed further north it will encounter harsher conditions, higher maintenance costs, and inevitable loss of both airframes and lives. The most extreme example of this is CFS Alert, an armed services station for signals and intelligence interception. Alert is the northernmost facility in the world that is permanently inhabited by human beings, who’s motto is “The People of the Land Beyond the Land Beyond”. The base is in complete darkness from mid October to the start of March, and experiences polar weather year-round; these conditions have caused the loss of two aircraft and fourteen lives in a period of 40 years.

The F35 is a single-engine aircraft; if that engine fails to light while in the air the pilot has little choice but to eject (like most modern fighters, the F35 does not have the ability to glide unpowered; its continued flight depends on constant thrust).

In summary, the F-35 is an airframe that is hugely expensive and remains untested in Canadian conditions. It is neither appropriate for Canada’s traditional military role of the last 60 years – that of peacekeeping support – nor can it fulfill the requirements for the nation’s strategic interests over the next half-century.

As I don’t believe in criticism without providing alternative solutions, an upcoming article will suggest a very different, less expensive and far more effective alternative for Canada’s Air Force.

Countdown to Skynet

The X-47B made its maiden flight a few days ago at Edwards Air Force base. What’s the X-47B? Just an autonomous, artificially intelligent, fully armed jet fighter.

At the same time, automated, human-seeking machine guns are patrolling the Korean DMZ, and researchers are developing an “Internet back channel” for robots to communicate with each other.

My New Geek Heroes

Forget the well-funded independent commercial American space operators like Space-X,Scaled Composites or XCOR. My new personal geek heroes are a small group of two dozen amateurs working out of an abandoned shipyard on a small island off the coast of Denmark. Funded by PayPal donations and the sale of T-shirts, Copenhagen Suborbitals are building a rocket as large as the V2 that bombed London during World War II. Oh, and in his spare time the founder built a frickin’ submarine.

Nautilus submarine