I am converting one of my spare bedrooms that previously was used for storage and my drumkit into a video production studio.
Rackspace is an American based web hosting company that, for many years, was touted as the premier managed web server hosting provider. They have never been “cheap”; however, in my opinion, you received far more than what you paid for, and their tech support was second to none. I guess that is why they previously branded themselves under the motto of “Fanatical Support.”
One of the tech startups I co-founded has been a Rackspace customer for a little over ten years now. In the beginning, we started with a single dedicated webserver, adding a dedicated Cisco firewall soon after. Within a year or so, we had outgrown our simple single webserver and upgraded and expanded throughout the years into multiple dedicated webservers to handle our needs.
As Rackspace expanded its product lines to cloud servers, email box hosting, and MongoDB services, we quickly signed up and increased our monthly financial commitment to them. I know we were never their largest customer by any means; however, in 2019, I have spent just over $82,966, so we are not their smallest customer either.
For many years we loved our service and happily (as happy as you can be to ever pay a bill) paid each month. Then . . . a few years ago, something happened. It is hard to put an exact date on when the change occurred; things were not the same as years past.
At first, the tech support would take a little longer to get back with you. Gone were the 20-minute support ticket replies; now it would be an hour or two before they emailed you back. Not usually a problem, but still a little annoying. Then I started to notice that on the weekends or late at nights, you were lucky if you heard anything back on your support ticket. You just accepted the fact that you had to wait until regular business hours to get a reply.
Many of the server techs that I knew by name, the same guys (and gals) that also knew my systems intimately; all disappeared, replaced with inexperienced technicians that were only barely a step above “did you plug in the computer?” level of support. Each time I chatted or spoke on the phone, I would have to explain everything in explicit detail and almost guide them along on how to do their job. I could no longer rely on them being the expert.
Things started to get worse, but we stuck with it.
Even with the less than stellar level of support, it was just easier to leave our webservers in place, upgrade them as needed, but not to make any drastic switches, as doing so would be a considerable project.
Back in 2015, our servers were hit with a distributed denial-of-service attack (DDoS attack) which took our websites, and the websites of our customers offline. The attack was intermittent and would come in waves of 2-3 minute attacks every 30 minutes to an hour or so. This gave just enough time for the webserver to come back online and let customers login, then bam, it was taken offline again.
After about a day of this on again off again DDoS, one of the Rackspace reps recommended a premium service they offered called ‘DDoS Mitigation.’ This premium service was an additional $1,500 per month and was like an insurance policy, “you hope you never need it, but if you do, it is there for you.” They explained their mitigation service was continually monitoring my servers 24x7x365. If another DDoS attack came in, they would react instantly to redirect the malicious traffic away from our infrastructure, keeping our network up for our customers.
We signed up for the service and moved on with life. Luckily for us, we went another three years without a single DDoS attack. We knew we were paying $1,500 a month for the service and haven’t used it, but it was an investment in the peace of mind knowing our servers would stay online, and our customer’s websites would not go down.
The Final Straw
Now we flash forward to Aug 2019; by this time, we have had the DDoS mitigation service for about three years and spent about $54,000 in total just on the DDoS mitigation service. On this day, we were hit with another DDoS attack. For comparison, the attack that first prompted us to signup for DDoS mitigation I will refer to as “1X” strength. August’s attack as 10X in strength.
Luckily, while the attack was “more powerful,” it was not very sophisticated. I was sitting at my desk and could hear our office tech support phone lines begin to ring like crazy. I knew there was a problem. Within a few minutes, I was getting frantic IMs sent to me from my tech support staff, informing me the server was down. I did some quick investigations and discovered we were under attack. I was able to quickly isolate the website that was targetted in the attack and was able to DNS that site over to Cloudflare’s network.
Our servers were down for about 30 minutes before I was able to get the issue mitigated on my own. Watching the Cloudflare reports, I could see the DDoS continued off-and-on for a few days, thankfully Cloudflare was able to prevent that traffic from hitting my network.
Then I contacted Rackspace to politely ask them, “WTF? Why didn’t the DDoS Mitigation kick in and protect us?” Their response, if not so infuriating, was laughable. “Ohh, it wasn’t big enough for us to even notice.”
“Excuse me, what?”
They acknowledged it was a more massive attack that the attack I had years ago, but since I mitigated it “so quickly,” they didn’t have the “time” to do anything. Wait? What? Our servers were down for 30 minutes, and they sold me on “24x7x365” monitoring. That was why I paid $54,000!
We would go on to argue with Rackspace for the next four months that I wanted to cancel the DDoS mitigation because it was useless. The only option they gave was for me to “buy my way out” of the contract. Even after I showed them quoted emails from their staff assuring me the DDoS Mitigation would protect me, they didn’t care.
Finally, after four months of arguing, they allowed me to cancel my contract, a whopping two months early. Big deal, thanks for nothing, guys. But they refused to give a refund for the prior four months that I had asked it to be canceled while they stonewalled.
So now, for $6,000 they stole and refused to refund, they are willing to give up on $80,000+ per year in revenue. This is even after I had been working with my sales rep to purchase more servers and bring my cost to about $100k per year.
Rackspace’s service has gone to shit over the years. I was putting up with it as it was “just easier” than switching everything. This was the final straw. It wasn’t that the $6,000 was a “make or break” issue for me financially. It was the principle of it. They knew they had sold me a service that would not provide the level of service they promised. Simply put, they did not care.
The Switch to Amazon Web Services (AWS)
I had already been using AWS for a few services that Rackspace didn’t offer for one of my other tech startups, and decided to move all of my web services over to AWS.
In future posts, I will describe my transition to AWS.
Facial recognition is a process of using computer vision based mathematics to detect and recognize a human face in a photograph or video. Using biometrics the facial recognition system maps facial features such as the location and shape of the eye, nose, mouth, distinguishable landmarks unique to the person and other geometric aspects to the face.
Complex mathematical algorithms are used to produce a numerical sequence that represents the face in a language a computer can understand. This “faceprint” is as unique as a human fingerprint and can be analyzed in real time to identify people as they walk past a video CCT security camera.
The facial recognition industry is expected to grow to over $7.7 billion annually in 2022, nearly doubling from $4 billion in 2017.
For privacy-minded individuals, you will want to pay close attention to how and by whom your faceprint data is used. The technology itself is not dangerous and people do not need to be paranoid about “big brother watching you.” Cautious? Yes, but paranoid about facial recognition technology, no.
Just like all technology, those that are criminally minded can abuse facial recognition. Luckily there are also good uses
In the past, most systems required the person to face the camera and have their face correctly framed within the photo. This may work for mug shots taken by the local Sheriff’s Office, but for real-time analysis of CCT security camera footage people as they walk by, newer technology needed to be invented.
Now, many systems can build a 3D representation of the human face based upon multiple photos known to be of that individual, giving the system a more detailed understanding of that person’s unique facial attributes.
All uses of this technology, for good and evil, requires advanced neural networks trained to work together to provide the user with the desired service.
Article Quick Links
- How Facial Recognition Works
- Law Enforcement Use of Facial Recognition
- Bringing Facial Recognition to the Masses
Breaking Down the Steps
Humans are great at recognizing people’s faces; we have had more than 2 million years of evolutionary adaptation behind us. Computers, even with all of their power cannot natively look at a person’s face and say “That’s Aunt Betty.”
Computer vision research started in the 1960s with scientists teaching computers how to detect and recognize a human face. Computers have come a long way since the 60s, but they still must be trained what to look at and how to distill that information down to binary computer language of 1s and 0s.
Before a computer can recognize a face in a photo, it must first determine if a human face is even in the picture. Facial recognition is very expensive in terms of computing power, whereas simple ‘is there a person in the photo’ is very inexpensive. For this reason, systems elect first to verify if there is a person in the photo before attempting the more complicated matching the photo to an actual person’s identity. This process is called Face Detection
1) Face Detection
A photo is analyzed to find the human face(s) within the image. Each face detected is isolated and processed separately. Open source libraries such as OpenCV and dlib are standard; of course, there are also many proprietary systems in use as well.
The face is cropped from the original picture with the eyes and mouth aligned in a common location for all photos. With these attributes in the same general area for all images, it makes the next steps easier for the computer.
The facial recognition software will then generate a unique and propriety identifier for this face based upon the facial landmarks detected. This identifier is usually represented by a series of floating point integers (computer geek speak for a number with a decimal point). In this step, the system detects and extracts all of the meaningful data that it will use to generate your faceprint.
Of course, these numbers do not really mean much to us as mere mortals, to a computer this is what a human face looks like. Each number in the array is a numerical representation showing the location of the eyes, nose, mouth, etc.
With this unique identifier, the facial recognition software will then query a database of hundreds, thousands, or even millions of known identities and return a list of possible matches.
This query to the database is performed by comparing the Euclidean distance between the positions of each array element’s floating point element compared to that of the same element of a different person’s array. If the Euclidean distance of all 128 points is less than 0.6 it is a strong candidate for a match.
Depending on how powerful the system is, in most cases the above four steps are processed and returned to the user in milliseconds. Now a human can manually verify the results returned by the system.
In our example, the girl from ‘Face 1’ is already known to the database and returns her information:
We have covered the basics of how to detect a face in a photo and searched our database to see if that person is known, but we skipped a very important step.
Before you can run Face Detection or Face Recognition, you must create and train computer vision models so that the computer knows what a human face “looks like.” There are some great python wrappers around dlib like face_recognition that are open source and are already trained. Or if you are more adventurous, you can train your own models, I will direct you to an excellent article by Dr. Adrian Rosebrock on how to train a facial recognition model.
An essential piece of a facial recognition system is the ability to intelligently group photographs of the same person into groups, or clusters. This can be accomplished through manual, semi-automated, and fully automated processes. By clustering these photos together, the system can better handle occlusion issues that partially block parts of the face in an individual photo such as sunglasses, hairstyles that cover part of the face, etc.
Clustering can also help by grouping photos of a person from dozens of different angles giving a complete representation of the person’s unique facial structure in 3-dimensional space.
Law Enforcement Use of Facial Recognition
MorphoTrust Idemia, is one of the largest vendors of face recognition and other biometric identification technology in the United States. It has designed systems for state DMVs, federal and state law enforcement agencies, border control and airports (including TSA PreCheck), and the state department.
Back in the day police departments had to manually search filing cabinets and mug books filled with photographs of suspects. Now they can use smartphone apps to take a quick photo of a suspect and instantly have their criminal records and personal information displayed on the screen for their review.
Many of these systems work together with social networking and other integrated surveillance systems to provide a complete look into a person’s identity.
Tech services that assist law enforcement and private sector litigators to search for and combat online revenge porn uploads from ex-boyfriends and ex-girlfriends.
Battling Child Porn
While not technically “Facial Recognition” per se, companies like Microsoft use computer vision algorithms to detect known child pornography online. They offer a free service called PhotoDNA to law enforcement agencies.
PhotoDNA creates a unique digital signature (known as a “hash”) of an image which is then compared against signatures (hashes) of other photos to find copies of the same image. When matched with a database containing hashes of previously identified illegal images, PhotoDNA is an incredible tool to help detect, disrupt and report the distribution of child exploitation material. PhotoDNA is not facial recognition software and cannot be used to identify a person or object in an image. A PhotoDNA hash is not reversible, and therefore cannot be used to recreate an image.Microsoft, PhotoDNA (https://www.microsoft.com/en-us/PhotoDNA)
Bringing Facial Recognition to the Masses
This face recognition technology is used in more products and services that you may realize
Online/Mobile Device Banking
Many national banks like Wells Fargo and Bank of America are using facial recognition in their mobile apps and in their ATMs to better verify and authenticate legitimate users
Thanks to the amazing assistance of AWS Support, I finally was able to install
1. Launch a new EC2 instance:
1) On “Choose AMI” screen, select “Amazon Linux AMI 2018.03.0 (HVM), SSD Volume Type” as your AMI.
2) On “Step 2: Choose an Instance Type” screen, I suggest to select a medium or large type (mine is t2.large). I tried to use t2.micro but the performance is not good enough for compiling
3) Click “Review and Launch”, then click “Launch”.
4) Use your key pair to finalize the configuration.
2. Use ssh to connect to your new EC2 instance with your private kay pair.
3. On your new EC2 Amazon Linux instance, please do following things:
1) Install gcc-c++: sudo yum install gcc-c++ -y
2) Install cmake: sudo yum install cmake -y
3) Install python36: sudo yum install python36 -y
4) Install python36-devel: sudo yum install python36-devel -y
4. Install Python dependencies:
1) Install Pillow: sudo python3 -m pip install Pillow
It will also install “PIL” for you.
2) Install face_recognition: sudo python3 -m pip install face_recognition
It will also install “face_recognition_modes”, “numpy”, and “dlib”.
Above modules will be installed to “/usr/local/lib64/python3.6/site-packages/”. Here is my file list for your reference:
[[email protected] ~]$ ls /usr/local/lib64/python3.6/site-packages/ -l
drwxr-xr-x 2 root root 4096 Feb 22 08:08 dlib-19.16.0.egg-info
-rwxr-xr-x 1 root root 9852368 Feb 22 08:08 dlib.cpython-36m-x86_64-linux-gnu.so
drwxr-xr-x 18 root root 4096 Feb 22 08:08 numpy
drwxr-xr-x 2 root root 4096 Feb 22 08:08 numpy-1.16.1.dist-info
drwxr-xr-x 4 root root 4096 Feb 22 08:00 PIL
drwxr-xr-x 2 root root 4096 Feb 22 08:00 Pillow-5.4.1.dist-info
5. Zip above mentioned “/site-packages” directory and download your zip file.
6. On your local machine, unzip above mentioned zip file. Then make some new folders for each module.
Here is my file structure for your reference:
└ dlib.so (I renamed “dlib.cpython-36m-x86_64-linux-gnu.so” to “dlib.so”)
│ face_recognition (directory and its contents)
└ face_recognition-1.2.3.dist-info (directory and its contents)
│ face_recognition_models (directory and its contents)
└ face_recognition_models-0.3.0.egg-info (directory and its contents)
│ numpy (directory and its contents)
└ numpy-1.16.1.dist-info (directory and its contents)
│ PIL (directory and its contents)
└ Pillow-5.4.1.dist-info (directory and its contents)
7. Zip each “python” directory under above five folders. You will get 5 “python.zip” files. These files are for Lambda layers.
8. Open AWS Lambda console. Create 5 new lambda layers for above zip files. Upload these zip files to each layer correspondingly. Here are some details:
PILPillow Python 3.6
dlib Not specified
face_recognition Python 3.6
face_recognition_models Python 3.6
numpy Python 3.6
Please note, “python.zip” of “face_recognition_models” is too large (approx. 100.6MB) to be uploaded from console webpage, you have to upload it to a S3 bucket first, mark it public, and then upload to Lambda layer from your S3 bucket.
9. Create a new Lambda function, add above 5 layers to your Lambda function, select Python 3.6 as runtime. Then upload your Python code with a test jpeg picture. Here is my “lambda_function.py” for your reference:
def lambda_handler(event, context):
fileName = “1.jpeg”
print(“start face recognition…”)
image = face_recognition.load_image_file(str(fileName))
encoding = face_recognition.face_locations(image)
print(“face recognition finish…”)
Please note, in my case, I use “1.jpeg” for testing only.
10. Increase your Lambda function’s memory to about 512 MB. The default 3s timeout is enough for simple request. In my case, the running status are:
Duration: 174.37 ms
Billed Duration: 200 ms
Memory Size: 512 MB
Max Memory Used: 400 MB
Please note, the memory and timeout should be updated as your functionality needs.
2. I firstly build/install everything on my EC2 with its pre-installed Python 2, but some libraries can’t be invoked by Lambda. Then I switch to Python 3, that’s what we did in step 3 and 4.
3. To build
directories = os.popen(“find /opt/* -type d –
5. Increasing memory allocation is important,
6. Be aware of Lambda total unzipped deployment files (includes your code/dependencies) size limitation, which is 250 MB. If your Lambda function has multiple layers, the total deployment size then is the sum of all referenced layers plus the Lambda itself. Please make sure the total size is within this limitation.
7. In production, the image file should be stored in storage services like S3. The jpeg file I upload to my Lambda function is just for testing.
8. Your functionality may vary from my POC, so you may need to increase the timeout and memory for advance processing.