AWS EC2: T3 vs. M5 Instances (with Benchmark)

Our link checker heavily relies on AWS and EC2 instances in particular. One of the more difficult decisions when dealing with EC2 is choosing the right instance type. Will burstable and (potentially) cheap T3 instances do the job, or should you pay more for general purpose M5 instances? In this blog post, I will try to shed some light on this and provide answers to the following questions:

  • How fast is a T3 instance compared to an equivalently sized M5 instance?
  • How does the T3 CPU credit model work, exactly?
  • What is T3 Unlimited and how reliable is it?

CPU Performance

Without further ado, let me dive directly into some actual testing. I create a t3.large instance (without Unlimited mode) in the us-east-2 region and select Ubuntu 18.04 as the operating system. The first thing I check is the processor the machine is running on:

$ cat /proc/cpuinfo | grep -E 'processor|model name|cpu MHz'

processor       : 0
model name      : Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz
cpu MHz         : 2500.000
processor       : 1
model name      : Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz
cpu MHz         : 2500.000

As you can see from the output, the T3 instance has access to two vCPUs (= hyperthreads) of an Intel Xeon Platinum 8175M processor.

In order to test the performance of the processor, I install sysbench …

$ sudo apt-get install sysbench

… and start the benchmark (which stresses the CPU for ten seconds by calculating prime numbers):

$ sysbench --threads=2 cpu run

sysbench 1.0.11 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 2
Initializing random number generator from current time


Prime numbers limit: 10000

Initializing worker threads...

Threads started!

CPU speed:
    events per second:  1623.24

General statistics:
    total time:                          10.0012s
    total number of events:              16237

Latency (ms):
         min:                                  1.16
         avg:                                  1.23
         max:                                  9.89
         95th percentile:                      1.27
         sum:                              19993.46

Threads fairness:
    events (avg/stddev):           8118.5000/0.50
    execution time (avg/stddev):   9.9967/0.00

The events per second (1623 in this case) is the number you should care about. The higher this number, the higher the performance of the CPU. As points of reference, here are the results from the same test on several other cloud machines as well as some desktop and mobile processors:

Instance Type Processor Threads Burst Baseline
t3.micro Intel Xeon Platinum 8175M 2 1654 175
t3.medium Intel Xeon Platinum 8175M 2 1653 343
t3.large Intel Xeon Platinum 8175M 2 1623 510
m5.large Intel Xeon Platinum 8175M 2 1647 1647
t3a.nano AMD EPYC 7571 2 1510 75
t2.large Intel Xeon E5-2686 v4 2 1830 538
a1.medium AWS Graviton 1 2281 2281
m4.large Intel Xeon E5-2686 v4 2 1393 1393
c4.large Intel Xeon E5-2666 v3 2 1583 1583
GCP f1-micro Intel Xeon (Skylake) 1 956 190
GCP e2-medium Intel Xeon 2 1434 717
GCP n1-standard-2 Intel Xeon (Skylake) 2 1435 1435
Linode 2GB AMD EPYC 7501 1 1236 1170
Hetzner CX11 Intel Xeon (Skylake) 1 970 860
Dedicated Server Intel Xeon E3-1271 v3 8 7530 7530
Desktop Intel Core i9-9900K 16 19274 19274
Laptop Intel Core i7-8565U 8 8372 8372

While bursting, T3 instances provide the same CPU performance as equally sized M5 instances, which answers my first question from above.

Investigating the T3 CPU Credit Model

For the next test, I let the t3.large instance sit unused for 24 hours until it has accrued the maximum number of CPU credits.

t3.large CPU Credit Balance after 24 hours

T3 instances always start with zero credits and earn credits at a rate determined by the size of the instance. A t3.large instance earns 36 CPU credits per hour with a maximum of 864 credits. According to the AWS documentation, one CPU credit is equal to one vCPU running at 100% for one minute. So how long should my t3.large instance be able to burst to 100% CPU utilization?

If the instance has 864 CPU credits, uses 2 credits per minute, and refills its credits at a rate of 0.6 (= 36/60) per minute, it should have the capacity to burst for 864 / (2 - 0.6) = 617 minutes = 10.3 hours. Let me put that to a test.

sysbench --time=0 --threads=2 --report-interval=60 cpu run

Just as expected, the performance drops to about 30% (the baseline performance for t3.large instances) after about ten and a half hours.

t3.large CPU Utilization

T3 Unlimited Mode

T3 instances with activated Unlimited Mode are allowed to burst even if no CPU credits are available. This comes at a price: a T3 instance that continuously bursts at 100% CPU costs approximately 1.5 times the price of an equally sized M5 instance. However, how reliable is Unlimited Mode? My worry is that AWS puts too many instances on a single physical machine, so not enough spare burst capacity is available. To answer this question, I launch a t3.nano instance with Unlimited Mode and let it run at full steam for about four days.

t3.nano CPU Utilization (Unlimited Mode)

As promised, there is no drop in CPU performance. The t3.nano instance delivers the full capacity of 2 vCPUs (almost) all the time. Quite impressive!

Network Performance

Instead of running my own network performance tests, I rely on the results that Andreas Wittig published on the cloudonaut blog. He used iperf3 to determine the baseline and burst network throughput for different EC2 instance types. Here are the values for different T3 instances and an m5.large instance:

Instance Type Burst (Gbit/s) Baseline (Gbit/s)
t3.nano 5.06 0.03
t3.micro 5.09 0.06
t3.small 5.11 0.13
t3.medium 4.98 0.25
t3.large 5.11 0.51
m5.large 10.04 0.74

Although an m5.large instance costs only about 15% more than a t3.large instance, it provides a 50% higher baseline throughput and more than double the burst capacity.

Conclusion and Recommendation

T3 instances are great! Even a t3.nano instance at a monthly on-demand price of less than $4 gives you access to the full power of two hyperthreads on an Intel Xeon processor and at burst runs as fast as a $70 m5.large instance. By activating Unlimited Mode, you can easily insure yourself against running out of CPU credits and being throttled.

If you don’t need the 8 GiB of memory (RAM) that an m5.large instance provides and can live with the lower network throughput, one of the smaller T3 instances with activated Unlimited Mode might be the much more cost-effective choice. In the end, it depends on how high the average CPU usage of your instance is. The table below lists the CPU utilizations up to which bursting T3 instances remain cheaper than an m5.large instance. Please note that the calculations are based on the on-demand prices and might be different when using reserved instances.

Instance Type Memory Cost-effective if average CPU usage is less than
t3.nano 0.5 GiB 95.8%
t3.micro 1 GiB 95.6%
t3.small 2 GiB 95.2%
t3.medium 4 GiB 74.4%
t3.large 8 GiB 42.5%

Last but not least, let me state that you shouldn’t rely solely on the benchmarks and comparisons in this post. You are welcome to use my finding as a first guide when choosing the right instance type, but don’t forget to run your own tests and make your own calculations.


How to Find Flash Content on a Website

The days of Flash are numbered. Chrome blocks Flash content by default since version 76 (released in July 2019), Firefox since version 69 (September 2019), and Google has announced it will stop indexing SWF files by the end of 2019. The final nail in the coffin will come later in 2020, when Adobe officially end-of-lifes the Flash Player.

According to W3Techs, only about 3 percent of all websites utilize Flash nowadays. This doesn’t sound like very much, but considering the huge number of websites on the web, it still means that millions of sites rely on Flash. If your website is among them, it’s definitely time to act!

The first step in migrating a website away from Flash is identifying the pages that need to be updated. In this post I will demonstrate how to use Dr. Link Check to crawl a website in order to find all SWF files and the pages linking to them.

Step 1: Start a link check

Click here to navigate to the Dr. Link Check home page, enter the address of your website into the input field, and click the Start Check button.

Start new link check

Step 2: Show all links

The service immediately starts crawling through your website, which may take a while. Once the crawl is complete, click the All Links item in the sidebar on the left.

All links

Step 3: Filter links by file extension .swf

Click the Add button in the Filter bar, select URL from the drop-down menu, enter .swf into the input field, and press Enter to confirm the input.

Add filter

The list will now only display links containing “.swf” in their URL. In order to see the pages linking to the Flash file, hover over the link and click the Details button.

Conclusion

Although Dr. Link Check is primarily a broken link checker, its flexibility also makes it an excellent tool for finding specific files on a website. You can not only use it to search for Flash, but also for Java, Silverlight, or any other type of content that’s identifiable by filename. Give it a try!


The Four Pillars of Modern SEO

Social media may have been the star of the digital marketing scene in recent times, but reports of SEO’s demise are greatly exaggerated. Effective search engine optimization can still drive plenty of profitable traffic to a website, but it needs to be approached correctly. What do you need to consider when developing a successful modern strategy?

4 Pillars of SEO

1. Link Building

Links have always been the foundation of the Google algorithm. Although the specifics have changed over time, a good link will always be a positive for your ranking. But what should you be looking for when building links?

  • Always strive for links that appear natural and freely given by the source site. If your link looks as if it’s been bought or traded, it will be downgraded in value, or may even work against you in extreme cases.
  • Links from highly trafficked sites are always preferable. Not only will these links tend to have a greater algorithmic impact, they’ll also send you traffic directly.
  • Your incoming links should have a wide range of anchor texts (the text used within the link itself). Aim for a mixture of descriptive keywords and phrases: generic terms such as “click here,” website addresses, and longer sentences which describe your page content. Also, mix it up by including links based around images rather than text.
  • Your link profile should be as diverse as possible. Aim for links from sites with high and low traffic, closely related themes and more general blogs, and from a variety of countries and organizations. The greater the mix, the more natural your profile will look, and the greater the impact your best links will have.

2. On-Page Optimization

On-page optimization is all about getting your page’s ducks in a row to leave the algorithm with no doubt about its theme. Although it’s nowhere near the make-or-break factor it once was, it’s still vital.

  • Optimize your page title and meta description so that they are readable, interesting, and accurately descriptive, and also include your target keywords.
  • Structure your content well, using headings and subheadings where appropriate, again including relevant keywords where they naturally fit.
  • Use image alt attributes to describe the image content accurately, while also adding the keywords for which you want to rank.
  • Tweak your content to naturally include your main terms and a good spread of related phrases. However, never do this at the expense of readability.
  • Add links to other pages on your website and to respected external sites, where they fit the text and add value for the user.

3. Technical SEO

Technical SEO is in some ways the ugly duckling of the optimization family. It’s rarely exciting, but it provides the foundation on which other parts of the discipline can build. In essence, it means ensuring your site is easily understood by the search engine algorithm, with no technical glitches or confusion to trip the spiders up. Here are the most important things to consider.

  • Identify and remove as much content as possible that’s duplicated across multiple pages.
  • Ensure that several different URLs can’t reach the same page. That is an easy trap to fall into when the URL is parsed by your website software to build a page from database entries.
  • Check that your pages load quickly and reliably with no server errors, even under heavy load. Google’s PageSpeed Insights tool and PingDom’s Website Speed Test can help you identify performance bottlenecks and improvement opportunities.
  • Ensure that your site works appropriately across a wide range of devices. Google now actively demotes sites which offer a poor mobile experience.
  • Use a broken link checker to track down and fix broken internal and external links. Broken links send poor quality signals to search engines, implying sloppy maintenance and a frustrating user experience.

4. Quality Content

The final part of the optimization jigsaw is high-quality content. Search engines strive to direct users to genuinely useful sites, and if your content is poor, you won’t fit this criterion. Not only that, but high-quality content increases user engagement, and this feeds directly back into the ranking system via tracking through Google’s advertising, analytics, and social media platforms.

  • Make your content as unique as possible. While it’s difficult to create something original in a crowded market, publishing rehashed material isn’t going to make you rise above your competitors.
  • Make your content genuinely useful, entertaining, or otherwise attractive. Think beyond the search engine spiders – user satisfaction should be your prime aim. Content that doesn’t engage or convert in some way is worth virtually nothing.
  • Make sure your content is easily readable and clearly presented. Your visitors immediately need to recognize that they’ve found what they’re seeking. If they don’t, they’ll hit the back button, inflating your bounce rates. Google will notice this and lower your rank accordingly.

Modern online marketing offers countless ways of driving traffic to a website, but search engine optimization remains one of the most powerful and cost-effective. But it’s not something you can take for granted or approach half-heartedly. Pay attention to these four pillars, and you’ll be giving your SEO efforts an essential underpinning for profitable success.


New Notification Features

We have added two new options to the Project Settings dialog that allow you to configure where and when to send notification emails about completed link checks. Previously, the results of a recurring check were always sent to the account holder’s email address. Now you can use different recipients for different projects.

Email notification settings

In order to have the emails delivered to multiple recipients, simply enter the addresses separated by a comma. The first email address will be used for the email’s To field, the remaining addresses go to the Cc field.

If you are not interested in getting emails about checks that didn’t identify any dead or bad links, tick the checkbox next to Only send if issues were found. Dr. Link Check will then keep quiet until something worth reporting is found.


How to Find All Non-HTTPS Links on a Website

The majority of the top 1 million websites now use HTTPS, and for good reason:

  • HTTPS keeps a user’s connection to a server safe from eavesdropping and tampering.
  • Non-HTTPS websites cause modern browsers to display an off-putting “Not Secure” warning.
  • Google Search favors HTTPS pages over unsecure HTTP ones.

Considering that SSL/TLS certificates don’t even cost you anything (see Let’s Encrypt for free certificates), there’s every reason to use HTTPS for a website.

When migrating an existing site from HTTP to HTTPS, it’s often difficult to find all the old http:// URLs that need to be changed to https://. Missing just a single link to an image, script, or other resource, results in a Mixed Content warning in the browser. In this post, I will demonstrate how to use Dr. Link Check for finding all http:// URLs on a website.

Step 1: Start a check at https://www.drlinkcheck.com/

Go to the Dr. Link Check homepage, enter the address of your website, and click on Start Check.

Start Link Check

Step 2: Click on “http:” under “Link Schemes”

After the check is complete, Dr. Link Check will display the number of http:// links under Link Schemes on the Overview page. Click the http: item to open a report listing the found links.

HTTP Links: Overview

Step 3: Add “Direction: Internal” filter item

The report lists all http:// links found on your website, including those to other websites. If you want to restrict the report to internal links, click the Add button in the Filter bar at the top of the report, and select Direction from the menu. Now, click on Outbound and select the Internal menu item.

HTTP Links: Report

As presented above, Dr. Link Check is not only a great broken link checker, but it can also help you migrate your website from HTTP to HTTPS. Give it a try!


Pricing Changes

We have updated the pricing model for Dr. Link Check to make it more transparent and easier to understand. You now buy the capacity to check websites of a certain size (based on the number of links it contains), and you can then check those websites as often as you want during the term of your subscription.

Previously, we were giving you a monthly quota of links that could be checked, and it was up to you how to use this quota. With a quota of 20,000 links, you could run one single check on a website with 20,000 links, or 20 checks for a smaller site with only 1,000 links. Once the quota was used up, you either had to upgrade your subscription or wait for it to reset with the next subscription cycle.

We now know that this is not what users want. If you are working on fixing your website’s links, you don't want to wait until next month to run another check to verify your changes. From now on, you can run as many checks on a website as you need, without having to worry about running out of your quota.

The prices of existing subscriptions won’t change, but you will benefit from the removal of the quota system. If your subscription previously included a quota of 20,000 links per month, it now has a limit of 20,000 links per website project.

In case you have any questions regarding the changes, please get in touch.


Older Posts Newer Posts