Kerry Baker leads Ookla's research and content efforts in North America. He has over 20 years experience in the telecom industry, primarily at T-Mobile studying network performance benchmarking and customer experience. Kerry also has founder’s experience with four technology-related startups. Kerry holds masters degrees from the University of Washington in Business and International Political Economics.
Replay the 2024 season and playoffs according to Ookla Speedtest Intelligence® data
Big league baseball is back! And like baseball, Ookla® Speedtest has lots of stats. So we’re taking the opportunity to replay 2024 through the lens of our Speedtest Connectivity reports for the United States and Canada from the second half of 2024. We use the mobile results for the baseball teams’ respective cities, as presented in these reports.
Here’s the framework:
Speedtest Season standings are based on our Consistency Score, because, over the course of 162 games in the regular season, consistency is certainly the key to success.
Playoffs Wild Card games are where teams can suddenly catch lightning in a bottle, so we determine the winners based on multi-server Latency.
Playoffs Quarterfinals winner is determined by median Upload Speed, because we have to save Download Speed for the next one…
Playoff Semifinals is won by the main workhorse network metric of median Download Speed
Speedtest Finals reaches outside of the Connectivity Report to identify the all-star Top 10% Download Speeds in each city. (Note: this metric is not in the Connectivity Report, but comes from the same Speedtest Intelligence data source.)
Speedtest Season – Consistency
As noted in our framework, Consistency is key to winning the regular season. Boston, Cleveland, Arlington (Texas), Philadelphia, Pittsburgh, and San Francisco are moving on.
But before starting our playoffs, let’s indulge in a little Inside Speedtest. It’s interesting to note how the Consistency metric for cities in the East tended to outperform cities in the West, particularly over California cities. Consistency is a minimum performance floor, measuring the percentage of Speedtest samples meeting or exceeding a threshold of 5 Mbps download and 1 Mbps upload throughput.
One reason to explain this is that older cities in the East developed to have more dense populations, and therefore the cellular networks are built accordingly. As a result, the opportunity for a given smartphone user to be closer to a cell site and better signal strength is more likely.
To highlight this with examples of opposites, vertiginous New York City has a population density over twenty-nine thousand (29,303) people per square mile. The sprawling city of Los Angeles, on the other hand, has over eight thousand (8,304) people per square mile (per 2020 US Census). Checking Speedtest Intelligence® mobile network quality data – not included in our Speedtest Baseball – the signal quality (RSRP) in the second half of 2024 for New York City was -93 dBm and for Los Angeles was -98 dBm. Like a pitcher’s ERA, lower is better for RSRP. A difference of 5 dBM may not sound like much, but it can be what is needed to have a cell signal indoors or be able to stream the baseball game. Likewise, Consistency for New York City was 90.6% and Los Angeles was 87.1%. A difference of 3.5% points may not sound like much, unless you are one of the 3.5% with balky network data speeds.
Advancing to the Speedtest Playoffs
Boston, Cleveland, Philadelphia, and Pittsburgh earn the bye, setting up the Wild Card matches for the rest of the cities.
Speedtest Playoff Wild Cards – Latency
For Speedtest Playoffs Round 1, the metric is Latency and, like RSRP and ERA, lower is better. New York edges out Arlington (Texas), Toronto sweeps Baltimore, and San Francisco squeaks past Washington. New York (again, but different) and Atlanta tied, so using Consistency for the tie breaker, New York (90.6%) is just able to move on past Atlanta (89.9%).
Speedtest Quarterfinals – Upload Speed
In actual big league baseball, the top regular season teams that earned the bye past the first round Wild Card often have found the time off not to their liking, with the break in their baseball routine blunting their competitive edge. Speedtest Baseball is similar, with Cleveland and Philadelphia both exiting versus Toronto and New York, respectively, based on Upload Speed. Upload Speed is great for posting your social media from the ballpark. Boston moves on over New York in that great rivalry, and Pittsburgh tops San Francisco.
Speedtest Semifinals – Download Speed
For measuring network performance, Download Speed is the champion. It’s not just about the individual’s data speed they see when running a Speedtest, it also indicates the network capacity available for everyone. Most people individually don’t need hundreds of megabits per second on their smartphone, but having that much speed available helps ensure a good network experience for all. Boston and Pittsburgh bring out their brooms to sweep their way into the Speedtest Finals.
Speedtest Finals – Fastest Download Speeds (top 10% of results)
Continuing from the Speedtest Semifinals, download speed is still on the field, but this time it’s just the all-star speeds – the top 10% – used to decide the winner. And with this, we have a change up in cities from the Speedtest Semifinals, where Pittsburgh was faster than Boston. Instead here in the Finals, Boston’s top speeds top Pittsburgh’s top speeds.
We didn’t see that coming, but an intrigue of baseball is that one can watch hundreds or even thousands of games, and still see something they had never seen before.
Congratulations to Boston, the victor of our Speedtest Baseball Finals.
Now, let’s Play Ball!
Ookla retains ownership of this article including all of the intellectual property rights, data, content graphs and analysis. This article may not be quoted, reproduced, distributed or published for any commercial purpose without prior consent. Members of the press and others using the findings in this article for non-commercial purposes are welcome to publicly share and link to report information with attribution to Ookla.
Originally published in Spanish on DPL News, February 20, 2025, under the title “La banda ancha en Latinoamérica: el camino ascendente”
To truly address the digital divide, meaningful connectivity starts with robust, accessible broadband internet for all
Telecommunications infrastructure in Latin American countries (LatAm) faces common challenges across the region. These include geographic barriers and economic disparities that hinder deployment. Further, these barriers directly contribute to the so-called “digital divide,” typically depicted as an urban-versus-rural condition. Not unique to LatAm, most countries and governments around the world are grappling with this same problem.
Ensuring broadband access alone is not adequate. Consumers are underserved by slow, unreliable internet connections. Many countries in LatAm are making steady progress toward providing high-speed broadband connectivity at scale.
The ‘Fiber Development Index Analysis 2024’ from the World Broadband Association (WBBA) highlights broadband connectivity’s foundational role in socioeconomic development. Broadband networks drive economic growth by supporting information and communications technology across all sectors, contributing significantly to GDP and employment. Furthermore, broadband enables the digitization of various industries, such as manufacturing, healthcare, and education, leading to increased productivity, efficiency, and business success. The World Bank cites research that finds when fast internet becomes available, employment increases by 13 percent and businesses nearly quadruple their exports.
The social benefits of broadband are equally important. In education, it provides access to better learning materials and facilitates communication between students and teachers, especially during remote learning situations. Other social benefits include remote work, telehealth services, and online banking for underserved populations. These benefits, while hard to quantify, are crucial for societal well-being.
Because of the numerous benefits it provides – including access to information, economic opportunities, education, healthcare, and entertainment – meaningful connectivity, characterized by high-quality and reliable broadband access, is increasingly recognized as a fundamental human right.
In LatAm and Europe, and in many regions across the globe, some countries are rapidly deploying fiber and 5G and emerging as digital leaders, while others are lagging behind; this disparity exists in both regions. For example, Europe has embedded universal gigabit broadband coverage by 2030 as a core pillar of its broader emerging pro-growth industrial strategy to boost the region’s competitiveness and differentiate it through world-class infrastructure. A common approach in LatAm and Europe, is the neutral network as a way to expand broadband coverage and foster competition. For example, in Mexico, Red Compartida is a nationwide wholesale-only wireless network. And in Spain, development of neutral networks through companies like Lyntia and MásMóvil, operate wholesale-only fiber networks used by multiple operators.
This article examines the high-level status of fixed and mobile connectivity of a selection of LatAm countries tending to have large populations, as well as Spain (Spain being the host of the Digital Summit LatAm 2025 in Madrid at the end of February).
Data sources are: (1) OoklaⓇ mobile SpeedtestⓇ data sets are queried, including the Speedtest Global Index of 152 countries, and Speedtest IntelligenceⓇ with focus on 5G insights. And, (2) the Fiber Development Index (FDI) – “a collaboration between Ookla, the WBBA, and Omdia – benchmarks fiber development and performance across 93 countries, the FDI provides crucial insights into actual broadband performance and availability worldwide” for industry stakeholders, policymakers and regulators, service providers and suppliers, supporting the development and growth of the fiber industries. The full FDI list of 93 countries is seen here.
Fixed Connectivity
The Speedtest Global Index current rankings (January 2025) of countries based on median download speeds for fixed connectivity presents many LatAm countries among a mix of countries from other regions for comparison. Three categories are evident. More advanced countries where fiber performance is already strong, including in Chile, seeing well over 200 Megabits per second (Mbps) median download speed. The second group is solidly above 100 Mbps median download speeds where its fiber is deployed. The last group is generally below 100 Mbps, slower to take off and not yet delivering the performance possible with fiber.
Speed is simple to understand, but it only truly matters when one has access to a network that delivers the speed. High speeds on a network with limited availability are probably not as beneficial as medium speeds on a widely available network. To judge the breadth of fiber deployment, the FDI provides a measurement to assist in evaluating this called Fiber-to-the-Home (FTTH).
Up-Right Trends
In the chart of Fiber Deployment and Speed, for comparison over five years time (2020 – 2024) and among a selected cohort of larger population countries (not all of LatAm, to keep the visual intelligible), FTTH penetration percentage is on the horizontal axis and median download speed of the fiber connection in Mbps is on the vertical axis. FTTH penetration is the number of FTTH subscriptions divided by the total number of households.
FTTH penetration is often higher in relatively smaller-geography countries because population density tends to be higher as well, resulting in favorable economic conditions for deployment and adoption. This does not describe the morphology of countries in this analysis. Year-to-year consistent progress is clearly evident for all countries (FTTH moving to the right), with Brazil, Chile and Mexico moving the farthest. However, none of the six large LatAm countries approached 50% FTTH penetration, let alone catching Spain.
Download speed showed steady progress as well, though at different rates of improvement. The same three categories as identified early are evident: Chile and Spain in the upper; Brazil, Peru and Colombia in the middle; Argentina and Mexico in the lower. Chile’s rise in performance is particularly commendable, from 55.42 Mbps in 2020 to 263.89 in 2024. On the other hand, Mexico at 76.85 Mbps in 2024, was still not where Spain was at 84.80 Mbps in 2020. In sum, among this cohort, speed performance presents a mix of successes and opportunities.
Excluding Spain, only Chile and Brazil improved on both FTTH and speed axes. Mexico expanded FTTH access, but its speed performance lagged. In the other direction, Peru and Colombia presented limited FTTH deployment, but it’s fast where available. Argentina trailed on both metrics.
It is important to consider that the speeds cited here are median speeds. The median is the middle, which means that half of the speeds were slower (and, of course, half were faster). This matters especially for regulators that are mandating minimum performance targets, often in the context of closing the digital divide. For example, in the European Union’s (EU) goals from its Gigabit Society and Digital Decade policy:
Broadband Europe promotes the Commission’s strategy on Connectivity for a European Gigabit Society by 2025 as well as the vision set by the Digital Decade for Europe’s digital transformation by 2030 to connect European citizens and businesses with very high-capacity networks, which will enable innovative products, services and applications to all citizens and business across the EU.
This Gigabit Society vision for 2025 relies on three main strategic objectives:
Gigabit connectivity for all of the main socio-economic drivers;
uninterrupted 5G coverage for all urban areas and major terrestrial transport paths;
access to connectivity offering at least 100 Mbps for all European households
The last two suggest a future where there is fixed service delivering 1,000 Mbps – that is to say, a Gigabit per second (Gbps). Incidentally, Brazil’s regulator Anatel also has a Gigabit target for fixed connectivity speed. And, for mobile networks in 2030, on the cusp of 6G, 5G will be pervasive in the same way that 4G is today. To that point, the next section examines 4G and 5G in these same cohort of countries.
Mobile Connectivity
Mobile is the primary internet access method for many people across LatAm, and sometimes their only internet access method. 4G can provide minimally adequate performance, with the benefit of being a mature and widely deployed technology. 4G networks and devices are abundant.
However, to fulfill the future promise, and enjoy the services and benefits that come with high-speed internet access, 5G is capable. 5G’s superior performance to that of 4G is competitive with fixed broadband, and can sometimes reach consumers where fixed networks might not yet have. Further, some consumers are making a financial decision – only affording one method of internet access. In this case, mobile usually wins this choice over fixed. (Mobile devices also have the benefit of being able to generate a Wi-Fi hotspot, whereas a fixed connection cannot become a mobile phone.)
This section will look at two “availability” metrics, analogous to FTTH penetration, to illustrate the accessibility of 5G. Then, just as with fixed performance above, compare the speed performance of 5G to that of 4G.
What the phone sees
Mobile Technology Generation, 2H 2024
Availability on Service-Active Devices shows the mobile network technology seen by all devices in a given county. This isn’t coverage or the amount of traffic (payload) on a given technology generation. Rather, this is the devices’ perspective of network technology generations (the “G”) available to connect with. In other words, this is a democratic view of available wireless technologies – every device is voting. Note that a 4G device will never see a 5G network, while a 5G device will see all the network technology generations. Just the same for a 3G or 2G device – devices are not “forward compatible” in mobile technology generations.
We can easily draw insights from the Availability on Service-Active Devices chart about the relative “G” availability for each country. Spain is further along in 5G deployment, Chile and Brazil are early in 5G, Mexico is just starting, and Colombia and Argentina are pre-deployment (perhaps in trials).
Less obvious is the velocity of change from one technology generation to the next. The installed base of 4G devices is large and slow to upgrade. Device quality is better and device costs are higher, leading consumers to use older technology for longer. In Spain, all of its operators had launched 5G networks in 2020, and thanks to proactive initiatives under the Universalization of Digital Infrastructures for Cohesion Program (UNICO), 5G coverage has expanded. Rural 5G coverage, as it is in most countries, is still a work-in-progress for UNICO. But also, five years into 5G, Spain still has many non-5G devices.
Thus, a key point this view makes, as obvious as it may be, is that the device is critical to the experience. (This point is also true of fixed networks. For example, an old Wi-Fi router would be a performance bottleneck on a Gigabit fiber connection.) In each new device, modern chipsets and advanced technologies, and spectrum bands and radios, combine to fulfill the potential capabilities of the network.
What the 5G phone sees
The 5G Availability time-series chart illustrates, for 5G devices, the percentage of those 5G devices that spend the majority of their time connected to a 5G network. Highly correlated with Availability on Service-Active Devices – Spain with the most 5G and Argentina with the least 5G – this view indicates the intersection of 5G network deployment and 5G devices.
Again, this is not a view of coverage, though it does suggest the degree of alignment between (network) deployment and (device) distribution. Rhetorically, what is the point of a 5G device without a 5G network?
However, let’s touch on coverage. In lay terms, coverage is usually thought of as geography, and this idea is reinforced with network maps. Networks, actually, and especially new ones, are deployed where the most people live. Scale economies dictate that network coverage is primarily about population, not geography.
For example, Anatel recently announced that Brazil had surpassed its 2027 coverage goal for 5G at 57.67% of the population – it now covers 62.98% of the population. With impressive precision, Anatel’s statement indicates that coverage is about people.
The Anatel news also contained the information that 5G subscriptions had doubled in 2024. Comparing this with data in the 5G Availability time series chart, from Q3-Q4 2023 to Q3-Q4 2024, Brazil 5G Availability went from 16.77% to 31.99% – it nearly doubled.
We have now touched on network availability, devices, and coverage. It’s time to turn to what 5G’s headline feature has been since before 3GPP release 15 in the prior decade. That is, of course, speed.
5G Speed – green means go fast
Based on Speedtest user data in each country in the second half of 2024, there is no question that 5G speeds are faster than 4G. And as we just learned with Brazil’s expanded 5G coverage and growth in devices and subscriptions, millions of Brazilian mobile users are enjoying mobile speed performance more than ten-times faster than their 4G friends. User experience will vary within each country based on the mobile service provider, among other factors.
Speed isn’t just about how fast an individual user gets to experience browsing or downloading. Speed is representative of network capacity. Moreover, speed is the result of the capacity (spectrum amount, technology, site count/density, backhaul) and user demand (number of users and, again, browsing or downloading or streaming…). So, when we look at Argentina’s 5G speed, we know from information above in this report that it is a very new deployment with very few users (Called “unloaded” in network jargon). Consequently, the speeds of newer networks must be considered in this context, to not over-estimate their potential and promise.
Spain offers a sensible counterbalance, though even a three-times faster experience should be encouragement for a Spaniard to ditch their 4G phone.
Don’t wait
Let’s consider meaningful connectivity and the digital divide. One could argue that, according to the findings in this analysis, fixed FTTH connectivity is under-penetrated and 5G mobile service is nascent to the extent that neither delivers meaningful connectivity. In this case, the digital divide, based on some aspirational access and performance objective, could mean most people. However, other countries’ experiences are likely to repeat. As penetration and adoption matures, attended by performance, to a state of meaningful connectivity, there will be those at risk of being left out. This is the real digital divide.
The opportunity is this – don’t wait. Programs and initiatives can pre-emptively address the systemic gaps that could otherwise result. Misses and successes of others can be instructive. There are billions in government funds aimed at this issue.
It is not the purpose of this article to survey each country’s digital divide initiatives, but to identify a few references for further exploration:
An example of pre-emptive success outside of LatAm, the state broadband office in South Carolina moved on the digital divide issue early and fast, resulting in a higher rural broadband adoption than in urban areas. An inverted divide! (How South Carolina Teamed with Ookla to Become a Rural Broadband Leader)
Brazilian regulator Anatel has produced its 2023 – 2027 Strategic Plan, wherein it lays out objectives and framework. (Other programs in Brazil can be found at this thorough list Hacking the Gap – Brazil’s Path to Digital Inclusion)
For more information about Ookla and Speedtest Intelligence data and insights, please get in touch.
Ookla retains ownership of this article including all of the intellectual property rights, data, content graphs and analysis. This article may not be quoted, reproduced, distributed or published for any commercial purpose without prior consent. Members of the press and others using the findings in this article for non-commercial purposes are welcome to publicly share and link to report information with attribution to Ookla.
Here’s what Chile’s 12-hour nationwide power outage looked like according to mobile Speedtest data
A major transmission line failure caused a massive power outage across most of Chile. The outage began in the afternoon of February 25th and lasted approximately 12 hours, into the early morning of February 26th. Critical facilities like hospitals had to rely on backup power generators and the outage was disruptive enough that a state of emergency was declared for the country.
How did mobile networks perform during the outage? This article looks at mobile networks’ performance through SpeedtestⓇ metrics of sample counts (i.e., the number of completed Speedtests) and download speeds.
Ookla examined the hourly Speedtest sample counts of the main four mobile carriers in Chile over a 72-hour period. Prior to noon on February 25, 2025, the networks performed in a normal state with fewer samples in the night and more in the day.
The outage reportedly started at 3:16 p.m., which matches the spikes in the counts of Speedtest users checking on the network. Comparing the 12-hour outage to the same time the day before, the rate of testing was over 2.5x. This behavior is typical of, and similar to, mobile network outages seen around the world. Staying connected in an emergency is critically important, so checking the status of the network makes good sense.
By the morning of February 26, 2025, when the outage is over and power restored, there is more testing than the prior two mornings. This is likely due to many people checking their phones upon waking up, so it isn’t unusual to see more aggressive testing than usual.
How did the mobile networks perform?
When mobile networks are “unloaded” — that is, not carrying much traffic — as they usually are at night, Speedtest users clock faster speeds. These are the peaks in the chart. The valleys are the times when most people are active and more loading occurs, resulting in slower speeds.
Selecting the 12 hours of the power outage (the labeled box) and the same 12 hours the prior day (the unlabeled box), the median download speed for the group of operators was effectively cut in half — 16.42 Mbps during the power outage compared to 33.68 Mbps before.
Taking a closer look at all the mobile service providers together, from the outage starting just after 3 p.m., median download speed initially held around 12+ Mbps, then dipped to its low point during the 6 p.m. hour at 7.49 Mbps. Thereafter and until recovery started around 11 p.m., download speeds remained in a narrow range around 9-10 Mbps through the 10 p.m hour. While these median download speeds wouldn’t be considered fast, they do suggest that the mobile networks were serviceable and customers were able to stay in touch.
However, noting that median means the middle, half of the Speedtest users were experiencing slower speeds than seen in the line chart. (And obviously, half were faster, but they wouldn’t have as poor an experience.) To get a sense of the bottom, the slowest 10th percentile for the day of the outage (February 25) compares unfavorably to the same three prior Tuesdays in February. On the day of the outage, the 10th percentile speed was below 1 Mbps – effectively unusable apart from a text message and hopefully holding a voice call. The prior three Tuesdays recorded 2.60 Mbps – still not great, but in older wireless network generations this was fast enough for watching a standard definition (SD) video.
Another observation from the closer look at All Providers Download Speed is the speed declined from early in the outage (3 p.m. to 4 p.m.) to the middle of the outage (5 p.m. to 10 p.m.). One explanation could be the network loading mentioned before, but this seems unlikely since the normal peak loading pattern would have been entirely disrupted by the power outage. A second explanation may be that backup battery power for the cell sites on mobile networks ran out. That is to say, if, for example, every site had a couple hours of battery backup, and some sites also had generators that came online after the batteries were used up, a pattern like this could be seen.
The Chilean telecoms regulator, SUBTEL, was cited in tech press to be looking into increasing the purported four-hour back up to six hours. Network hardening, whether against a power outage or other disruptions like natural disasters, is critical for network resilience.
Ookla data can offer insights into network performance, reliability, and resiliency. To find out more about Speedtest Intelligence® data and insights, please contact us here.
Cómo afectó un apagón eléctrico a la resiliencia de la red móvil de Chile
Así fue el apagón nacional de 12 horas en Chile según los datos de Speedtest móvil
Una importante avería en una línea de transmisión provocó un apagón masivo en la mayor parte de Chile. El corte comenzó en la tarde del 25 de febrero y duró aproximadamente 12 horas, hasta la madrugada del 26 de febrero. Las instalaciones críticas, como los hospitales, tuvieron que recurrir a generadores de energía de reserva y el apagón fue tan disruptivo que se declaró el estado de emergencia en el país.
¿Cómo funcionaron las redes móviles durante el apagón? Este artículo examina el rendimiento de las redes móviles a través de las métricas de recuento de muestras de SpeedtestⓇ (es decir, el número de Speedtests completados) y velocidades de descarga.
Ookla analizó los recuentos horarios de muestras Speedtest de los cuatro principales operadores de telefonía móvil de Chile durante un período de 72 horas. Antes del mediodía del 25 de febrero de 2025, las redes funcionaban de manera normal, con menos muestras durante la noche y más durante el día.
Al parecer, la interrupción comenzó a las 15.16 horas, lo que coincide con los picos en los recuentos de usuarios de Speedtest que comprueban la red. Si se compara la interrupción de 12 horas con la misma hora del día anterior, la tasa de comprobación se multiplicó por más de 2,5. Este comportamiento es típico y similar al de las interrupciones de la red móvil en todo el mundo. Mantenerse conectado en caso de emergencia es de vital importancia, por lo que comprobar el estado de la red tiene mucho sentido.
En la mañana del 26 de febrero de 2025, cuando terminó el apagón y se restableció el suministro eléctrico, se registraron más test que las dos mañanas anteriores. Esto se debe probablemente a que muchas personas comprueban sus teléfonos al despertarse, por lo que no es raro ver más pruebas de lo habitual.
¿Cómo funcionaron las redes móviles?
Cuando las redes móviles están “descargadas”, es decir, no tienen mucho tráfico, como suele ocurrir por la noche, los usuarios de Speedtest registran velocidades más rápidas. Esto es lo que representan son los picos en el gráfico. Los valles son los momentos en los que la mayoría de las personas están activas y se produce una mayor carga en la red, lo que da como resultado velocidades más lentas.
Seleccionando las 12 horas del apagón (el recuadro con la línea más gruesa) y las mismas 12 horas del día anterior (el recuadro con la línea más fina), la velocidad media de descarga del grupo de operadores se redujo efectivamente a la mitad: 16,42 Mbps durante el apagón frente a 33,68 Mbps antes.
Si analizamos todos los proveedores de servicios móviles en su conjunto, desde el apagón que comenzó justo después de las 15.00 horas, la velocidad mediana de descarga se mantuvo inicialmente en torno a los 12 Mbps o más, para descender a su punto más bajo a las 18.00 horas, con 7,49 Mbps. A partir de entonces, y hasta que comenzó la recuperación hacia las 23.00 horas, las velocidades de descarga se mantuvieron en un estrecho margen de entre 9 y 10 Mbps hasta las 22.00 horas. Aunque estas velocidades medianas de descarga no se pueden considerar rápidas, sugieren que las redes móviles estaban operativas y los clientes pudieron mantenerse en contacto.
Sin embargo, teniendo en cuenta que la mediana significa el medio, la mitad de los usuarios de Speedtest experimentaron velocidades más lentas que las que se ven en el gráfico lineal (y obviamente, la mitad eran más rápidos, pero no tendrían una experiencia tan mala). Para hacerse una idea de la parte inferior, el percentil 10 más lento para el día de la interrupción (25 de febrero) se compara desfavorablemente con los mismos tres martes anteriores de febrero. El día del apagón, la velocidad del percentil 10 estaba por debajo de 1 Mbps, prácticamente inutilizable salvo para enviar un mensaje de texto y, con suerte, mantener una llamada de voz. Los tres martes anteriores se registraron 2,60 Mbps, lo que sigue sin ser bueno, pero en las antiguas generaciones de redes inalámbricas era lo bastante rápido para ver un vídeo de definición estándar (SD).
Otra observación que se desprende del análisis de la velocidad de descarga de todos los proveedores es que la velocidad disminuyó desde el principio del apagón (de 15.00 a 16.00 horas) hasta la mitad del mismo (de 17.00 a 22.00 horas). Una explicación podría ser la carga de la red mencionada antes, pero parece poco probable, ya que el patrón normal de picos de carga se habría visto totalmente interrumpido por el apagón. Una segunda explicación podría ser que se agotaran las baterías de reserva de los emplazamientos de las redes móviles. Es decir, si, por ejemplo, cada emplazamiento tuviera un par de horas de batería de reserva, y algunos emplazamientos también tuvieran generadores que entraran en funcionamiento una vez agotadas las baterías, podría observarse un patrón como éste.
La SUBTEL, organismo regulador de las telecomunicaciones en Chile, ha sido citada en la prensa tecnológica por estudiar la posibilidad de aumentar a seis horas la supuesta reserva de cuatro horas. El refuerzo de la red, ya sea contra cortes de electricidad u otras interrupciones como catástrofes naturales, es fundamental para la resistencia de la red.
Los datos de Ookla pueden ofrecer información sobre el rendimiento, la fiabilidad y la resistencia de la red. Para obtener más información sobre los datos y perspectivas de Speedtest Intelligence, póngase en contacto con nosotros ®aquí.
Ookla retains ownership of this article including all of the intellectual property rights, data, content graphs and analysis. This article may not be quoted, reproduced, distributed or published for any commercial purpose without prior consent. Members of the press and others using the findings in this article for non-commercial purposes are welcome to publicly share and link to report information with attribution to Ookla.
More than a million Canadians could be left out in the cold
In the head spinning geopolitical trade news, among many items was the canceling and uncanceling of Starlink in Ontario, Canada. In response to President Trump’s proposed tariffs on Canada, Ontario Premier Doug Ford stated on social media platform X that he planned to cancel the province’s contract with Starlink, which is owned by Elon Musk, who is working closely with Trump on a number of initiatives. However, within 24 hours of making that statement, Trump had delayed his planned tariffs and Ford said he would pause his retaliatory measures.
In Ontario and other provinces with relatively denser urbanization in their south, consumers have many options for broadband internet service. But in rural and remote areas of provinces and in much of The Territories, cancelling Starlink could result in the loss of internet connectivity entirely. Looking at Speedtest Intelligence data of Starlink users in Canada, we can estimate what might be at stake. We compared Starlink Speedtest user tests to the total number of fixed Speedtest user tests to estimate adoption trends among each Province and Territory from 2020 through 2024.
New Brunswick, Manitoba, Alberta and Ontario were the first to see traces of Starlink Speedtest users in 2020. Soon after in the first half of 2021, British Columbia, Nova Scotia, Quebec, and Saskatchewan appeared, with Prince Edward Island and Newfoundland and Labrador lighting up in the second half of 2021 as well. Then the rest of Canada came in the back half of 2022 with the territories – Northwest Territories, Nunavut and Yukon – leaping into the picture.
Take off to the Great White North
A pattern emerges among the provinces and territories with the service launch of an adoption followed by a relative stabilization in the share of Starlink Speedtest user samples of the total fixed internet Speedtest samples.
Quite obviously the rates of adoption differ. As one would intuitively expect, the more-rural and less-population-dense areas see the steepest adoption curves. Nunavut in particular (population 37 thousand, whom all could fit inside the Rogers Center – home of the Toronto Blue Jays – yet similar in area to Mexico) rockets to the top of the chart and settles into the low-to-mid-40%s of Starlink Speedtest user sample share. Northwest Territories and Yukon, launching in the same time frame as Nunavut, also follow the more-rural-less-population-dense logic, reaching 27.8% and 20.9% share in 2H 2024.
Vertical scale notwithstanding, the pattern is this: a relatively quick market adoption with stabilization after a year roughly. Share stabilization could be due to the Starlink service fulfilling latent market demand or the satellite constellation’s capacity limits being reached (halting further sales of the service, like around Edmonton currently, for example, per starlink.com/map), or a combination of the two.
Ontario (the protagonist), British Columbia, and Quebec have the lowest percentage of rural populations in Canada, and that is reflected in their lower samples shares – 4.9%, 4.5% and 2.2%, respectively.
Bringing it back together – what is at stake? Imagine that Starlink Speedtest user share is projectable to the population of Canada. That calculates to roughly 2 million Canadians who could feel the effects of such a cancellation. While many of those affected could switch to another internet service provider, for some in Canada – Nunavut especially – satellite connectivity is sometimes the only means of accessing the internet. Some portion of these people would be casualties in a trade war.
Ookla retains ownership of this article including all of the intellectual property rights, data, content graphs and analysis. This article may not be quoted, reproduced, distributed or published for any commercial purpose without prior consent. Members of the press and others using the findings in this article for non-commercial purposes are welcome to publicly share and link to report information with attribution to Ookla.
Customers of the two biggest cable mobile virtual network operators (MVNO) in the U.S. – Charter’s Spectrum Mobile and Comcast’s Xfinity Mobile – are both enjoying an impressive upward trajectory in their overall network performance. In their respective service areas over the past two years, Spectrum Mobile and Xfinity Mobile users’ download speeds increased by more than 100 Mbps. The performance lift is largely thanks to Wi-Fi offloading capabilities through Spectrum Mobile’s Speed Boost and Xfinity Mobile’s PowerBoost.
Charter began augmenting speeds for its Spectrum Mobile customers leveraging their home Charter Wi-Fi in spring 2022 and, in the fall the same year offered speeds up to 1 Gigabit per second (Gbps). Likewise, Comcast was aggressive in tapping into its installed base of customer hotspots, and in 2023, Charter and Comcast offered their combined footprints to each other’s MVNO customers. That is, Charter’s Spectrum Mobile customers have access to the network of Comcast Wi-Fi hotspots within the Comcast service area – and visa-versa for Xfinity Mobile customers in the Charter service area.
Comcast said it operates more than 23 million Wi-Fi hotspots in its footprint. Charter says it has 43 million Wi-Fi access points in its and its partners’ networks, the latter including the Comcast figure. Both cable companies say that between their own Wi-Fi networks and those of partners, with minimal overlap, they provide near-nationwide (population) coverage.
Then, in April 2024, Comcast took the lid off with WiFi Boost offering speeds up to 1 Gbps, and just last week have rebranded the feature to PowerBoost. These steps by Charter and Comcast to converge the mobile and Wi-Fi networks are meaningfully boosting their customer experience in home and out-and-about.
We examined the last eight quarters (Q1 2023 – Q4 20241) of Speedtest IntelligenceⓇ data for the median download speed performance of Speedtest users for customers of Spectrum Mobile, Xfinity Mobile, U.S. mobile providers as a group, and Verizon; the latter is the host mobile network for both MVNOs. We analyzed the Spectrum Mobile and Xfinity Mobile results within their two respective geographic service areas of Charter and Comcast, where “public” Wi-Fi is available to enhance available speeds of their mobile customers. (“Public,” in our usage here, means access points managed by Charter and Comcast, available to Spectrum Mobile and Xfinity Mobile customers.) We do not have knowledge about the wholesale network relationships or any performance implications as part of their agreements.
Spectrum Mobile – They Have Separation
Charter’s earlier start in promoting Speed Boost had its Spectrum Mobile users consistently clocking higher overall download speeds in its service area than Comcast did for Xfinity Mobile users in its service area. (Charter and Comcast service areas do not overlap.)
During these past two years, Spectrum Mobile customers saw their median download speeds increase by more than double from 84.35 Mbps to 188.63 Mbps – a 104.28 Mbps increase. The pause in its quarterly climb in Q2 and Q3 2024 is explained, at least partly, by the underlying host mobile network slowing down.
Xfinity Mobile – Up and Up
Xfinity Mobile, on the other hand, was not affected the same way. Its later introduction of WiFi Boost, compared to Spectrum Mobile’s Speed Boost, was still providing gains during this period. Speed Boost, however, had already realized the speed gains from Wi-Fi (also explaining why Spectrum Mobile was faster than Xfinity Mobile), thus being more sensitive to the mobile network performance. Interestingly, the Xfinity Mobile and Spectrum Mobile speeds became closer to one another in the second half of 2024, as WiFi Boost (now PowerBoost)was catching up.
Xfinity Mobile users, comparing year-on-year Q4 2023 to Q4 2024, saw their download speeds increase more than 2.5 times from 66.60 Mbps to 170.39 Mbps.
Quarter-on-quarter, Xfinity Mobile Speedtest users experienced a relentless climb in their download speed performance starting from Q3 2023 at 54.81 Mbps:
• Q4 2023 up 11.79 Mbps to 66.60 Mbps • Q1 2024 up 29.60 Mbps to 96.20 Mbps • Q2 2024 up 24.85 Mbps to 121.05 Mbps • Q3 2024 up 21.86 Mbps to 142.91 Mbps • Q4 2024 up 27.48 Mbps to 170.39 Mbps
The Verizon network provides the Spectrum Mobile and Xfinity Mobile customer experience, when not on Wi-Fi. During this same two-year period, nationwide Verizon’s overall median download speeds increased from 66.81 Mbps in Q1 2023 to 97.45 Mbps in Q4 2024 – up 30.64 Mbps. Increased network capacity from Verizon’s ongoing C-Band spectrum deployment is a rising tide that floats the boat for Verizon’s customers, as well as for Spectrum Mobile and Xfinity Mobile customers.
To Xfinity, Spectrum and Beyond
Hardware is a key component to faster speeds, and PowerBoost benefits from a new access point with Wi-Fi 6E technology that triples the available bandwidth versus the prior generation. Comcast began offering its 6E-capable xFi Advanced Gateway in 2022.
Comcast also is making improvements to its underlying network. The company is currently upgrading from DOCSIS 3.1 to DOCSIS 4.0, doubling downstream capacity and quadrupling upstream capacity. In September 2024 the cable company said it had deployed DOCSIS 4.0 to parts of 10 markets and 1 million homes.
And, as older handsets are replaced with newer and faster ones, the technology ecosystem as a whole comes together to deliver an ever-better customer experience.
And just as technology advancements mentioned above have provided the foundation for the performance gains these past two years, more is on the way.
The XB10 gateway, coming in 2025, will be Comcast’s fastest and most powerful device yet – supporting WiFi 7 and DOCSIS 4.0 – and will deliver multi-gigabit symmetrical speeds over WiFi, with the unprecedented capacity to connect to 300 devices simultaneously. The XB10 also includes AI technology that will help ensure a consistent experience. Comcast WiFi
Implicit in the march of technology is not just the network foundation, but also the consumer adoption. New access points require new cabling, and newer devices offer improved capabilities, and often paying for the “right” bundled service offering is needed. For example:
For Spectrum Speed Boost, you need Spectrum Internet, Spectrum Mobile and Advanced WiFi. Spectrum Speed Boost
It’s not a secret that the vast majority of mobile device data traffic is actually served by Wi-Fi, rather than the mobile network. Comcast President Michael Cavanagh noted in a recent call with investors, “My final thought on broadband is the importance of bundling with mobile, with 90% of Xfinity Mobile smartphone traffic traveling over our Wi-Fi network.”
Noteworthy is Cavanagh’s mention of “bundling,” which, in the context of the foregoing analysis, is the combining of networks in a way to best benefit the customer experience. Comcast and Charter are proving this to be the case.
This approach of combining the networks is part of a strategic narrative happening in the U.S. telecoms industry called network “convergence.” Were it not for AI, convergence could have been the industry’s top buzzword in 2024. And, like with AI, there will be plenty more said about convergence in 2025.
Ookla retains ownership of this article including all of the intellectual property rights, data, content graphs and analysis. This article may not be quoted, reproduced, distributed or published for any commercial purpose without prior consent. Members of the press and others using the findings in this article for non-commercial purposes are welcome to publicly share and link to report information with attribution to Ookla.
The Maine Connectivity Authority (MCA) recently made it possible for homes and businesses in remote locations of the state to become eligible to receive low-Earth orbit satellite internet service from Starlink. According to Ookla data, Starlink Speedtest users in Maine are able to get median download speeds of 116.77 Mbps, which surpasses the FCC’s threshold for broadband.
Maine is a Top 10 state in Ookla’s U.S. Broadband report, ranking 9th with a 30% year-on-year increase in Speedtest users who achieved broadband speeds (1H 2023 vs 1H 2024). Yet only 37.6% of Rural Speedtest users experienced speeds at or above the FCC thresholds of 100 Mbps download and 20 Mbps upload speeds for broadband, lagging its Urban counterparts by 13.3 percentage points.
Taking a look to see how the MCA Starlink program might help, here are current Starlink Speedtest user results in Maine:
1 Oct – 15 Dec 2024
10th percentile (Slower)
Median
90th percentile (Faster)
DL (Mbps)
22.81
116.77
250.96
UL (Mbps)
6.79
18.17
27.17
Latency* (ms)
72
47
37
* Multi-server latency
With a median of 116.77 Mbps download speed, a majority of Starlink Speedtest users are already exceeding the FCC’s 100 Mbps download speed target minimum. The median upload speed of 18.17 Mbps is close to the FCC’s 20 Mbps upload speed target minimum.
Taking a look at the slower results (that is, the 10th percentile, where 90% of Speedtest user tests results are faster), this cohort saw download speeds of 22.81 Mbps and upload speeds of 6.79 Mbps. Coincidentally, there exists a lower threshold of 25 Mbps download speed and 3 Mbps upload speed that divides the Unserved (below 25/3 Mbps) from the Underserved (above 25/3 Mbps; over 100/20 Mbps is considered Served by broadband internet).
These results are based upon the existing end-user equipment and the current satellite constellation. Newer customer premise equipment and recent (and future) satellite launches could improve these results.
Satellites are an important solution for addressing the digital divide in rural areas, as the MCA program with Starlink represents. The Starlink Speedtest results demonstrate this, and we look forward to revisiting this data as this program takes flight.
Ookla retains ownership of this article including all of the intellectual property rights, data, content graphs and analysis. This article may not be quoted, reproduced, distributed or published for any commercial purpose without prior consent. Members of the press and others using the findings in this article for non-commercial purposes are welcome to publicly share and link to report information with attribution to Ookla.
More than 80% of Canadians have access to fixed broadband networks, but for rural Canadians that figure drops to just 60%. Yet, this gap between who has broadband access and who doesn’t is closing in rural areas at a rate nearly three times faster than in urban areas. Canada is narrowing its rural broadband divide thanks to a clearly articulated and well-funded connectivity strategy to ensure high-speed internet access for all.
Using Speedtest Intelligence® data, this report identifies Canada’s Provinces and Territories that are delivering the minimum standard for fixed broadband speeds, as established by the Canadian Radio-television and Telecommunications Commission (CRTC). Based on data from the 1H 2024 (and compared to 1H 2023), it also analyzes performance in other geographic splits, including Urban-vs-Rural and the Remoteness Index.
Key Takeaways
As many as 2 million more Canadians enjoyed broadband speeds in the first half of 2024 compared to the first half of 2023.
Rural Canadian Speedtest users saw a 23% increase in those with broadband speeds in 1H 2024over 1H 2023.
Satellite internet service plays a key role in closing the broadband divide for Canada’s vast geography. In the U.S., regulators were ambivalent about allowing satellite internet to qualify for government broadband funding (this attitude has recently begun to change in favor). Canada knew many years ago that satellite internet was critical.
Broadband in the Spotlight
While the COVID-19 pandemic shined a light on the digital divide, the government of Canada has long been investing in broadband deployments to close the gap. In 2014 it established the Connecting Canadians program (CCP), allocating C$305M to improve connectivity for 300,000 underserved households.
The Connect to Innovate (CTI) program was launched in December 2016 with C$500M (and C$85M added to CTI in 2019) to expand high-speed Internet in communities underserved by the private sector. The CTI aimed to improve access for over 380,000 homes.
Deepening its investments in 2020, perhaps in response to COVID-19 lockdowns, the Canadian government launched the C$3.225B Universal Broadband Fund (UBF).
UBF
Universal Broadband Fund (2020)
$3.225B
CTI
Connect To Innovate (2016)
$585M
CCP
Connecting Canadians Program (2014)
$305M
With over C$4B from these programs alone, the CRTC has a goal of connecting 98% of Canadians to high-speed internet (broadband) delivering at least 50 Mbps download (DL) and 10 Mbps upload (UL) speeds (50/10 Mbps) by 2026, and 100% by 2030.
Conquer Divide
Canada’s population of 41 million is concentrated in a handful of large, urban metropolitan areas near the U.S. border. However, despite the country’s large land mass, it is highly urbanized. Approximately two-thirds, or 27 million, of Canadians live within 100 kilometers of the U.S. border —about the distance from the border to Winnipeg —yet this is only 4% of Canada’s land mass.
Imagine a line across lower Canada, 100 km north of Canada’s southern border – crossing Winnipeg in the middle of the population distribution map. In the area below the line and above the border with the U.S., 82.2% of Speedtest users are getting the CRTC’s minimum standard for fixed broadband speeds (50/10 Mbps). For Speedtest users north of the 100 km line, 76.3% met (or exceeded) the standard. Just a 5.9 percentage points gap.
However, the 100 km line is rather blunt and unsophisticated (and imaginary). Instead let’s look at the Urban-vs-Rural division as well as a more-granular designation based on the Remoteness Index.
Compared with the 100 km line’s gap, the Urban-vs-Rural digital divide aligns more closely with expectations of a larger disparity between these geographic areas. Specifically, 83.8% of Urban users meet the 50/10 Mbps standard, compared to 60.0% of Rural users resulting in a 23.8 percentage point gap.
The good news is the gap has closed from the prior year with Rural users seeing a 23% improvement compared to 5% for Urban users. To emphasize this further, in the first half of 2023 fewer than 50% of Rural users were able to get broadband speeds of 50/10 Mbps. This indicates that efforts to address this gap (i.e., UBF) were targeting the right places.
Continuing across the table, median download and upload speeds are roughly three times faster among urban than rural Speedtest users (DL 251.62 Mbps vs 90.76 Mbps and UL 64.82 Mbps vs 18.88 Mbps, respectively). Comparing the rural median speeds to the 50/10 Mbps threshold shows that half of Speedtest users in rural Canada enjoy download speeds that are 40.76 Mbps (i.e., 90.76 minus 50) and upload speeds that are 8.88 Mbps (i.e., 18.88 minus 10) faster than the target. (This is not to say that an individual Speedtest user experiences both upload and download speeds over the threshold, which is required for the target.)
The difference in Urban and Rural latency literally demonstrates the difference in physical distances (which not only means farther, but also more hops or switching; even buffering time based on the capacity of the data transport) .
The Remoteness Index presents a similar picture in finer geographic slices. (As one would expect, since the underlying data is the same.) Not surprisingly, the percentage of Speedtest users that meet the 50/10 Mbps threshold decreases as their location moves from least remote to most remote. Here too, speed and latency degrade at each step of remoteness, correlating with those meeting the 50/10 Mbps threshold percentages.
The percentage change from the prior year tells a similar story of greater improvement in the more remote geographies, but with some nuance. While we might expect Most Remote to demonstrate even more improvement and be greater than More Remote, deployment challenges and associated costs to deliver telecommunications infrastructure in the most remote geographies can be exponentially prohibitive. Thus, this extreme lack of population density coupled with difficult topography explain why government programs like CTI are needed, and why other solutions like satellite are viable.
The Provinces and The Territories
In the above analysis, broadband speeds were assessed using an imaginary 100 km demarcation, illustrating the concentration of population along Canada’s southern border. Before examining broadband performance and the digital divide among Canada’s provinces and territories, here are additional facts about Canada’s relative sparseness of people in its vast geography.
Landmass: The combined land area of the territories (Yukon, Northwest, Nunavut) is larger than the land area of India, the world’s 7th largest (and most populous) country
Population: The population of the territories is equivalent to the total number of births across Canada every four months.
The territories hold just 0.3% of Canada’s population on 39% of its land. Canada is often compared to the U.S., but Australia is a better comparison, with both Canada and Australia ranking among the least densely populated countries in the world.
Now let’s examine the 50/10 Mbps threshold in the provinces and territories.
Province / Territory
% Meeting 50/10 Mbps 1H 2024
Change from 1H 2023
Urban-Rural Gap 1H 2024, %pts
Newfoundland and Labrador
81.7
6%
22.5
British Columbia
79.8
4%
21.9
New Brunswick
78.1
2%
14.0
Nova Scotia
77.8
9%
8.2
Québec
76.5
8%
9.8
Ontario
76.0
8%
33.9
Alberta
75.7
5%
30.0
Manitoba
71.7
11%
20.8
Prince Edward Island
71.2
14%
18.5
Saskatchewan
64.7
17%
33.0
Northwest Territories
57.3
8%
-7.1
Yukon Territory
53.2
14%
6.7
Nunavut
36.2
94%
Not meaningful
In this table, by geography, percentage of Speedtest users achieving the CRTC broadband speed targets in 1H 2024, compared with the same period in prior year, and the digital divide.
Following the logic of the Remoteness Index, the territories have the fewest Speedtest users meeting the 50/10 Mbps threshold. The percentage changes in the territories from 2023 do not follow the same pattern seen in the Urban-vs-Rural chart and Remoteness Index because these geographies are a mix of these geographic definitions (as it is in the provinces). Similarly, the Urban-vs-Rural gap result is confounded by sparse population. In Nunavut, more than half of the population is defined as rural.
On the top of the table, Newfoundland and Labrador, New Brunswick, and Nova Scotia may be unexpected leaders in meeting the 50/10 Mbps threshold. New Brunswick and Nova Scotia have relatively higher population density among the provinces and territories which (economically) encourages the deployment of telecommunications infrastructure. This appears to play out in the Urban-vs-Rural gap as well, ranking among the lowest gaps in the provinces.
The exceptional performance of Newfoundland and Labrador (81.7% meeting 50/10 Mbps) can partly be attributed to the vast majority of its population residing on the island of Newfoundland, and half of them, in turn, residing on the Avalon peninsula (see the population distribution map above and St. John’s in the east). This concentration of population underscores the fundamental reality of economics in telecommunications deployment. And in the opposite direction, the Urban-vs-Rural gap (22.5%pts) also makes this same point for the need for the funding programs like UBF to address the digital divide.
Breaking Down the Digital Divide
Percentage of Urban and Rural Speedtest users in each Province/Territory with broadband speeds of at least 50/10 Mbps, 1H 2024, Nunavut: Urban n too small; Rural 43.6%
Look, Up in the Sky
As addressed in the discussion about Most Remote, because of Canada’s topographical challenges, fiber and electricity are cost prohibitive in many deployment cases. In 2019, Canada added C$85M to its CTI program because it recognized that it needed support for low-Earth orbit (LEO) satellites to reach its connectivity goals (50/10 Mbps connectivity to 95% of Canadians by 2026, and the hardest-to-reach Canadians by 2030).
Briefly looking across Canada for Speedtest users of satellite internet services during the first half of 2024, over half saw download speeds of 72.90 Mbps or greater, and upload speeds of 12.47 or greater. Moreover, in the territory of Nunavut the speeds were basically identical (75.16 Mbps and 12.50 Mbps, respectively), which makes sense since Nunavut is equally Urban or Rural (or More Remote or Less Remote) to an orbiting satellite a few hundred miles overhead. Clearly LEO is a viable solution technically and economically.
Whether fiber or satellite, broadband connectivity means nothing without power. The cost of electricity in the north can be ten times more expensive than in southern cities. In some cases, diesel fuel burned for both heat and electricity is flown, shipped by sea, or by tanker over frozen lakes and rivers in the winter. As with broadband, the Canadian government has a plan to invest in Rural and Northern Communities to make available affordable and clean energy.
Another barrier to fully-connected communities that affects Urban as well as the Most Remote is digital literacy (though the latter faces compounding factors). Here again the Canadian government is addressing the issue with initiatives such as its Digital Literacy Exchange Program.
No matter the geographic lens – Urban-vs-Rural, Remoteness Index, or Provinces and Territories – the goal is to get to 100% in 2030. But even 100% broadband connectivity from a purely technical perspective, would not be 100% in spirit without other programs and initiatives like these. Communication is achieved when the signal is received, not just sent.
Recently, Ookla also looked at the broadband and digital divide in the U.S. – How the 50 U.S. States Stack up in Broadband Speed Performance: 1H 2024 | Ookla®. We look forward to providing more updates on the U.S. and Canada’s progress to provide high-speed internet connectivity for all. For more information about Speedtest Intelligence data and insights, please get in touch.
Le Rétrécissement de la Fracture Numérique au Canada
Plus de 80 % des Canadiens ont accès à des réseaux fixes à large bande, mais pour les Canadiens des régions rurales, ce chiffre tombe à seulement 60 %. Pourtant, cet écart entre ceux qui ont accès à la large bande et ceux qui n’en ont pas se rétrécit dans les régions rurales à un rythme près de trois fois plus rapide que dans les zones urbaines. Le Canada réduit la fracture numérique dans les régions rurales grâce à une stratégie de connectivité clairement articulée et bien financée pour assurer l’accès à l’Internet haute vitesse pour tous.
À l’aide des données de Speedtest Intelligence®, le présent rapport identifie les provinces et les territoires du Canada qui fournissent la norme minimale pour les vitesses à large bande fixes, telle qu’établie par le Conseil de la Radiodiffusion et des Télécommunications Canadiennes (CRTC). Sur la base des données du 1er semestre 2024 (et par rapport au 1er semestre 2023), il analyse également les performances dans d’autres catégories géographiques, y compris les catégories urbaine et rurale ainsi que l’indice d’éloignement (Remoteness Index).
Principaux points à retenir
Pas moins de 2 millions de Canadiens supplémentaires ont bénéficié de l’accès à Internet haute vitesse au premier semestre 2024 par rapport au premier semestre 2023.
Les utilisateurs ruraux canadiens de Speedtest ont vu une augmentation de 23 % de ceux qui ont accès à Internet haute vitesse au cours du premier semestre 2024par rapport au premier semestre 2023.
Le service Internet par satellite constitue un levier essentiel pour réduire la fracture numérique et améliorer l’accès à la large bande sur l’immense territoire canadien. Aux États-Unis, les organismes de réglementation étaient ambivalents quant à autoriser le financement par le gouvernement de l’Internet par satellite (cette attitude a récemment commencé à changer). Le Canada savait il y a de nombreuses années que l’Internet par satellite était essentiel.
La large bande sous les feux de la rampe
Alors que la pandémie de COVID-19 a mise en lumière la fracture numérique, le gouvernement du Canada investit depuis longtemps dans le déploiement de la large bande pour réduire cet écart. En 2014, il a mis sur pied le programme Un Canada branché, allouant 305 millions de dollars canadiens pour améliorer la connectivité de 300 000 ménages mal desservis.
Le programme Brancher pour innover a été lancé en décembre 2016 avec 500 millions de dollars canadiens (85 millions de dollars canadiens supplémentaires alloués en 2019) pour étendre l’accès a l’Internet à haut débit dans les communautés mal desservies par le secteur privé. Ce programme visait à améliorer l’accès à plus de 380 000 foyers.
Renforçant ses investissements en 2020, peut-être en réponse aux confinement lié au COVID-19, le gouvernement canadien a lancé le Fonds universel pour la large bande, doté de 3,225 milliards de dollars canadiens.
UBF
Fonds universel pour la large bande (2020)
3,225 milliards de dollars
CTI
Brancher pour innover (2016)
585 M$
CCP
Programme « Un Canada branché » (2014)
305 M$
Avec plus de 4 milliards de dollars canadiens provenant de ces programmes uniquement, le CRTC a pour objectif de connecter 98 % des Canadiens à l’Internet haute vitesse (large bande) offrant des vitesses d’au moins 50 Mbps en téléchargement (DL) et 10 Mbps en téléversement (UL) (50/10 Mbps) d’ici 2026, et 100 % d’ici 2030.
Vaincre la fracture
La population de 41 millions d’habitants du Canada est concentrée dans une poignée de métropoles près de la frontière américaine. Malgré l’immensité du pays, le Canada est fortement urbanisé. Environ les deux tiers, soit 27 millions, des Canadiens vivent à moins de 100 kilomètres de la frontière américaine, soit environ la distance entre la frontière et Winnipeg; mais cette zone ne représente que 4 % de la superficie du Canada.
Imaginez une ligne traversant le Canada, à 100 km au nord de sa frontière méridionale, et traversant Winnipeg au milieu de la carte de répartition de la population. Dans la zone située sous la ligne de démarcation et au-dessus de la frontière avec les États-Unis, 82,2 % des utilisateurs de Speedtest obtiennent la norme minimale du CRTC pour les vitesses des services à large bande fixe (50/10 Mbps). Pour les utilisateurs du Speedtest au nord de la ligne des 100 km, 76,3 % respectent (ou dépassent) la norme. Soit un écart de seulement 5,9 points.
Cependant, la ligne des 100 km reste une mesure assez simpliste, peu sophistiquée et, de surcroît, imaginaire. Examinons plutôt la division entre les zones urbaines et rurales, ainsi qu’une classification plus détaillée basée sur l’indice d’éloignement.
Comparée à celle définie par la ligne des 100 km, la fracture numérique entre les zones urbaines et rurales correspond davantage aux attentes, reflétant une disparité plus marquée entre ces zones géographiques. Plus précisément, 83,8 % des utilisateurs urbains respectent la norme de 50/10 Mbps, comparativement à 60,0 % des utilisateurs ruraux, ce qui donne un écart de 23,8 points.
La bonne nouvelle est que l’écart s’est réduit par rapport à l’année précédente, les utilisateurs ruraux ayant constaté une amélioration de 23 %, contre 5 % pour les utilisateurs urbains. Pour souligner davantage ce point, au cours du premier semestre de 2023, moins de 50 % des utilisateurs ruraux ont pu obtenir des vitesses à large bande de 50/10 Mbps. Cela indique que les efforts déployés pour combler cette lacune (c’est-à-dire le Fonds universel pour la large bande) ont été utilisés à bonne fin.
Si l’on poursuit la lecture du tableau, les vitesses médianes de téléchargement et de téléversement sont environ trois fois plus élevées chez les utilisateurs urbains que chez les utilisateurs ruraux de Speedtest (DL 251,62 Mbps contre 90,76 Mbps et UL 64,82 Mbps contre 18,88 Mbps, respectivement). La comparaison des vitesses médianes en milieu rural avec le seuil de 50/10 Mbps montre que la moitié des utilisateurs de Speedtest dans les régions rurales du Canada bénéficient de vitesses de téléchargement de 40,76 Mbps (c.-à-d. 90,76 moins 50) et de vitesses de téléversement de 8,88 Mbps (c.-à-d. 18,88 moins 10) plus rapides que l’objectif fixé. (Cela ne signifie pas pour autant qu’un utilisateur individuel de Speedtest bénéficie à la fois de vitesses de téléchargement et de téléversement dépassant le seuil requis, ce qui est nécessaire pour atteindre l’objectif.)
La différence entre les temps de latence en milieu urbain et en milieu rural illustre littéralement la différence entre les distances physiques (ce qui signifie non seulement des distances plus grandes, mais aussi un plus grand nombre de sauts de traffic).
L’indice d’éloignement présente une image similaire dans des tranches géographiques plus fines. (Comme on pouvait s’y attendre, puisque les données sous-jacentes sont les mêmes.) Il n’est pas surprenant de constater que le pourcentage d’utilisateurs Speedtest qui atteignent le seuil de 50/10 Mbps diminue au fur et à mesure que cette indice augmente. Ici aussi, la vitesse et la latence se dégradent avec l’augmentation de l’indice, en corrélation avec le pourcentage d’utilisateurs atteignant de seuil de 50/10 Mbps.
La variation (en pourcentage) par rapport à l’année précédente offre des similarités, avec une amélioration dans les zones géographiques les plus éloignées, mais avec tout en apportant une certaine nuance. On pourrait s’attendre à ce que les régions les plus éloignées s’améliorent, mais les difficultés de déploiement et les coûts associés à la mise en place d’une infrastructure de télécommunications dans les zones géographiques les plus reculées peuvent être exponentiellement prohibitifs. Ainsi, ce manque extrême de densité de population, associé à une topographie difficile, explique pourquoi des programmes gouvernementaux sont nécessaires et pourquoi d’autres solutions comme l’accès par satellite sont viables.
Les provinces et les territoires
Dans l’analyse ci-dessus, les vitesses à large bande ont été évaluées à l’aide d’une démarcation imaginaire de 100 km, illustrant la concentration de la population le long de la frontière sud du Canada. Avant d’examiner les performances de la large bande et la fracture numérique parmi les provinces et territoires du Canada, voici quelques faits supplémentaires concernant la relative faiblesse de la densité de population dans l’immensité géographique du pays.
Masse continentale : La superficie terrestre combinée des territoires (Yukon, Nord-Ouest, Nunavut) est plus grande que la superficie de l’Inde, le 7e pays le plus grand (et le plus peuplé) du monde
Population : La population des territoires équivaut au nombre total de naissances au Canada tous les quatre mois.
Les territoires ne représentent que 0,3 % de la population du Canada mais 39 % de son territoire. Le Canada est souvent comparé aux États-Unis, mais l’Australie est une meilleure comparaison; le Canada et l’Australie se classant parmi les pays les moins densément peuplés du monde.
Examinons maintenant le seuil de 50/10 Mbps dans les provinces et les territoires.
Province / Territoire
% Atteignant l’objectif 50/10 Mbps 1H 2024
Changement par rapport à 1H 2023
Écart entre les zones urbaines et rurales 1H 2024, %pts
Terre-Neuve-et-Labrador
81,7
6 %
22,5
Colombie-Britannique
79,8
4 %
21,9
Nouveau-Brunswick
78,1
2 %
14,0
Nouvelle-Écosse
77,8
9 %
8,2
Québec
76,5
8 %
9,8
Ontario
76,0
8 %
33,9
Alberta
75,7
5 %
30,0
Manitoba
71,7
11 %
20,8
Île-du-Prince-Édouard
71,2
14 %
18,5
Saskatchewan
64,7
17 %
33,0
Territoires du Nord-Ouest
57,3
8 %
-7,1
Territoire du Yukon
53,2
14 %
6,7
Nunavut
36,2
94 %
Pas significatif
Pourcentage d’utilisateurs de Speedtest atteignant les objectifs de vitesse de large bande du CRTC au premier semestre 2024 par région par rapport à la même période de l’année précédente.
Suivant la logique de l’indice d’éloignement, les territoires ont le moins d’utilisateurs Speedtest répondant au seuil de 50/10 Mbps. Les variations en pourcentage dans les territoires à partir de 2023 ne suivent pas la même tendance que celle observée dans le graphique urbain/rural et l’indice d’éloignement, car ces géographies sont un mélange de ces définitions géographiques (comme c’est le cas dans les provinces). De même, le résultat de l’écart entre les zones urbaines et les zones rurales est faussé par la faible densité de population. Au Nunavut, plus de la moitié de la population est définie comme rurale.
En haut du tableau, Terre-Neuve-et-Labrador, le Nouveau-Brunswick et la Nouvelle-Écosse pourraient être des chefs de file inattendus dans la réalisation du seuil de 50/10 Mbps. Le Nouveau-Brunswick et la Nouvelle-Écosse ont une densité de population relativement plus élevée parmi les provinces et les territoires, ce qui encourage (économiquement) le déploiement de l’infrastructure de télécommunications. Cela semble également se jouer dans l’écart entre les régions urbaines et rurales, qui se classe parmi les écarts les plus faibles dans les provinces.
La performance exceptionnelle de Terre-Neuve-et-Labrador (81,7 % atteignant 50/10 Mbps) peut en partie être attribuée au fait que la grande majorité de sa population réside sur l’île du Terre Neuve, dont la moitié vit sur la péninsule d’Avalon (voir la carte de répartition de la population ci-dessus et la ville de St. John’s à l’est). Cette concentration de population souligne la réalité fondamentale de l’économie dans le déploiement des télécommunications. Et dans la direction opposée, l’écart entre les zones urbaines et rurales (22,5 % de points) souligne également la nécessité de programmes de financement pour combler la fracture numérique.
Briser la fracture numérique
Pourcentage d’utilisateurs de Speedtest fixe en milieu urbain et rural dans chaque province et territoire ayant accès à des vitesses à large bande de 50/10 Mbps, 1H 2024, Nunavut: Urbain n trop petit; Rural 43.6%
Regardez vers les étoiles
Comme mentionné précédemment, en raison des défis topographiques du Canada, la fibre et l’électricité peuvent avoir des coûts de déploiement prohibitifs dans de nombreux cas. En 2019, le Canada a ajouté 85 millions de dollars canadiens à son programme Brancher pour innover (CTI), car il a reconnu qu’il avait besoin de satellites en orbite basse (LEO) pour atteindre ses objectifs en matière de connectivité (connectivité de 50/10 Mbps pour 95 % des Canadiens d’ici 2026, et pour les Canadiens les plus difficiles à atteindre d’ici 2030).
En examinant brièvement les utilisateurs de Speedtest des services Internet par satellite à travers le Canada durant la première moitié de 2024, plus de la moitié ont enregistré des vitesses de téléchargement de 72,90 Mbps ou supérieures, ainsi que des vitesses de téléversement de 12,47 Mbps ou supérieures. De plus, dans le territoire du Nunavut, les vitesses étaient pratiquement identiques (75,16 Mbps et 12,50 Mbps, respectivement), ce qui est logique, puisque le Nunavut est à la fois urbain et rural (ou plus éloigné et moins éloigné) pour un satellite en orbite à quelques centaines de kilomètres au-dessus. Il est clair que le LEO est une solution viable techniquement et économiquement.
Qu’il s’agisse de fibre ou de satellite, la connectivité à large bande ne signifie rien sans électricité. Le coût de l’électricité dans le nord peut être dix fois plus élevé que dans les villes du sud. Dans certains cas, le carburant diesel utilisé pour le chauffage et l’électricité est transporté par avion, par bateau ou par camion-citerne sur des lacs et des rivières gelés en hiver. Comme pour la large bande, le gouvernement canadien a un plan pour investir dans les collectivités rurales et nordiques afin de rendre disponible une énergie propre et abordable.
Un autre obstacle à des communautés pleinement connectées, qui touche tant les zones urbaines que les régions les plus éloignées, est la maîtrise des outils numériques (bien que ces dernières fassent face à des facteurs aggravants). Là encore, le gouvernement canadien s’attaque au problème avec des initiatives comme son Programme d’échange en matière de littératie numérique.
Peu importe la perspective géographique – urbain par rapport à rural, indice d’éloignement ou provinces et territoires – l’objectif est d’atteindre 100 % en 2030. Mais même une connectivité à large bande à 100 %, d’un point de vue purement technique, ne serait pas à 100 % dans l’esprit sans d’autres programmes et initiatives comme ceux-ci. La communication est réalisée lorsque le signal est reçu, pas seulement envoyé.
Récemment, Ookla a également examiné la fracture numérique et l’accès à la large bande aux États-Unis. – Comment les 50 États américains se positionnent en matière de vitesse de la large bande : 1H 2024 | Ookla®. Nous sommes impatients de fournir d’autres mises à jour sur les progrès réalisés par les États-Unis et le Canada pour fournir une connectivité Internet haute vitesse pour tous. Pour plus d’informations sur les données et les analyses de Speedtest Intelligence, veuillez prendre contact.
Ookla retains ownership of this article including all of the intellectual property rights, data, content graphs and analysis. This article may not be quoted, reproduced, distributed or published for any commercial purpose without prior consent. Members of the press and others using the findings in this article for non-commercial purposes are welcome to publicly share and link to report information with attribution to Ookla.
Network performance stats in the tech press catch our attention here at Ookla – especially download and upload speeds or latency. Here’s a quick look into this press release about the Starlink satellite internet service being offered by Claro Colombia. The release states that Starlink will offer download speeds ranging from 40 to 220 Mbps, upload speeds ranging from from 8 to 25 Mbps, and latency in the range of 20 to 60 ms.
Speed and Latency Ranges Are In Reach
On the surface, Starlink’s performance claims appear possible for users to achieve, even though Q3 2024 Speedtest user data for Starlink in Colombia is not yet there.
The article doesn’t state how its ranges were determined, so, to make a comparison, we constructed our own ranges by examining the lowest and highest 10th percentiles, along with medians, of Starlink Speedtest users.
For download speeds, the Speedtest range (8 – 140 Mbps) fell short of the Starlink range (40 – 220 Mbps) and the median Speedtest (54 Mbps) was well short of the middle of the Starlink range (130 Mbps). Still, there was a healthy overlap, and there are other factors that can explain slower speeds in Speedtest user test than what Starlink is providing, such as older user equipment or Wi-Fi propagation path loss.
To get a picture of download speeds, we mapped our Starlink Speedtest users’ tests based on the 40 – 220 Mbps range with a color break at the 130 Mbps midpoint. The light red and light green squares coalesced, forming an area of tests within the range – a clear visual majority. And while many tests fell below the range (dark red squares), a few were also above range (dark green squares).
Turning to upload speeds, the Speedtest range (4 – 27 Mbps) and the Starlink range (8 – 25 Mbps) were much closer than download speeds. Likewise, Speedtest median (14 Mbps) and the Starlink midpoint (17 Mbps) are close, too.
Last, the latency Speedtest range (30 – 116 ms) and the Starlink range (20 – 60) were also closer than they may appear. Latency tests are susceptible to extreme values (very high latency) that result in a long-tail effect. More insightful for latency is the closeness of the Speedtest median (44 ms) and the Starlink mid-point (40 ms).
Quick Take Away
Starlink began service in Colombia in February 2023. It continues to add capacity to its constellation and introduce new terminals to support more use cases.
The Speedtest perspective finds the speed and latency ranges in the press release to be credible, and we look forward to revisiting the data to see more Speedtest users enjoying the benefit of the Claro Colombia and Starlink alliance.
Low for Starlink is the bottom of its stated range for speed and top for latency; for Speedtest it is the slowest 10th percentile for speed and highest 10th percentile for latency ↩︎
Mid for Starlink is the simple average of its stated range; for Speedtest it is the median speed ↩︎
High for Starlink is the top of its stated range for speed and bottom for latency; for Speedtest it is the fastest 10th percentile for speed and lowest 10th percentile for latency ↩︎
Ookla retains ownership of this article including all of the intellectual property rights, data, content graphs and analysis. This article may not be quoted, reproduced, distributed or published for any commercial purpose without prior consent. Members of the press and others using the findings in this article for non-commercial purposes are welcome to publicly share and link to report information with attribution to Ookla.
