Analyzing the impact of local caching on YouTube video streaming performance in the Middle East
Video streaming is one of the most popular services globally, driving a large share of internet traffic. Global content providers efficiently distribute assets via caching architectures, while operators aim to localize traffic paths to optimize user experience and mitigate international transit costs. We use Ookla® Consumer QoE™ data to analyze the level of in-country video content caching in select countries in the Middle East (Bahrain, Jordan, Kuwait, Qatar, Saudi Arabia, and the U.A.E.) and assess how it impacts streaming quality metrics.
Key takeaways:
- Video start time performance in the benchmarked countries is very good overall. Ookla data from February 2026 shows that the selected countries had short video start times ranging from 1.45s in Kuwait to 1.69s in Qatar; that means that the delay between the viewer clicking to play and the first frame appearing was minimal. These values confirm a strong YouTube video streaming experience for most users.
- Local caching ratios are high but vary significantly between operators. While countries like Kuwait and the U.A.E. have in-country caching ratios (the proportion of video tests that use local cache over the total number of tests) above 94%, Jordan and Saudi Arabia rely more on out-of-country caching. Since the in-country caching ratio is generally high, operators should focus on resolving local network issues before exploring external ones.
- Video streaming indicators sharply degrade when content is sourced internationally. Routing traffic to international nodes generally incurs a latency penalty on video KPIs, causing increased delays, with out-of-country video start time and rebuffering time (a component of start time representing the delay to load data before the first frame appears) significantly higher than the local cache. The penalty can range from less than 0.5 seconds to more than six seconds, like in Qatar.
- Operators can maximize cache efficiency. Operators should analyze in-country traffic to identify and address bottlenecks, and coordinate with market peers to install dedicated cache nodes to reduce reliance on expensive international transit links.
Video streaming, particularly YouTube, is the main driver of internet traffic globally
Live or on-demand video streaming accounts for a large share of current internet traffic due to its high bandwidth requirements, the growing number of users, the popularity of short-form videos, and social media platform proliferation. According to AppLogic Networks’ report, YouTube was responsible for 10% of global internet traffic in 2025, while TikTok contributed 5-7%.
Video-on-demand has also been a major driver of broadband adoption and revenue growth for both telecom operators and content platform providers. The global video streaming market was valued at US$156.86 billion in 2025 and is projected to grow to reach US$692.68 billion by 2033. Video is also increasingly used for marketing and branding. The digital video advertising market is forecast to increase by US$636.3 billion between 2024 and 2029, driven by the increasing popularity of in-app advertising and increased spending on online video advertising.
The surge in demand, volume, and quality of streamed content places significant pressure on operators’ networks and content providers, and increases delivery costs, specifically for content delivery network (CDN) bandwidth, network upgrades, and maintenance. If content providers cannot address this adequately, it will inflame the ‘fair use’ debate, pressuring streaming companies and large traffic generators to contribute a share of the transport cost.
Shortening the traffic path is beneficial for operators, content providers, and consumers
Global content caches serve popular content repeatedly to multiple concurrent users, reducing traffic demand and eliminating upstream traffic to the content’s origin. Caching also reduces the cost of international bandwidth or IP transit when content originates abroad. Local caches allow operators to serve higher definition video to end users, as traffic is localized and less likely to be congested. For content providers, it also means they need less bandwidth to serve more users and improve customer retention. And for users, it can result in a better, faster experience.
A good example is Google Global Cache (GGC). This distributed caching architecture is located at ISPs and edge points, forming part of Google’s infrastructure that serves YouTube videos. Popular content gets temporarily cached on GCC nodes so that when a user requests a video, the request is routed via Domain Name System (DNS) and IP routing mechanisms to the nearest (or most optimal) GCC node based on the user’s location and ISP. Google also deploys other technologies, such as adaptive streaming and load balancing, to reduce buffering and latency.
If the content is not cached, then it’s retrieved from edge Points of Presence (PoPs) at interconnection facilities (Google has over 100 interconnection facilities globally), which are connected to Google’s various data centres located in the Americas, Europe, and Asia. Other global streaming video service providers, such as Netflix, have similar caching architectures to efficiently distribute their content globally.
Disparities in local content caching across Middle Eastern nations highlight the potential to optimize end-user experience.
According to Ookla’s aggregated Consumer QoE™ data from February 2026, the Middle Eastern countries in this analysis demonstrated fast video start times (the duration a user waits for playback to begin). Results generally indicate a high-quality YouTube streaming experience with average times spanning from 1.45s in Kuwait to 1.69s in Qatar. Nevertheless, these national averages hide notable variations among individual operators, which stem from differences in traffic routing strategies and the effectiveness of local caching.
We determine if an in-country cache was used by analyzing video server IP addresses from the Consumer QoE data. Our data shows that the U.A.E., Kuwait, and Bahrain have the highest in-country caching ratios, all recording values above 90%. Conversely, Jordan and Saudi Arabia exhibit a higher out-of-country caching proportion of 18.1% and 28.4%, respectively.
When traffic is routed outside the country, it incurs a latency penalty on video start times, resulting in longer buffering times. In Jordan, Saudi Arabia, and Kuwait, the out-of-country video start time is between 25% and 87% higher than for in-country traffic; the difference rises to around 140% in Bahrain and the U.A.E.
Qatar showed significantly more severe video delays, representing the largest latency penalty observed. Video start time and rebuffering time were 5.5x and nearly seven times higher, respectively, compared to when local cache was used. This is due to the use of Google’s servers in Ireland and the U.S. when requested content is not readily available or local cache servers are saturated. The low proportion of video requests from international CDNs (11.5%) translates into a modest but notable impact on video streaming KPIs for Qatar.
There are also significant differences when comparing video delays associated with in-country CDNs. Given that most markets maintain a high in-country caching ratio, operators should prioritize the investigation of performance bottlenecks within their own and their peers’ local CDN networks before addressing out-of-country issues.
On-peak load in Saudi Arabia forces the delegation of video traffic to international CDNs
We analyzed our Saudi Arabian video streaming data to determine why its local caching ratio of 71.6% is lower than that of its neighbors. Our data highlights a significant disparity among providers, with one operator relying more heavily on international caches than its peers. Specifically, 44.0% of Operator A’s YouTube traffic was served from out-of-country servers, whereas other operators maintained this figure at just over 6%. This could be a result of one or a few cache servers hosted by Operator A being forced to offload their traffic to Google servers outside the country due to capacity constraints.
Operator A had a large share of video cache requests exiting the country (44.0%), but a relatively low penalty of 0.5s compared to the other two. This suggests Operator A has excellent IP transit links or a closer proximity to international PoPs. However, the wider performance gap between in-country and out-of-country cache for Operators B and C suggests these providers face a greater performance challenge when traffic is forced international, despite their superior overall caching ratios (just over six percent).
Median Initial Buffering Duration and Median Start Times, per Traffic Route, per Operator, Saudi Arabia
Source: Consumer QoE® | February 2026
Median Initial Buffering Duration and Median Start Times, per Traffic Route, per Operator, Saudi Arabia
Consumer QoE data allows us to identify daily traffic patterns that lead to an increase in internationally routed traffic. For the three operators, we notice four distinct video traffic peaks in demand: at 5 AM, 12 PM, 3 PM, and 5 PM, aggregated over February. This shows that at these times, more users stream YouTube videos, creating demand for Google’s CDNs, and leading some users to be directed towards international CDNs.
However, Operator A’s traffic pattern suggests a much higher level of video traffic offload to servers outside the country than the other two operators, and this trend has been persistent. This shows that while local servers can handle video traffic during off-peak hours, they struggle to handle the on-peak load, especially later in the day (at 3 PM and 5 PM). To address the high volume of internationally routed traffic, Operator A could upgrade the capacity and connectivity speeds of its local cache servers, expand their footprint, and improve their geographical distribution.
Video Test Traffic per Time of the Day, Destination, and Operator, Saudi Arabia
Source: Consumer QoE® | February 2026
Video Test Traffic per Time of the Day, Destination, and Operator, Saudi Arabia
Reducing international traffic and maximizing cache efficiency
Ookla’s data shows Middle East operators excel in video streaming KPIs, but opportunities remain to integrate deeper, local caching solutions and optimize national caching servers. These strategies offer clear benefits: content providers serve higher quality videos, operators realize lower infrastructure costs and improved customer retention, and end users enjoy high-resolution, lag-free videos. In short, optimizing traffic routes ensures the profitable use of network infrastructure as video demand continues to grow.
Please contact us for more details on how tools such as Consumer QoE can help provide actionable insights into network performance and resilience.
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.
