Testing High-Speed (Multi-Gigabit) Internet Access QoS and QoE

When Internet speeds surpass interface, hardware and processing capabilities of consumer electronics (laptop, tablet, phone, stream box, WiFi), subscribers using free speed test services may get the wrong QoS impression. It can be costly for the ISP

Recent worldwide shelter-in-place events (2020) have shown that even best effort services such as broadband Internet access can become essential services. That is, when most of the population works, learns and communicates remotely, the network must cope with usage surge requirements that perhaps were never considered a plausible scenario.

Building for typical rather than peak demand is considerably less expensive. As we’ve seen, network planning is mostly based on predicted exponential growth of Internet applications and services, which require greater bandwidth. In order to cope with such demand, Internet service providers (ISPs) started deploying new access technologies such as PON fiber-to-the-home (FTTH), DOCSIS 3.1, R-PHY/DAA, Extended Spectrum DOCSIS (ESD) and/or DOCSIS 4.0, allowing them to make gigabit Internet services available to their home and enterprise subscribers as well as preparing their networks for the unstoppable increasing surge of bandwidth demand. Although most of the focus is on access technologies, optimizing bandwidth sharing, metro, core and transport network upgrades are equally important to carry all the aggregated traffic.

New broadband Internet services, particularly those surpassing 500 Mbps and emerging ones the gigabit speed limits, have brought new challenges to ISPs, such as bottlenecks in the access network that slow down, and sometimes disrupt subscribers’ services. These bottlenecks can also be caused by legacy home or enterprise equipment, like 802.11n wireless routers that can’t keep up with the contracted Internet service speed. All of this translates into potentially higher volumes of service calls from upset subscribers claiming they are not receiving the promised speed, unnecessary truck rolls to subscribers’ premises, increased OPEX and notably important, bad user feedback on quality of experience (QoE).

Tech-savvy customers tend to test their Internet service speed with off-the-shelf equipment, like smart devices, consumer-oriented laptops or even high-end laptops. The problem with these CPU-based devices is that they lack the power and capabilities to test gigabit speeds at full-line rate. For instance, speed tests for line rates above 500 Mbps become unreliable when performed by even high-end laptops and are virtually useless at the 1 Gbps mark or above, since they are limited by the physical port or connection (e.g., 1000BASE-T or legacy Wi-Fi).

With gigabit Internet services growing at such a fast pace, ISPs face problems even in best case scenarios where service technicians use similar off-the-shelf devices to validate Internet speed at the time of installation. In worst case scenarios, crews don’t have test devices due to lack of or reduced budgets. Troubleshooting and isolating issues at the customer premise continues to be a challenge as the market lacks the necessary tools to test and measure gigabit Internet services accurately.

Traditional telcos and cable operators, which have been gradually updating their access networks to offer higher top speeds, tend to focus too much on what they know best — their cable plant, the physical layer. Although their access networks have been cleaned, groomed, fine-tuned and optimized, they may have paid little attention to data traffic engineering. It is not unusual to hear customer service representatives say these frustrating words to suffering and already upset customers, “I just ran some tests and everything looks fine on my end. The signal is good!” Having a “good pipe” or knowing how much speed the access network can handle in ideal conditions is not the point. It is about actual traffic, oversubscription and bottlenecks affecting subscribers and their families who are working from home and streaming all day, which could bring shared Internet access services to a crawl.

But how can ISPs cope with all the challenges mentioned above? The answer is simple, ISPs have an urgent need to improve the way they test Internet access speeds at the customer premises, with tools that are quick and easy to use, providing accurate, repeatable and reliable test results. That requires some hardware-based test gear.

VeEX’s V-TEST™ feature is a truly scalable, low cost and reliable Internet access speed test solution, with an intuitive user interface that every field technician can use it within minutes.

Opposite to free speed test apps and websites, V-TEST is a high performance, FPGA-based test solution integrated in Ethernet test equipment, which eliminates the need to carry additional expensive test gear or "companion" boxes. With up to 100 Gbps full line rate capability, the truly scalable V-TEST uses the required physical interfaces of the measurement instrument to test links at 100/1000BASE-T, 1000BASE-X, 2.5/5/10GBASE-T, 10GBASE-X, 25GE, 50GE, 100GE, etc. One hardware-based centralized test unit (Remote Test Unit or RTU) can use on of its 100GE links to run multiple lower rate tests to serve up to 128 field test sets, with guaranteed performance for each stream. The V-TEST solution enables repeatability and reliability in testing methodology and procedures.

The Layer 4+ option in VeEX's test sets feature Layer 4 Internet speed tests (V-TEST) and iPerf compatible RFC6349 tests (V-PERF), measuring true customer QoE. Internet access speed tests require proprietary HTTP servers or Speedtest® by Ookla® servers. With its multiple testing modes, the Layer 4+ test bundle gives the flexibility to load the user’s proprietary list of servers, manually enter server details or use Ookla servers.

Service providers delivering traditional carrier Ethernet services have adopted the RFC6349 testing methodology. This approach requires high performance and dedicated TCP and UDP servers, field portable test equipment (for the service assurance/service turn-up phase) and centralized test heads capable of supporting speeds from 1G to 100G.

VeEX’s RTU-320 (10G), the newer RTU-340 (16G) and the RTU-600x (100G) multi-service test modules operate as both TCP/UDP servers and clients when testing against field test equipment. The RTU-300+ test modules perform bi-directional upstream and downstream tests, IPv6 V-PERF, bidirectional V-PERF that combines client-to-server and server-to-client tests into one. With up to two independent test blades, the RTU-320 and RTU-340 enable 4x 10 Gbps test interfaces, while the RTU-600x enables up to 2x 100 Gbps test interfaces, ensuring reliable and repeatable data. (Each 100GE test port, of the RTU-600x module, can handle more than 100 simultaneous test sessions, to server large field tech crews. For example, 100 field techs running a hundred 1GE tests towards a single RTU-300+ centralized test head.)

In summary, legacy equipment and home networking issues are proving challenging for ISPs. To avoid deployment failure and leave no room for doubt that the gigabit Internet service meets customer expectations, ISPs should equip their service technician crews with reliable test equipment that perform consistent speed tests and stateful TCP throughput like RFC6349. Using test solutions like VeEX’s Layer 4+ functionality and one RTU-320, RTU-340 or RXT-600x (centralized) remote test units, field technicians can verify and assure true customer QoE.

Testing High-Speed (Multi-Gigabit) Internet Access QoS and QoE

When Internet speeds surpass interface, hardware and processing capabilities of consumer electronics (laptop, tablet, phone, stream box, WiFi), subscribers using free speed test services may get the wrong QoS impression. It can be costly for the ISP

Related Test Solutions