Skip to main content

VDSL2 and GPON study finds sweet spots


The business case for both VDSL2 vectoring and GPON can be compelling, and Bell Labs Consulting has been helping fixed access providers discover their unique sweet spot for each. A new study analyzes a variety of subscriber mixes, morphologies, and outside plant conditions to guide operators toward the optimal approach in the medium and long term.


Bell Labs has conducted numerous studies comparing multiple solutions worldwide for customers in a variety of service areas and demographics. This particular study focused on comparing VDSL2 and GPON technologies where subscribers demand ultra-broadband access speeds across a range of serving areas with characteristics that varied according to various parameters, including:

  • Customer base
  • Area topology and morphology
  • Existing network and facilities
  • Cost factors
  • Competition

Representative distribution areas

It isn’t feasible to analyze every distribution area since there can be hundreds of thousands in a fixed access network. So as a 1st step, representative areas needed to be identified for study.

After reviewing all candidate distribution areas, consultants classified them according to a series of key parameters and defined a set of 8 to 12 different area types. In this case, the key determining parameters were subscriber density, % of residential vs. business customers and copper loop quality. For each type, the fixed access technologies were evaluated to compare capital expenses, operational expenses, total cost of ownership (TCO), and the business cases.


In service areas with demand for ultra-broadband access, is copper based VDSL2 or GPON/fiber-to-the-home the better choice? Some might assume that a fiber based solution is the answer, but there’s more than meets the eye.

Looking at Figure 1, we see lab-based access speeds as a function of copper loop lengths. If loop length is 400 m or less, subscribers get up to 100 Mbps when VDSL2 vectoring is used. If the length extends to up to 1 km, this bandwidth can still be achieved using more than one loop pair with bonding.

Figure 1. Results from VDSL2 lab trial

One caveat: A variety of factors, such as cable quality, can reduce a loop’s effective length making it shorter than its physical length. Using the Nokia Motive Network Analyzer–Copper, this study measured the copper-loop lengths in order to determine the percentage of subscribers less than 1 km from the street cabinet.

Although 1 km may seem limited, that’s not the full story. As shown in Figure 2, a VDSL2 architecture with street cabinets can reach subscribers up to 5 kms from the central office.

Figure 2. Street cabinet-based VDSL2 architecture


Unlike greenfield areas where fiber is the clear choice, this study focused on brownfield areas where the choice of access technology is not obvious. Bell Labs consultants evaluated the initial capital cost of deploying each technology. They considered 3 scenarios where the maximum speed upgrade would be 20 Mbps, 50 Mbps, and 100 Mbps.

The result? In cases where 75% of the distribution plant was aerial, the capital cost of VDSL2 was approximately US$400-500 per subscriber, whereas GPON was approximately $700-750 per subscriber.1 As shown in Figures 3 and 4, the top 5 items account for 80% of the total cost.

Figure 3. Capital cost components for VDSL2 (20 Mbps per subscriber)

Figure 4. Capital cost components for GPON

Copper based access costs are sensitive to the maximum bandwidth offered while, for their part, fiber access doesn’t vary as much. For instance, the capital cost of VDSL2 for 20 Mbps service is between $350-400 per subscriber. For 50 Mbps, the cost increases to $400-480 and for 100 Mbps the cost is $520-600.

The CAPEX bottom line: Even in the most conservative case, the per subscriber difference in capital outlay between the 2 technologies is hundreds of dollars. For this reason, an accurate picture of the services and bandwidth demand is critical when VDSL2 is deployed.


Copper infrastructure maintenance is typically the most costly of all network types. By comparison, GPON is significantly cheaper — upwards of 45% less to operate than VDSL2. This difference is mainly attributable to greater power consumption and loop maintenance costs associated with the latter.

That’s why replacing aging copper plant with fiber has significant operational benefits. New fiber is more reliable than old copper while, at the same time, consumes much less energy. Also, with GPON operational costs are further reduced with the elimination of the digital subscriber line access multiplexer (DSLAM) from the access architecture.

However, on an annual, per-subscriber basis, the absolute difference in on-going OPEX between the two technologies in this case study represents about $5 USD per subscriber per year. It should be noted that this applied to a Latin American country where the labor rates were relatively low. Actual costs will vary on a case by case basis depending on a variety of factors such as labor rates, fuel costs and deployment types (e.g., indoor vs. outdoor cabinets).

However, even when accounting for these variations and considering a 5- or 10-year horizon, the small difference in OPEX doesn’t tip the scales in favor of GPON. According to this Bell Labs analysis, the high capital costs of GPON are so great for the brownfield scenarios that VDSL2 wins out.


The viability of VDSL2 in brownfield service areas depends upon reuse of existing copper loops. So what about areas where a significant percentage of the loops need to be refurbished. Is VDSL2 still more cost effective than GPON?

Bell Labs consultants modeled operations maintenance costs based on a service provider’s actual ticket volumes from existing FTTH and xDSL service areas. Taking into account loop quality conditions, the study included the capital cost of rehabilitating the copper distribution plant. Plus, trouble ticket tracking data was analyzed to identify faulty customer drops along with the capital cost of replacing the drop cables.

The study shows that loop quality rehabilitation does drive up VDSL2 deployment costs by approximately US$100 per subscriber. That said, even with this additional expense it still has a lower TCO and higher net present value than fiber based access.


Let’s consider the business case for ultra-broadband access. As we’ve discussed, for brownfield operators upgrading copper access to VDSL2 provides a compelling case with lower CAPEX and shorter time-to-market than a fiber based solution. But what about the end-to-end business case over 5 years?

Each case depends on the specific input criteria. Subscriber demand and willingness to pay will depend on a variety of factors including local economics, demographics and broadband consumption culture; competition, pricing bundles and availability of quality content.

Figure 5 shows a brownfield study result based on Bell Labs studies performed with customers in Latin America and Eastern Europe. Here we see that GPON barely breaks even in 5 years, and only if a portion of the customers subscribe to at least 50 Mbps. While with VDSL2, the service provider breaks even between 2 and 4 years, depending on the speeds offered.

Figure 5. VDSL2 versus GPON business case comparison


Although this Bell Labs economic study shows the viability of VDSL2 for ultra-broadband access, there are still several cases where fiber trumps copper. These include:

  • Greenfield areas where it is most cost-effective to serve the premises with fiber
  • Distribution areas, such as commercial districts or business parks, where a significant percentage of subscribers are enterprises with demands that exceed the speeds that can be supported by the copper infrastructure
  • Areas with intense competition where the service provider wants to showcase services that require fiber and ultra-high access speeds


In many cases, the choice may not be clear between 1 technology or another. Rather, subscriber bandwidth demands in different serving areas will require the flexibility to offer a mix of technologies concurrently.

Service providers will need to deploy copper and fiber access concurrently within the same network. Service providers also need to have the flexibility to quickly upgrade from VDSL2 to GPON, if and when demand increases, such as when a competitor enters an area currently served by an incumbent.

With this in mind, Bell Labs consultants developed variations of the above business cases for 2 10--year evolution scenarios. Scenario 1 is a serving area upgrade with deployment of GPON from day 1. Scenario 2 is deployment of VDSL2 followed by an upgrade to GPON in year 6 when bandwidth demand warrants it.

As shown in Figure 6, scenario 2 has a much shorter break-even point, with approximately 5% lower TCO and 8% higher Cumulative Discounted Cash Flow (CDCF) in year 10. It turns out that, even with the capital costs to deploy VDSL2 and later GPON, scenario 2 is still more lucrative for several reasons:

  • VDSL2 allows for faster time-to-market compared to GPON, so time-to-revenue is shorter
  • The capital investment to deploy FTTH is deferred in large part until year 6
  • VDSL2 feeder fibers and street cabinets are reused when upgrading to GPON

Figure 6. Comparison of the business case for initial versus deferred GPON deployment in a 100 Mbps scenario


For ultra-broadband access in the above presented brownfield situations, this Bell Labs Consulting study has found that the business case and TCO for VDSL2 technologies clearly outperforms GPON. As each case may vary widely depending on local conditions including costs (labor, fuel, etc.), revenues (customer base, competition), and even variations within particular serving areas, every case should be closely analyzed individually.


Economic studies from Bell Labs Consulting

Bring ultra-broadband to any user fixed access solution

The authors wish to acknowledge the contribution of Beth Polonsky to this article.


1 For a maximum bandwidth of 20 Mbps per subscriber.

Our authors look forward to your questions and comments.

Mohamed El-Sayed

About Mohamed El-Sayed

Mohamed El-Sayed is the consulting manager of the Network Strategy and Technology Evolution Practice of Bell Labs Consulting in Murray Hill, New Jersey. His current interests include architecture and design of 3G wireless networks, IP Multimedia Subsystems (IMS), intelligent optical networks, data/optical integration, and converged networks. His current group works with many customer teams in North America, Europe, and the Asia/Pacific and Caribbean/Latin American regions on projects related to network evolution and optimization. Dr. El-Sayed holds a Ph.D. in systems engineering from Case Western Reserve University in Cleveland, Ohio, and a B.Sc. and M.Sc. in electrical engineering from Cairo University, Egypt.

Samrat Kulkarni

About Samrat Kulkarni

Samrati Kulkarni is in the Converged High Performance Networks Modeling R&D Practice of Bell Labs Consulting in Murray Hill, New Jersey. He focuses on optical transport network planning and modeling of next-generation SONET/SDH, WDM, and fiber-to-the-x (FTTx) solutions for telecom service provider networks, providing optimized network designs, product configurations, cost modeling, and systems engineering of capacity planning tools. He has a B.S. in electrical engineering from Birla Institute of Technology and Science in Pilani, India, and an M.Sc. in electrical engineering from Purdue University in West Lafayette, Indiana. He also serves as a chair on the local Institute of Electrical and Electronics Engineers New Jersey Coast Chapter.

Robert Novo

About Robert Novo

Mr. Novo has more than 25 years of experience with Bell Labs and is an Alcatel-Lucent Technical Academy member. He has been the recipient of various Bell Labs and Alcatel-Lucent awards. He has published several articles and presented at technical conferences in English and Spanish. At Bell Labs, Mr. Novo has led network design, modeling and optimization projects in diverse technologies, including Voice over Internet Protocol (VoIP), IP Multimedia Subsystem (IMS), access, IPTV, wireless, optical, and circuit switched voice. He has also conducted studies in a wide range of areas, including capacity analysis, architectures and designs, readiness assessments, audits, best practices, optimization, transformation, migration, and bandwidth recovery. Mr. Novo has a M. Eng. degree in electrical engineering from Cornell University, Ithaca, New York, and a B.S. in computer and systems engineering from Rensselaer Polytechnic Institute, Troy, New York.

Article tags