In today’s digital landscape, the term “Network Transport” is frequently mentioned, often surrounded by a cloud of confusion. Powered by readily available, yet sometimes simplistic, explanations, particularly from sources like ChatGPT, it’s crucial to delve into a more robust and engineer-led understanding. This guide offers a detailed overview of network transport options, aiming to clarify the concepts and empower you when making decisions about connecting your critical locations. While drawing examples from Flexential’s portfolio, the principles and solutions discussed are broadly applicable across various providers and customer scenarios.
Defining Network Transport: Private vs. Public Connectivity
Let’s begin by clearly defining network transport. At its core, network transport refers to private connectivity designed to link specific, predetermined locations. This is fundamentally different from internet transit or general internet access, which provides public connectivity to virtually any location across the globe. Think of network transport as building your own private roads between your offices or data centers, as opposed to using the public highway system (the internet).
This distinction is crucial. Network transport prioritizes security, predictability, and often, higher performance for connections between your key infrastructure points. A wide array of products and technologies fall under the umbrella of network transport, each offering unique advantages, costs, and technical characteristics. Common categories include:
- Layer 2 Point-to-Point Connectivity: Dedicated links between two sites at the data link layer.
- Layer 2 Multipoint Connectivity: A network allowing multiple sites to connect as if on the same local network at Layer 2.
- Layer 3 Multipoint Connectivity: Similar to Layer 2 Multipoint but operating at the network layer (Layer 3), utilizing routing protocols.
- Layer 1 Wavelengths or Dark Fiber: Physical layer connections offering raw bandwidth capacity.
- SD-WAN, IPSEC, and Tunnel-based approaches: Software-defined solutions leveraging tunnels for secure connectivity, often over the internet.
This list provides a solid foundation and covers the most prevalent network transport methods deployed today. Let’s explore each category in detail.
Layer 2 Point-to-Point: Dedicated Connections
Layer 2 Point-to-Point transport solutions establish direct, dedicated connections between two locations at Layer 2 of the OSI model. Technologies like Ethernet Private Line (EPL) and L2VPN Pseudowire are prominent examples. You might also encounter Ethernet Virtual Private Line (EVPL), an evolution that allows for multiple point-to-point connections to be aggregated onto a single physical interface, increasing efficiency.
While subtle technical variations exist, the core function of these products is consistent: securely and reliably linking two sites as if they were directly cabled together. Layer 2 point-to-point solutions are known for their maturity, high reliability, and cost-effectiveness, particularly for connecting a limited number of sites or specific workloads.
For instance, Flexential’s Data Center Interconnect (DCI) service, based on L2VPN Pseudowire technology, exemplifies this category. DCI is frequently chosen for disaster recovery (DR) scenarios, connecting primary and backup data centers. In such use cases, the focus is on high capacity and resilience for critical workloads between two key locations, rather than broad, multi-site connectivity.
Standard DCI DR use case and deployment
However, Layer 2 point-to-point connectivity encounters challenges as network scale increases, especially in fully meshed network designs where every site needs direct communication with every other site. The complexity escalates rapidly.
Point-to-point at scale
The number of connections required in a fully meshed point-to-point network grows exponentially with the number of sites. The formula c = s * (s – 1) / 2 illustrates this, where ‘c’ is the total connections and ‘s’ is the number of sites. Connecting just ten sites in this manner necessitates a staggering 45 individual point-to-point connections, assuming single hand-offs at each location. This complexity can become operationally and financially prohibitive as networks expand.
Layer 2 Multipoint Connectivity: The Mesh Approach
Layer 2 Multipoint connectivity addresses the scalability limitations of point-to-point solutions by enabling a many-to-many connection model. Often categorized under the umbrella term E-LAN (Ethernet LAN), you may also see references to Ethernet Virtual Private LAN (EVPLAN) or technologies like VPLS (Virtual Private LAN Service) that underpin these services. Despite varied marketing names, these solutions fundamentally provide Layer 2 multipoint-to-multipoint connectivity. From an operational perspective, it’s as if all your sites are plugged into the same virtual switch, operating on the same VLAN.
IX Mesh – Layer 2 operator experience
Multipoint-to-multipoint architecture offers significant advantages, primarily in scalability and simplified failover management. Adding new sites to a Layer 2 multipoint network is straightforward, requiring no changes or additional connections to existing sites. Furthermore, managing redundancy and failover is often more intuitive compared to point-to-point setups.
With point-to-point DCI connections, implementing failover typically involves routing traffic across different DCI circuits and devices, as each DCI represents a separate network segment. In contrast, Layer 2 multipoint places all connections within the same network domain. This is particularly beneficial for clients using active/passive firewall clusters at their network edge, where multipoint simplifies failover mechanisms considerably.
Flexential’s Interconnection Mesh product is a prime example of a Layer 2 multipoint offering. The diagram below highlights the complexities clients often encounter when implementing failover with traditional point-to-point solutions.
Limitations of Point-to-Point Failover
However, Layer 2 multipoint solutions like Interconnection Mesh generally come with a higher price point compared to point-to-point options, especially when connecting only a few (2-3) sites. The cost-benefit analysis shifts as the number of connected locations increases, making multipoint increasingly attractive for larger deployments.
A crucial technical decision when considering multipoint solutions is choosing between Layer 2 and Layer 3 implementations. This isn’t solely a matter of cost versus benefit, but rather a trade-off between management responsibility and network control.
- Layer 2 Multipoint: Clients retain full control and responsibility for all network routing within their multipoint network. They manage routing protocols and configurations across all connected sites.
- Layer 3 Multipoint: Clients participate in routing but rely on the service provider as a central routing point for all connected sites. The provider handles the core routing infrastructure and management.
The choice hinges on the level of routing control desired. Clients seeking maximum control over their network routing infrastructure often prefer Layer 2 multipoint, while those prioritizing ease of management might lean towards Layer 3 options.
Layer 3 Multipoint Connectivity: Provider-Managed Routing
Layer 3 Multipoint Connectivity, commonly known as L3VPN (Layer 3 VPN) or IPVPN (IP VPN), provides multipoint connectivity at the network layer. It’s important to address a common industry misnomer: while some mistakenly refer to this as “MPLS,” MPLS (Multiprotocol Label Switching) is actually an underlying protocol that supports L3VPN, not the product itself.
The operational experience with Layer 3 Multipoint is similar to Layer 2 Multipoint E-LAN, but instead of the provider presenting a Layer 2 switch, they offer a Layer 3 router interface at each site. Functionally, Layer 2 and Layer 3 multipoint solutions can often address similar networking challenges. The ultimate decision often boils down to client preference, desired level of control, and specific use cases.
L3VPN solutions offer significant value for clients with limited in-house routing expertise or resources, particularly those new to dynamic routing protocols like BGP (Border Gateway Protocol). By positioning the provider as a central router, L3VPN simplifies multi-site network management. Clients typically manage a single BGP peering session with the provider, rather than needing to establish and manage peering or routing relationships directly with every remote site. Scaling and making network changes are also simplified. Adding a new site to a Layer 2 multipoint network requires logical configuration updates at existing sites, while L3VPN often eliminates the need for any changes at existing sites, both physical and logical.
The trade-off with L3VPN is reduced control and potentially limited feature sets. Clients who require granular control over their routing policies or need advanced BGP features not supported by the provider might find L3VPN restrictive. Many standard L3VPN products lack support for advanced BGP functionalities like community-based traffic engineering or highly granular route filtering. For organizations demanding precise routing control, Layer 2 solutions are generally more suitable.
Flexential offers both Layer 2 (E-LAN) and Layer 3 (L3VPN) multipoint options within its Interconnection Mesh product family. Customer choices between these options often reflect their priorities regarding management simplicity versus routing control, with both being viable solutions for large, multi-site deployments.
L3VPN also presents a distinct advantage in specific use cases, particularly when connecting to cloud environments or large numbers of Layer 2 devices. Some clients prefer L3VPN to avoid complexities associated with Layer 2 protocols like spanning-tree in large or heterogeneous networks. While this is a niche concern, network engineers familiar with spanning-tree will recognize its potential challenges. In essence, any Layer 2 network deployment must carefully consider and manage spanning-tree protocol behavior. For certain deployments, adopting a Layer 3 product like L3VPN provides a simpler path by circumventing spanning-tree issues altogether. Flexential often recommends L3VPN for hybrid and cloud connectivity scenarios for this very reason.
Layer 2 vs. Layer 3 IX Mesh
Layer 1 Wavelengths or Dark Fiber: Raw Capacity
For organizations requiring massive bandwidth and ultimate control, Layer 1 Wavelengths or Dark Fiber offer network transport at the most fundamental physical layer. Dark Fiber is precisely what it sounds like: unused fiber optic cable. Clients purchasing dark fiber are responsible for supplying and managing all optical equipment needed to transmit data across the fiber. Wavelengths, while also operating at Layer 1, are carrier-managed services. The provider manages the optical layer, providing a dedicated portion of the optical spectrum, typically a specific wavelength, to the client, rather than the entire fiber strand as with dark fiber. The operational experience is akin to having a physical cable directly connecting your devices, without any intermediary switching or routing.
The primary driver for choosing Layer 1 transport is the need for enormous bandwidth capacity. Standard wavelength services often start at 100Gbps, and concepts like “committed information rates” are typically not applicable. Clients essentially lease the entire capacity of the wavelength.
The majority of Flexential’s clients do not require this level of raw bandwidth and control, so Layer 1 offerings are more limited in their portfolio. In dense metropolitan areas like Hillsboro, Oregon, Flexential does maintain some Dark Fiber offerings, but it is not a product line broadly available across all markets. Layer 1 solutions are typically reserved for very specific high-bandwidth, low-latency applications.
SD-WAN, IPSEC, and Tunnel-based Approaches: Over-the-Top Transport
SD-WAN (Software-Defined WAN) represents a more abstract and software-centric approach to network transport. The SD-WAN market is dynamic and evolving, with numerous vendors and a lack of standardized definitions. “SD-WAN” is not a single technology but rather a category of solutions employing diverse techniques.
A common SD-WAN methodology involves establishing IPSEC tunnels over the public internet to create secure connections between sites. IPSEC tunnels inherently provide both site-to-site connectivity and private communication via encryption, thus fitting within the broader definition of “network transport.” Historically, relying on internet tunnels for critical transport was viewed with skepticism due to concerns about internet reliability. However, advancements in internet infrastructure and SD-WAN technologies have changed this landscape. Modern SD-WAN solutions incorporate proprietary error detection and correction mechanisms to mitigate the inherent variability of internet performance. Consequently, some organizations are now strategically replacing traditional dedicated transport with internet-based tunnel solutions. While cost used to be a primary driver for this shift, the declining cost of dedicated transport has made the economic argument less clear-cut.
The inherent variability of internet performance and the lack of universal standards in SD-WAN mean that the effectiveness of this transport approach can vary significantly. The quality of the underlying internet connection is paramount.
For clients adopting SD-WAN and tunnel-based transport, investing in robust and reliable internet service is crucial. Flexential’s IP Bandwidth product line is designed to address this need, offering a blended internet service aggregating multiple Tier 1 ISPs with full physical and logical redundancy and a 100% uptime SLA for redundant configurations. Even the most sophisticated SD-WAN with advanced error correction cannot compensate for a fundamentally unreliable internet connection.
While SD-WAN and tunnel-based transport encompass a far broader range of features and considerations, within the context of network transport, the key takeaway is the critical dependence on a high-quality underlying internet service.
Conclusion: Choosing the Right Network Transport Solution
Selecting the optimal network transport solution is a critical decision, and there is no one-size-fits-all answer. Partnering with a provider that offers a comprehensive portfolio of services is essential to ensure you can build and adapt your network to meet evolving needs.
Flexential provides a wide spectrum of interconnection services, including:
- Interconnection Mesh: A multipoint, any-to-any network with both Layer 2 and Layer 3 options, offering scalability and flexibility.
- DCI (Data Center Interconnect): A Layer 2 point-to-point service for dedicated, high-performance connections between specific sites.
- Dark Fiber cross connects: Enabling direct Layer 1 fiber connections for maximum bandwidth and control within supported locations.
Connecting your various locations involves careful consideration of numerous factors. Understanding the high-level categories of network transport outlined in this guide is a crucial first step in navigating the options for site-to-site connectivity. Effective network planning and optimization require expertise, and Flexential Professional Services offers workshops designed to assist you in developing a robust network strategy. For further questions or to explore Flexential’s specific interconnection products in detail, please contact your Flexential representative.
Thank you for reading! We hope this guide has provided valuable clarity on the landscape of network transport. Happy networking!
Learn more about Network Transport and the full portfolio of Flexential Interconnection offerings, register for a workshop, or contact us for more information.
