Optimal VARP Deployment

Optimal Layer 3 Forwarding

Layer 3 forwarding is crucial in ensuring efficient and seamless network data transmission. Optimal Layer 3 forwarding, in particular, is an essential aspect of network architecture that enables the efficient routing of data packets across networks. In this blog post, we will explore the significance of optimal Layer 3 forwarding and its impact on network performance and reliability.

Layer 3 forwarding directs network traffic based on its network layer (IP) address. It operates at the network layer of the OSI model, making it responsible for routing data packets across different networks. Layer 3 forwarding involves analyzing the destination IP address of incoming packets and selecting the most appropriate path for their delivery.

Table of Contents

Highlights: Optimal Layer 3 Forwarding

The Role of Optimal Layer 3 Forwarding:

Optimal Layer 3 forwarding ensures that data packets are transmitted through the most efficient path, improving network performance. It minimizes packet loss, latency, and jitter, enhancing user experience. By selecting the best path, optimal Layer 3 forwarding also enables load balancing, distributing the traffic evenly across multiple links, thus preventing congestion.

Example: Arista with Large Layer-3 Multipath

Arista EOS supports hardware for Leaf ( ToR ), Spine, and Spline data center design layers. They have a wide product range supporting large layer-3 multipath ( 16 – 64-way ECMP ) with excellent optimal Layer 3-forwarding technologies. Unfortunately, multi-protocol Label Switching ( MPLS ) limits to static MPLS labels, which could become an operational nightmare, and as of yet, no Fibre Channel over Ethernet ( FCoE ) support.

Arista supports massive Layer-2 Multipath with ( Multichassis Link aggregation ) MLAG. Validated designs with Arista Core 7508 switches ( offer 768 10GE ports ) and Arista Leaf 7050S-64 support over 1980 x 10GE server ports with 1:2,75 oversubscription. That’s a lot of 10GE ports. Do you think layer 2 domains should be designed to that scale?

 

Related: Before you proceed, you may find the following helpful:

  1. Scaling Load Balancers
  2. Virtual Switch
  3. Data Center Network Design
  4. Layer-3 Data Center
  5. What Is OpenFlow

 



Optimal Layer 3 Forwarding

Key Optimal Layer 3 Forwarding Discussion Points:


  • Introduction to optimal layer 3 forwarding and what is involved.

  • Highlighting the details of using deep buffers.

  • Critical points on the use case of Arista and virtual ARP.

  • Technical details on load balancing enhancements and LACP fallback.

  • Technical details on Direct Server Return and detecting server failures.

 

Back to Basics: Router operation and IP forwarding

Every IP host in a network is configured with its IP address and mask and the IP address of the default gateway. Suppose the host wants to send traffic, which, in our case, is to a destination address that does not belong to a subnet that the host is directly attached to; the host passes the packet to the default gateway. For example, the default gateway would be a Layer 3 router.

The Role of The Default Gateway 

A standard misconception is how the address of the default gateway is used. People mistakenly believe that when a packet is sent to the Layer 3 default router, the sending host sets the destination address in the IP packet as the default gateway router address. However, if this were the case, the router would consider the packet addressed to itself and not forward it any further. So why configure the default gateway’s IP address, then?

First, the host uses the Address Resolution Protocol (ARP) to find the specified router’s Media Access Control (MAC) address. Then, having acquired the router’s MAC address, the host sends the packets directly to it as data link unicast submissions.

 

Benefits of Optimal Layer 3 Forwarding:

1. Enhanced Scalability: Optimal Layer 3 forwarding allows networks to scale effectively by efficiently handling a growing number of connected devices and increasing traffic volumes. It enables seamless expansion without compromising network performance.

2. Improved Network Resilience: By selecting the most efficient path for data packets, optimal Layer 3 forwarding enhances network resilience. It enables networks to quickly adapt to network topology or link failure changes, rerouting traffic to ensure uninterrupted connectivity.

3. Better Resource Utilization: Optimal Layer 3 forwarding optimizes resource utilization by distributing traffic across multiple links. This enables efficient utilization of available network capacity, reducing the risk of bottlenecks and maximizing the network’s throughput.

4. Enhanced Security: Optimal Layer 3 forwarding plays a role in network security by ensuring traffic is directed through secure paths. It enables the implementation of firewall policies and access control lists, protecting the network from unauthorized access and potential security threats.

 

Implementing Optimal Layer 3 Forwarding:

To achieve optimal Layer 3 forwarding, various technologies and protocols are utilized, such as:

1. Routing Protocols: Dynamic routing protocols, such as OSPF (Open Shortest Path First) and BGP (Border Gateway Protocol), enable networks to exchange routing information automatically and determine the best path for data packets.

2. Quality of Service (QoS): QoS mechanisms prioritize network traffic, ensuring that critical applications receive the necessary bandwidth and reducing the impact of congestion.

3. Network Monitoring and Analysis: Continuous network monitoring and analysis tools provide real-time visibility into network performance, enabling administrators to identify and resolve potential issues promptly.

 

Arista deep buffers: Why are they important?

A vital switch table you need to be concerned with for large 3 networks is the size of Address Resolution Protocol ( ARP ) tables. When ARP tables become full and packets are offered with the destination ( next hop ) that isn’t cached, the network will experience flooding and suffer performance problems.

Arista Spine switches have deep buffers, ideal for bursty and latency-sensitive environments. In addition, deep buffers are perfect when you have little knowledge of the application traffic matrix, as they can handle most types efficiently.

Finally, deep buffers are most useful in Spine layers as this is where traffic concentration occurs. If you are concerned that ToR switches do not have enough buffers, physically direct servers to chassis-based switches in the Core / Spine layer.

 

Optimal layer 3 forwarding  

Every data center has some mix of layer 2 bridging and layer 3 forwardings. The design selected depends on layer 2 / layer 3 boundaries. Data centers that use MAC-over-IP usually have layer 3 boundaries on the ToR switch. While fully virtualized data centers require large layer two domains ( for VM mobility ); VLANs span Core or Spine layers.

Either of these designs can result in suboptimal traffic flow. Layer 2 forwarding in ToR switches and layer 3 forwarding in Core may result in servers in different VLANs connected to the same ToR switches being hairpinned to the closest Layer 3 switch.

Solutions that offer optimal Layer 3 forwarding in the data center were available. These may include stacking ToR switches, architectures that present the whole fabric as a single layer 3 elements ( Juniper QFabric ), and controller-based architectures (NEC’s Programmable Flow ). While these solutions may suffice for some business requirements, they don’t have optimal Layer 3 forward across the whole data center while using sets of independent devices.

Arista Virtual ARP does this. All ToR switches share the same IP and MAC with a common VLAN. Configuration involves the same first-hop gateway IP address on a VLAN for all ToR switches and mapping the MAC address to the configured shared IP address. The design ensures optimal Layer 3 forwarding between two ToR endpoints and optimal inbound traffic forwarding.

Optimal VARP Deployment
Diagram: Optimal VARP Deployment

Load balancing enhancements

Arista 7150 is an ultra-low latency 10GE switch ( 350 – 380 ns ). It offers load-balancing enhancements other than the standard 5-tuple mechanism. Arista supports new load-balancing profiles. Load-balancing profiles allow you to decide what bit and byte of the packet you want to use as the hash for the load-balancing mechanism—offering more scope and granularity than the traditional 5-tuple mechanism.

 

LACP fallback

With traditional Link Aggregation ( LAG ), LAG is enabled after receiving the first LACP packet. This is because before receiving LACP packets, the physical interfaces are not operational and are down / down. This is viable and perfectly OK unless you need auto-provisioning. What does LACP fallback mean?

If you don’t receive an LACP packet and the LACP fallback is configured, one of the links will still become active and will be UP / UP. Continue to use the Bridge Protocol Data Unit BPDU ) guard on those ports, as you don’t want a switch to bridge between two ports and create a forwarding loop.

 

 

Direct server return

7050 series supports Direct Server Return. The load balancer in the forwarding path does not do NAT. Implementation includes configuring VIP on the load balancer’s outside IP and the internal servers’ loopback. It is essential not to configure the same IP address on server LAN interfaces, as ARP replies will clash. The load balancer sends the packet unmodified to the server, and the server sends it straight to the client.

Requires layer 2 between the load balancer and servers; load balancer needs to use MAC address between the load balancer and servers. It is possible to use IP, which is called Direct Server Return IP-in-IP. Requires any layer 3 connectivity between the load balancer and servers.

Arista 7050 IP-in-IP Tunnel supports essential load balancing, so one can save the cost of not buying an external load-balancing device. However, it’s a scaled-down model, and you don’t get the advanced features you might have with Citrix or F5 load balancers.

 

Link flap detection

Networks have a variety of link flaps. Networks can experience fast and regular flapping; sometimes, you get irregular flapping. Arista has a generic mechanism to detect flaps so you can create flap profiles that offer more granularity to flap management. Flap profiles can be configured on individual interfaces or globally. It is possible to have multiple profiles on one interface.

 

Detecting failed servers

The problem is when we have scale-out applications, and you need to detect server failures. When no load balancer appliance exists, this has to be with application-level keepalives or, even worse, Transmission Control Protocol ( TCP ) timeouts. TCP timeout could take minutes. Arista uses Rapid Indication of Link Loss ( RAIL ) to improve performance. RAIL improves the convergence time of TCP-based scale-out applications.

 

OpenFlow support

Arista matches 750 complete entries or 1500 layer 2 match entries, which would be destination MAC addresses. They can’t match IPv6 or any ARP codes or inside ARP packets, which are part of OpenFlow 1.0. Limited support enables only VLAN or layer 3 forwardings. If matching on layer 3 forwarding, match either the source or destination IP address and rewrite the layer 2 destination address to the next hop.

Arista offers a VLAN bind mode, configuring a certain amount of VLANs belonging to OpenFlow and another set of VLANs belonging to standard Layer 3. Openflow implementation is known as “ships in the night.”

Arista also supports a monitor mode. Monitor mode is regular forwarding with OpenFlow on top of it. Instead of allowing the OpenFlow controller to forward forwarding entries, forwarding entries are programmed by traditional means via Layer 2 or Layer 3 routing protocol mechanism. OpenFlow processing is used in parallel to conventional routing—openflow then copies packets to SPAN ports, offering granular monitoring capabilities.

 

DirectFlow

Direct Flow – I want all traffic from source A to destination A to go through the standard path, but any HTTP traffic goes via a firewall for inspection. i.e., set the output interface to X and a similar entry for the return path, and now you have traffic going to the firewall but for port 80 only.

It offers the same functionality as OpenFlow but without a central controller piece. DirectFlow can configure OpenFlow with forwarding entries through CLI or REST API and is used for Traffic Engineering ( TE ) or symmetrical ECMP. Direct Flow is easy to implement as you don’t need a controller. Just use a REST API available in EOS to configure the flows.

 

Optimal Layer 3 Forwarding: Final Points

Optimal Layer 3 forwarding is a critical network architecture component that significantly impacts network performance, scalability, and reliability. Efficiently routing data packets through the best paths enhances network resilience, resource utilization, and security.

Implementing optimal Layer 3 forwarding through routing protocols, QoS mechanisms, and network monitoring ensures a robust and efficient network infrastructure. Embracing this technology allows organizations to deliver seamless connectivity and a superior user experience in today’s increasingly interconnected world.

Summary: Optimal Layer 3 Forwarding

In the ever-evolving networking world, maximizing efficiency and performance is crucial. One key aspect of network optimization is layer 3 forwarding. In this blog post, we explored the concept of optimal layer 3 forwarding and its significance in enhancing network performance.

Section 1: What is Layer 3 Forwarding?

Layer 3 forwarding, also known as IP forwarding, is the process of routing network traffic between different networks or subnets. It involves analyzing the destination IP address of incoming packets and determining the best path for forwarding them to their intended destinations. Layer 3 forwarding operates at the network layer of the OSI model and forms the backbone of modern network communication.

Section 2: The Importance of Optimal Layer 3 Forwarding

Optimal layer 3 forwarding is vital in achieving efficient and reliable network performance. Making intelligent routing decisions ensures that network traffic is efficiently distributed across available paths, minimizing congestion and maximizing throughput. This results in improved network response times and reduced latency, leading to a seamless user experience.

Section 3: Factors Affecting Layer 3 Forwarding Efficiency

Several factors influence the efficiency of layer 3 forwarding. One such factor is the quality and accuracy of routing tables. Outdated or incomplete routing information can lead to suboptimal forwarding decisions and inefficient network utilization. Network topology, link capacities, and traffic patterns also impact layer 3 forwarding efficiency.

Section 4: Techniques for Optimizing Layer 3 Forwarding

To achieve optimal layer 3 forwarding, network administrators can employ various techniques. One common approach uses dynamic routing protocols, such as OSPF or BGP, which continuously exchange routing information and adapt to network changes. Additionally, implementing traffic engineering techniques, such as Quality of Service (QoS) and load balancing, can enhance forwarding efficiency.

Section 5: Case Study: Enhancing Network Performance with Optimal Layer 3 Forwarding

To illustrate the real-world benefits of optimal layer 3 forwarding, let’s consider a case study. A multinational enterprise with geographically dispersed offices experienced network performance issues due to suboptimal routing decisions. Implementing dynamic routing protocols and fine-tuning their routing configurations significantly improved network response times and reduced packet loss.

Conclusion:

Optimal layer 3 forwarding is a cornerstone of network optimization. Intelligently routing traffic enhances network performance, reduces latency, and improves the overall user experience. Network administrators should strive to implement best practices and leverage modern techniques to maximize the efficiency of layer 3 forwarding in their networks.