Many people have asked how to optimize WAN Virtualization. A common questions centers on deciding between a solution centered around either Multiprotocol Label Switching (MPLS) or SD-WAN. Here are a few major differences between the two technologies that will help decide which system is better for certain applications.
Price quotes for MPLS were at the level you know would never be approved; along the lines of $5,300 per month. In a predicament, you ask fellow IT admins for their experiences and advice on alternatives to MPLS that would fit your budget and be effective in terms of reliability and continuity.
The comparison of MPLS vs SD-WAN focuses on four key area: packet loss and availability, Quality of Service (QOS), security, and connectivity.
Even though MPLS networks are a shared infrastructure, they still provide highly reliable packet delivery. MPLS places special labels on each packet that isolates them from other traffic on the network, so even shared systems will not cause interference. This can be extremely useful when there are many customers on a shared network. High traffic times will result in lost packets, packet collisions, and overall degradation of signal. While this technology does help reduce packet loss, it is still operated on a shared network which means there will be competition for bandwidth at times. This congestion can result in the loss of packets.
MPLS networks can see anywhere between 0.1% and 1% packets dropped at any time. Compared this to the internet which can experience 1% or more loss depending on a huge variety of conditions. Service Level Agreements (SLA) will specify the percentage of allowable lost packets from the MPLS provider. Unmanaged networks will not have an SLA and therefore have no guarantees on packet availability.
MPLS providers, not only guarantee a certain level of availability, they will offer QOS buckets so that your most vital data is prioritized over general traffic and has less of a chance of being disrupted. Other internet solutions do not provide any guarantees for packet delivery or QOS, thus are less reliable in terms of having an SLA guarantee.
A common procedure that most customers use to avoid packet loss is to utilize two different links from two unique internet providers. This redundancy is useful overall and allows them to maintain around 99.99% availability and provides the customer with two unique access paths to the internet. While two providers increases availability, it is useful to measure packet loss individually across each provider. If there is a noticeable issue with any particular provider, steps can then be taken to simply and reduce traffic across that link. Forward Error Correction (FEC) is an excellent tool that can help address problems if there are packet loss issues across both providers simultaneously.
FEC is like a form of RAID but used for network packets. As packets are sent out, parity packets are mixed into the data. These parity packets can be compared at their destination and help identify when and where lost data packet occurred. Any lost packets can be identified and resent which improves the overall quality of the connection. These techniques allow QOS to be applied to the most critical packets with FEC providing reliability so that all packets are delivered even across a WAN connection. ...
MPLS or SD-WAN?
You're interested in upgrading the connectivity of your multiple offices. At the time you were using a standard IPsec VPN and a business level broadband connection. One office would soon be upgrading to managed fiber and that PRI would be discontinued at the main headquarters. You had previously used VoIP across these VPN connections without having too many issues.Price quotes for MPLS were at the level you know would never be approved; along the lines of $5,300 per month. In a predicament, you ask fellow IT admins for their experiences and advice on alternatives to MPLS that would fit your budget and be effective in terms of reliability and continuity.
The comparison of MPLS vs SD-WAN focuses on four key area: packet loss and availability, Quality of Service (QOS), security, and connectivity.
Packet Loss and Availability
The biggest advantage of MPLS is its capabilities to deliver packets and provide a high QOS. MPLS excels at managing packet loss which is valuable for those interested in real-time protocols such as VoIP, video conferencing, and virtual desktops. The high packet availability of MPLS allows these real-time systems to work without loss of quality or degradation of signal.Even though MPLS networks are a shared infrastructure, they still provide highly reliable packet delivery. MPLS places special labels on each packet that isolates them from other traffic on the network, so even shared systems will not cause interference. This can be extremely useful when there are many customers on a shared network. High traffic times will result in lost packets, packet collisions, and overall degradation of signal. While this technology does help reduce packet loss, it is still operated on a shared network which means there will be competition for bandwidth at times. This congestion can result in the loss of packets.
MPLS networks can see anywhere between 0.1% and 1% packets dropped at any time. Compared this to the internet which can experience 1% or more loss depending on a huge variety of conditions. Service Level Agreements (SLA) will specify the percentage of allowable lost packets from the MPLS provider. Unmanaged networks will not have an SLA and therefore have no guarantees on packet availability.
MPLS providers, not only guarantee a certain level of availability, they will offer QOS buckets so that your most vital data is prioritized over general traffic and has less of a chance of being disrupted. Other internet solutions do not provide any guarantees for packet delivery or QOS, thus are less reliable in terms of having an SLA guarantee.
SD-WAN Capabilities
Packet delivery across the internet has seen significant improvements due to SD-WAN. Before QOS can even be considered, networks must control packet loss issues. The main idea being that QOS prioritized packets are not effective if they are lost in delivery.A common procedure that most customers use to avoid packet loss is to utilize two different links from two unique internet providers. This redundancy is useful overall and allows them to maintain around 99.99% availability and provides the customer with two unique access paths to the internet. While two providers increases availability, it is useful to measure packet loss individually across each provider. If there is a noticeable issue with any particular provider, steps can then be taken to simply and reduce traffic across that link. Forward Error Correction (FEC) is an excellent tool that can help address problems if there are packet loss issues across both providers simultaneously.
FEC is like a form of RAID but used for network packets. As packets are sent out, parity packets are mixed into the data. These parity packets can be compared at their destination and help identify when and where lost data packet occurred. Any lost packets can be identified and resent which improves the overall quality of the connection. These techniques allow QOS to be applied to the most critical packets with FEC providing reliability so that all packets are delivered even across a WAN connection. ...
