A new breed of network.
AcceleRoute is a data network architecture disclosed in a family of patents offering an evolutionary path for network vendors and operators.
Starting with the well known spine and leaf concept, AcceleRoute adds next generation architectural capabilities that are well suited for scaling data centers and cloud services for decades to come.
AcceleRoute is a patent family for a network design that features a completely bufferless core that can offer extremely low latency and massive scalability in a simple architecture.
AcceleRoute is an architecture for a physical underlay network. SD-WAN software would manage connectivity downward into an AcceleRoute network in an analogous manner to other physical networks such as the Internet, private MPLS networks and 5G.
Yes, AcceleRoute would fit into the Spine and Leaf architecture category however it offers a number of significant advancements over conventional Spine and Leaf implementations that enable much larger and geographically distributed scalability.
The main causes of network latency are buffering and network hops. The AcceleRoute core network is bufferless and fully meshed. All system buffering is at the edge of the network and most network transits are a single hop.
AcceleRoute networks can expand to include tens of thousands of access routers with high throughput of hundreds of petabits/second.
An AcceleRoute core switch element can be designed using conventional ICs typically found in ethernet switches and routers. Once built, this core switching element can be used to build out a large capacity symmetrical bufferless core network.
The AcceleRoute architecture has also been designed to construct a fully optical network that would require the development of an optical switching element.
There are currently 35 patents spanning 9 patent families.
Global time alignment is a key feature of AcceleRoute that allows every Access Node on the network to intercommunicate through a vast lineup of coherent time slots which eliminates the need for buffering in the network core.
The patents for AcceleRoute are owned by its inventor, Dr. Maged Beshai who is represented by InventionShare Inc. for the purposes of commercialization.
A fused network is an architectural method that was invented and patented by network designer, Dr. Maged Beshai as part of the AcceleRoute patent portfolio. It is used to achieve substantially larger scalability of a Spine and Leaf network.
Currently, AcceleRoute is a network design that has been extensively patented and simulated. It has not yet been constructed.
A contiguous network exhibits consistent behavior and architectural symmetry at all points of scale.
A new breed of network.
AcceleRoute is a data network architecture disclosed in a family of patents offering an evolutionary path for network vendors and operators.
Starting with the well known spine and leaf concept, AcceleRoute adds next generation architectural capabilities that are well suited for scaling data centers and cloud services for decades to come.
AcceleRoute is a patent family for a network design that features a completely bufferless core that can offer extremely low latency and massive scalability in a simple architecture.
AcceleRoute is an architecture for a physical underlay network. SD-WAN software would manage connectivity downward into an AcceleRoute network in an analogous manner to other physical networks such as the Internet, private MPLS networks and 5G.
Yes, AcceleRoute would fit into the Spine and Leaf architecture category however it offers a number of significant advancements over conventional Spine and Leaf implementations that enable much larger and geographically distributed scalability.
The main causes of network latency are buffering and network hops. The AcceleRoute core network is bufferless and fully meshed. All system buffering is at the edge of the network and most network transits are a single hop.
AcceleRoute networks can expand to include tens of thousands of access routers with high throughput of hundreds of petabits/second.
An AcceleRoute core switch element can be designed using conventional ICs typically found in ethernet switches and routers. Once built, this core switching element can be used to build out a large capacity symmetrical bufferless core network.
The AcceleRoute architecture has also been designed to construct a fully optical network that would require the development of an optical switching element.
There are currently 35 patents spanning 9 patent families.
Global time alignment is a key feature of AcceleRoute that allows every Access Node on the network to intercommunicate through a vast lineup of coherent time slots which eliminates the need for buffering in the network core.
The patents for AcceleRoute are owned by its inventor, Dr. Maged Beshai who is represented by InventionShare Inc. for the purposes of commercialization.
A fused network is an architectural method that was invented and patented by network designer, Dr. Maged Beshai as part of the AcceleRoute patent portfolio. It is used to achieve substantially larger scalability of a Spine and Leaf network.
Currently, AcceleRoute is a network design that has been extensively patented and simulated. It has not yet been constructed.
A contiguous network exhibits consistent behavior and architectural symmetry at all points of scale.
The current state of networks.
Today’s networks and data centers are faced with applications ravenous for more bandwidth. Everything from unlimited cloud storage and 8K Netflix to quantum computing and Bitcoin mining. Combine that with increasingly faster devices and ultra-fast access over WiFi 6 and 5G cellular and maintaining expected quality of experience becomes a huge challenge. How can we design networks to continuously scale at this rate.
Extensively used today are buffered core networks that require multiple hops for data to reach a destination. Consequently, these networks are prone to internal congestion which leads to highly variable latency and constrained throughput.
More recently, spine and leaf networks have gained increasing attention. In their simplest form, these networks offer good performance, however their scalability is limited by the number of ports and capacity of the spine switches.
The current state of networks.
Today’s networks and data centers are faced with applications ravenous for more bandwidth. Everything from unlimited cloud storage and 8K Netflix to quantum computing and Bitcoin mining. Combine that with increasingly faster devices and ultra-fast access over WiFi 6 and 5G cellular and maintaining expected quality of experience becomes a huge challenge. How can we design networks to continuously scale at this rate.
Extensively used today are buffered core networks that require multiple hops for data to reach a destination. Consequently, these networks are prone to internal congestion which leads to highly variable latency and constrained throughput.
More recently, spine and leaf networks have gained increasing attention. In their simplest form, these networks offer good performance, however their scalability is limited by the number of ports and capacity of the spine switches.
Next generation spine and leaf.
AcceleRoute uses a fundamentally different network topology to expand the capacity of spine and leaf networks. Unlike conventional networks that scale “vertically” by adding increasingly larger spine switches and additional spine layers, AcceleRoute scales “horizontally” using lateral connections to interconnect spines. This patented method, called Network Fusion, enables spine and leaf networks to be architecturally joined to form a seamless spine fabric. This “Lego” approach to expanding networks has a range of operational advantages:
- Entire networks can be constructed using relatively small spine switches that make it easy and efficient to stock spares for construction and repair with minimal obsolescence.
- Smooth expansion is incremental and virtually unlimited with no service disruption.
- Every dollar invested buys gear that can stay in the network a long time and can be easily repurposed.
- Zero buffering in the core network delivers near-zero latency and high stability under high traffic fluctuations.
- Evolvable to optical switching for significant power and footprint reduction.
Next generation spine and leaf.
AcceleRoute uses a fundamentally different network topology to expand the capacity of spine and leaf networks. Unlike conventional networks that scale “vertically” by adding increasingly larger spine switches and additional spine layers, AcceleRoute scales “horizontally” using lateral connections to interconnect spines. This patented method, called Network Fusion, enables spine and leaf networks to be architecturally joined to form a seamless spine fabric. This “Lego” approach to expanding networks has a range of operational advantages:
- Entire networks can be constructed using relatively small spine switches that make it easy and efficient to stock spares for construction and repair with minimal obsolescence.
- Smooth expansion is incremental and virtually unlimited with no service disruption.
- Every dollar invested buys gear that can stay in the network a long time and can be easily repurposed.
- Zero buffering in the core network delivers near-zero latency and high stability under high traffic fluctuations.
- Evolvable to optical switching for significant power and footprint reduction.
AcceleRoute Patent Family Overview
The AcceleRoute portfolio of 34 patents has been independently assessed by IP.com as offering high monetization potential and technology value. This portfolio is a must-have IP asset for networking companies.
First family
Scalable network architecture based on bufferless switches
Second family
Time-coherent global networks
Third family
High capacity network of multiple switch planes
Fourth family
Large-scale data center/geographically distributed network
Fifth family
Electronic switching nodes, stand-alone or modules of a large-scale data center
Sixth family
Packet switching based on contention-free switch modules and distributed control systems
Seventh family
A contiguous switching system for cloud communications / large-scale data centers
Eighth family
Optical spectral-temporal connector interconnecting, a large number of nodes in a full-mesh structure
Ninth family
Distributed control of a vast communication network
