Method and apparatus for allocating processing capacity of system processing units in an extranet gateway
View Patent ↗The expected processing bandwidth of System Processing Units (SPUs) in an extranet gateway is first assessed, for example by multiplying a CPU processor speed by a conversion factor or determining the type of accelerator, and then adjusting the capacity to take into account other processing requirements on the SPU. Once assessed, a load distribution system assigns tunnels to the SPUs based on their processing capacity. As tunnels are assigned and torn down, the estimated available processing capacities of the SPUs are adjusted to account for the assignment or release of the system processing resources attributable to those VPN services. Tunnel assignment may take into account the ability of a SPU to handle the type of tunnel, the absolute or relative capacity of the SPUs, the type of tunnel being assigned, features of the tunnel such as the presence of compression, and efficiencies of the SPU at handling that tunnel type with selected features.
1. A method of allocating processing capacity of system processing units in an extranet gateway, the method comprising the steps of:
establishing a first initial expected available processing bandwidth of a first of the system processing units, the first expected available processing bandwidth representing a first amount of VPN tunnel bandwidth which the first of the system processing units is expected to be able to handle;
establishing a second initial expected available processing bandwidth of a second of the system processing units, the second expected available processing bandwidth representing a second amount of VPN tunnel bandwidth the second of the system processing units is expected to be able to handle; and
assigning a Virtual Private Network (VPN) tunnel to one of the first and second system processing units for processing according to estimated current available bandwidths of the first and second system processing units, the estimated current available bandwidths being estimated by assessing the initial expected available bandwidths for each system processing unit and decrementing the initial expected available bandwidth of each system processing unit by other processing requirements assigned to that respective system processing unit;
wherein the step of assigning the VPN tunnel to one of the first and second system processing units looks to assign the VPN tunnel to the system processing unit with the highest estimated current available bandwidth, the highest current available bandwidth being based on an absolute bandwidth capacity basis, the absolute bandwidth being calculated by determining which system processing unit has a largest amount of estimated current available bandwidth;
wherein the first initial expected available bandwidth is established by determining whether the first system processing unit is an accelerator or a central processing unit (CPU); and
wherein if the first system processing unit is a CPU, the step of establishing the first initial expected available bandwidth comprises determining a type of CPU and CPU speed, obtaining a first conversion factor for the type of CPU, and multiplying the conversion factor with the CPU speed.
2. The method of claim 1 , wherein if the first system processing unit is an accelerator, the step of establishing the first initial expected available bandwidth comprises determining a type of accelerator and obtaining expected available bandwidth information for that type of accelerator from an initial expected bandwidth table.
3. The method of claim 1 , wherein the first conversion factor is based on an amount of bandwidth passable by that processor type per unit CPU speed.
4. The method of claim 1 , wherein the other processing requirements of a system processing unit comprise:
overhead processing requirements assigned to that system processing unit; and
processing requirements associated with other VPN tunnels already assigned to that system processing unit.
5. The method of claim 4 , wherein the processing requirements associated with other VPN tunnels assigned to that system processing unit comprise encryption and de-encryption processing requirements for the other VPN tunnels.
6. The method of claim 4 , wherein the other processing requirements of a system processing unit further comprise processing requirements associated with other VPN tunnels assigned to other system processing units.
7. The method of claim 4 , wherein the actual load on the other VPN tunnels assigned to the system processing unit is not monitored or used in connection with estimating the current available bandwidth of the system processing unit.
8. A method of allocating processing capacity of system processing units in an extranet gateway, the method comprising the steps of:
establishing a first initial expected available processing bandwidth of a first of the system processing units, the first expected available processing bandwidth representing a first amount of VPN tunnel bandwidth which the first of the system processing units is expected to be able to handle;
establishing a second initial expected available processing bandwidth of a second of the system processing units, the second expected available processing bandwidth representing a second amount of VPN tunnel bandwidth the second of the system processing units is expected to be able to handle; and
assigning a Virtual Private Network (VPN) tunnel to one of the first and second system processing units for processing by assessing current available bandwidths of the first and second system processing units, the current available bandwidths being determined by assessing the initial expected available bandwidth for that system processing unit as decremented by other processing requirements for that system processing unit;
wherein the step of assigning the VPN tunnel to one of the first and second system processing units looks to assign the VPN tunnel to the system processing unit with the highest estimated current available bandwidth, the highest current available bandwidth being based on a relative bandwidth capacity basis, the relative bandwidth being calculated by determining which system processing unit has a highest percentage of available capacity;
wherein the first initial expected available bandwidth is established by determining whether the first system processing unit is an accelerator or a central processing unit (CPU); and
wherein if the first system processing unit is a CPU, the step of establishing the first initial expected available bandwidth comprises determining a type of CPU and CPU speed, obtaining a first conversion factor for the type of CPU, and multiplying the conversion factor with the CPU speed.
9. The method of claim 8 , wherein if the first system processing unit is an accelerator, the step of establishing the first initial expected available bandwidth comprises determining a type of accelerator and obtaining expected available bandwidth information for that type of accelerator from an initial expected bandwidth table.
10. The method of claim 8 , wherein the first conversion factor is based on an amount of bandwidth passable by that processor type per unit CPU speed.
11. The method of claim 8 , wherein the other processing requirements of a system processing unit comprise:
overhead processing requirements assigned to that system processing unit; and
processing requirements associated with other VPN tunnels already assigned to that system processing unit.
12. The method of claim 11 , wherein the processing requirements associated with other VPN tunnels assigned to that system processing unit comprise encryption and de-encryption processing requirements for the other VPN tunnels.
13. The method of claim 11 , wherein the other processing requirements of a system processing unit further comprise processing requirements associated with other VPN tunnels assigned to other system processing units.
14. The method of claim 11 , wherein the actual load on the other VPN tunnels assigned to the system processing unit is not monitored or used in connection with estimating the current available bandwidth of the system processing unit.