Efficient data transfer between servers and remote peripherals

Abstract

Methods and apparatus are provided for transferring data between servers and a remote entity having multiple peripherals. Multiple servers are connected to a remote entity over an Remote Direct Memory Access capable network. The remote entity includes peripherals such as network interface cards (NICs) and host bus adapters (HBAs). Server descriptor rings and descriptors are provided to allow efficient and effective communication between the servers and the remote entity.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates data transfer between servers and remote entities including remote peripherals. In one example, the present invention relates to methods and apparatus for efficiently transferring data between servers and remote peripherals such as network interface cards (NICs), host bus adapters (HBAs), and cryptographic accelerators.

2. Description of Related Art

A server or computing system generally includes one or more processors, memory, and peripheral components and peripheral interfaces. Examples of peripheral components include cryptographic accelerators, graphics accelerators, and eXtensible Markup Language (XML) accelerators. Examples of peripheral interfaces include network interface cards (NIC), serial ATA (SATA) adapters, and host bus adapters (HBA). Processors, memory, and peripherals are often connected using one or more buses and bus bridges. To provide fault-tolerance, individual servers are often configured with redundant resources.

Because resources such as peripheral components and peripheral interfaces are assigned on a per server or a per processor basis, other servers do not typically have access to these resources. In order to provide adequate resources for each server, resources are typically over-provisioned. For example, more hardware acceleration is provided than is typically needed. More network interface bandwidth is allocated than is typically used simply to handle worst-case or expected worst-case scenarios. Resources are over-provisioned resulting in overall waste and low utilization. Resource assignment on a per server or a per processor basis also limits the ability to reconstruct or reconfigure a resource environment.

A variety of peripherals are sometimes provided remotely over a network. However, protocols for communications between servers and remote peripherals are often inefficient, inadequate, or even unavailable. Consequently, the techniques and mechanisms of the present invention allow efficient data transfer between servers and remote peripherals over a networking technology.

SUMMARY OF THE INVENTION

Methods and apparatus are provided for transferring data between servers and a remote entity having multiple peripherals. Multiple servers are connected to a remote entity over an Remote Direct Memory Access capable network. The remote entity includes peripherals such as network interface cards (NICs) and host bus adapters (HBAs). Server descriptor rings and descriptors are provided to allow efficient and effective communication between the servers and the remote entity.

In one embodiment a technique for transferring data from a server to a remote entity is provided. A write index is sent to virtualization logic associated with the remote entity. The remote entity includes multiple peripheral interfaces. The write index is provided from the server to virtualization logic over an RDMA capable networking technology. One or more server descriptor entries are provided to virtualization logic. An RDMA read data buffer request is received from virtualization logic. Data is transferred in response to the RDMA read data buffer request.

In another example, a technique for transferring data from a remote peripheral to a server is provided. A write index is sent to virtualization logic associated with a remote entity including multiple peripherals. The write index is provided from the server to virtualization logic over an RDMA capable networking technology. A server descriptor entry referenced by the write index is provided to virtualization logic. An RDMA write response buffer is received from virtualization logic. A completion message is received from virtualization logic.

A further understanding of the nature and advantages of the present invention may be realized by reference to the remaining portions of the specification and the drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention may best be understood by reference to the following description taken in conjunction with the accompanying drawings, which are illustrative of specific embodiments of the present invention.

FIG. 1 is a diagrammatic representation showing individual servers.

FIG. 2 is a diagrammatic representation showing individual servers connected to remote peripherals.

FIG. 3 is a diagrammatic representation depicting server descriptor rings.

FIG. 4 is an exchange diagram showing data transmission from a server.

FIG. 5 is an exchange diagram showing data transmission to a server.

FIG. 6 is a diagrammatic representation showing a command descriptor.

FIG. 7 is a diagrammatic representation showing an initialization descriptor.

FIG. 8 is a diagrammatic representation depicting a completion descriptor.

FIG. 9 is a diagrammatic representation showing a remote entity.

FIG. 10 is a diagrammatic representation showing a remote entity having virtualization logic and resources.

DETAILED DESCRIPTION OF SPECIFIC EMBODIMENTS

Reference will now be made in detail to some specific examples of the invention including the best modes contemplated by the inventors for carrying out the invention. Examples of these specific embodiments are illustrated in the accompanying drawings. While the invention is described in conjunction with these specific embodiments, it will be understood that it is not intended to limit the invention to the described embodiments. On the contrary, it is intended to cover alternatives, modifications, and equivalents as may be included within the spirit and scope of the invention as defined by the appended claims.

For example, the techniques of the present invention will be described in the context of RDMA and Infiniband as one example of an RDMA capable network. However, it should be noted that the techniques of the present invention can be applied to a variety of different standards and variations to RDMA and Infiniband. Similarly, a server is described throughout. However, a server can be a single processing system, a multiple processor system, a guest operating system, a system image, or a virtual machine. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The present invention may be practiced without some or all of these specific details. In other instances, well known process operations have not been described in detail in order not to unnecessarily obscure the present invention.

Furthermore, techniques and mechanisms of the present invention will sometimes be described in singular form for clarity. However, it should be noted that some embodiments include multiple iterations of a technique or multiple instantiations of a mechanism unless noted otherwise. For example, a processor is used in a variety of contexts. However, it will be appreciated that multiple processors can also be used while remaining within the scope of the present invention unless otherwise noted.

Servers having processors and memory typically have access to physical peripherals, including peripheral components and peripheral interfaces. However, physically tying resources to particular servers can lead to a number of drawbacks, including management and usage inefficiencies. In some examples, peripherals are moved from individual servers onto a remote device. In some embodiments, resources such as hardware accelerators, NICs, and HBAs are removed from individual servers and aggregated at a remote entity. The remote entity can be configured to include multiple peripherals and virtualization logic allowing server access to the peripherals.

In some examples, the remote entity is an appliance or switch that is connected to multiple servers over a network. However, offloading resources onto a remote entity requires that the servers be able to access the resources in an efficient manner. Typical networks do not allow efficient access to the resources in a manner providing low-latency, high-throughput, processor efficiency, support for quality of service (QoS). Consequently, the techniques and mechanisms of the present invention contemplate providing efficient access to remote peripherals including remote storage adapters, remote network adapters, and remote hardware accelerators. In one example, the remote entity receives requests from individual servers over an remote direct memory access (RDMA) capable network such as Infiniband and provides resources to handle individual requests.

Examples of RDMA capable networks include Infiniband and RDMA over Ethernet. The RDMA capable network provides a low latency, high bandwidth connection between the multiple servers and the resources in the remote entity. The technique takes into account the fact that the RDMA-capable network adapter within each server may have different performance characteristics than the specialized device that resides on the I/O entity. According to various embodiments, a higher processing burden is placed on the target device within the remote entity rather than on the RDMA network adapter within each server. This improves performance and reduces the overhead imposed on the computer and its RDMA network adapter.

According to various embodiments, an RDMA network adapter on a server has a large set of capabilities, but is limited because of the need to support a variety of protocols. The RDMA adapters on the remote entity, however, can be configured to be much more specialized. Each remote entity RDMA adapter does not need to support all host functionality. In some examples, each remote entity RDMA adapter only needs to support a particular data transfer protocol. However, the remote entity RDMA adapter needs to support high bandwidth and low latency, as a remote entity RDMA adapter may be configured to support many server adapters at the same time. According to various embodiments, the techniques of the present invention place processing burden on the remote entity adapter and not on the server. RDMAs are issued by the remote entity adapter and not by the server adapter.

Various embodiments of the present invention provide a unified infrastructure for supporting storage, network, and other types of remote input/output (I/O) traffic. Message coalescing is used to improve performance by reducing the number of messages. Support for QoS is maintained and the techniques are suitable for either hardware or software implementation. According to various embodiments, the remote entity allows dynamic reconfiguration and remapping of virtual resources to servers to support true stateless computing and utility computing and allows charge back by application, user, departments, etc. Storage resources including local disks can be shared and virtualized to allow stateless computing.

Requests from multiple servers may also be forwarded to the same NIC. An administrator can provision and partition resources including server resources, peripheral component resources, and peripheral interface resources at the remote entity based on particular needs and requirements. According to various embodiments, a remote entity can also perform packet inspection or deep packet inspection to classify server traffic. Traffic policies can be applied to allow application level quality of service. Quality of service (QOS) and traffic engineering schemes can be implemented. Traffic associated with particular devices or servers can be given priority or guaranteed bandwidth. The total amount of resources can be decreased while increasing resource utilization. The remote entity can be introduced into existing server racks and farms with little disruption to system operation.

FIG. 1 is a diagrammatic representation showing a typical system for connecting servers to an IP network. Server 101 includes a processor 103, memory 105, a cryptographic accelerator 107, and NIC 109. The processor 103 communicates with other components and interfaces in the system using a network and associated I/O controllers. One or more bus bridges may also be used. In typical implementations, communications between components and interfaces in server 101 occur over a network such as Infiniband. Server 111 includes processor 113, cryptographic accelerator 117, memory 115, and NIC 119. Communication within server 111 similarly occurs over one or more I/O buses. Server 121 includes a processor 123, memory 125, HBA 127 and NIC 129. In order to allow communication with an IP network through a local area network and through network router 141, NICs 109, 119, and 129 are provided. To allow communication with a fibre channel storage area network through fibre channel switch 143, an HBA 127 is provided. In one example, a processor 103 passes data to a cryptographic accelerator 117 to encrypt data prior to transmitting the data onto the local area network 161. Similarly, data received from a NIC 109 is passed to a cryptographic accelerator 117 for decryption when data is received by the processor 103.

The various NICs 107, 119, and 129 are also associated with IP addresses and media access control (MAC) addresses. Each server and associated NIC encapsulates data into IP packets for transmission to a network router 141. Encapsulation may involve adding appropriate Telnet Control Protocol (TCP) and IP headers and addresses. Each NIC is also configured to remove TCP/IP headers and addresses and provide data to an associated processor over a system bus when IP packets are received from an IP network. An HBA 127 may be configured to perform fabric login (FLOGI) and port login (PLOGI) processes associated with operating in a fibre channel storage area network. Each server 101, 111, and 121 may be individual computers, UNIX based low-end to high-end servers, mainframes, or blade servers. Any system including a processor and memory and an address space is referred to herein as a server. A server may include components noted above as well as a variety of other components.

In some examples, a series of servers each include a processor and a number of resources such as HBAs, NICs, iSCSI TOE, SATA as well as other peripherals interfaces and peripheral components. Cables and lines are then connected to each server to allow access to Ethernet switches, Fibre Channel switches, appliances, and mainframe or KVM switches. In some examples, redundant cables are used to connect each server to each switch. In an example with 40 servers, 80 cables would be used to connect the servers to an Ethernet switch. An additional 80 cables would be used to connect the servers to Fibre Channel switch, etc. The resource allocation and system management inefficiencies are magnified by the physical complexities of routing the individual cables.

FIG. 2 is a diagrammatic representation showing separate servers connected to a remote entity. Server 201 includes processor 203 and memory 205. Server 211 includes processors 213 and 217 and memory 215. Server 221 includes only processor 223 and memory 225. Components and peripherals in each server 201, 211, and 221 are connected using one or more I/O buses. It should be noted that a server can be a virtual server such as a guest operating system, an application, or a virtual machine. According to various embodiments, an RDMA capable network is used to allow access to a remote entity 243. Each server is configured to communicate with a remote entity 243 using virtualization logic 241.

Virtualization logic 241 may be a standalone entity, integrated within a particular server, or provided with a remote entity 243. According to various embodiments, components such as HBA 253, NIC 255, a cryptographic accelerator 257, an XML accelerator 259, Internet SCSI (iSCSI)/TCP Offload Engine (TOE)/Serial ATA (SATA) 251, can be offloaded from servers 201, 211, and 221 onto a remote entity 243. The resources including HBA 253, NIC 255, cryptographic accelerator 257, XML accelerator 259, and iSCSI/TOE/SATA 251 are maintained in a shared and virtualized manner on a remote entity 243. Links are provided between the remote entity and external switches/routers such as an IP network switch/router.

According to various embodiments, a series of servers is connected to the remote entity using an Infiniband network. Infiniband allows interconnection of processing nodes and I/O nodes. The Infiniband stack includes support of multiple queues including send and receive queues. In one example, the queues used are descriptor rings. Descriptors are used to hold instructions, commands, and data addresses. For example, descriptors hold instructions to transmit data and receive data.

By using a remote entity, the number of resources and links can be significantly reduced while increasing allocation efficiency. Resources can be accessed at the driver level. Virtual resources including service engines can be dynamically bound to servers including system images, guest operating systems, and virtual machines. In one instance, a cryptographic accelerator is bound to a particular system image.

In this example, to perform cryptographic operations associated with sending or receiving data, a server 201 passes data to the remote entity cryptographic accelerator 257 as though the remote entity cryptographic accelerator 257 were included in the server 201. To perform XML operations, a server 201 passes data to the XML accelerator 259 as though the XML accelerator 259 is included in the server 201. Alternatively, XML and cryptographic operations can be performed as data is received at a remote entity from a network. Consequently, additional server 201 processor involvement is not needed.

FIG. 3 is a diagrammatic representation showing descriptor rings that can be used for communication between servers and a remote entity over an Infiniband network. According to various embodiments, each server may include one or more descriptor rings. Each server descriptor 303, 313, and 323 is associated with a write index and a read index. Server descriptor 303 has a write index 331 and a read index 333. The write index 331 is modified by the server and is associated with the descriptor the server 303 has last written. The read index 333 is modified by virtualization logic and associated peripheral components and peripheral interfaces. The read index 333 is associated with the descriptor the virtualization logic has last read.

Similarly, descriptor ring 313 has write index 341 and read index 343 and descriptor ring 323 has read index 353 and write index 351. In one embodiment, the server descriptor rings 303, 313, and 323 are associated with different servers. Each server descriptor ring may also have an associated ring size value and a base address value.

FIG. 4 is an exchange diagram showing one technique for transmitting data from a server to a remote entity having multiple resources to allow access to a peripheral such as a port adapter. In one example, a server may be transmitting data onto a network through a NIC provided at a remote entity. A server 401 communicates with virtualization logic 403 associated with remote peripherals. The server 401 sends a write index 411 to the virtualization logic 403. In response to the write index, the virtualization logic 403 sends an RDMA read to the server 401. The server 401 provides one or more descriptor entries to the virtualization logic 403. A wide variety of descriptor formats are available. In many embodiments, a descriptor includes a pointer or memory address. The number of bytes in the buffer or memory may also be provided. A start of memory, middle of memory, or end of memory indicator can also provided in the event that the data block is fragmented.

The virtualization logic 403 sends an optional RDMA write to provide a read index 417 to the server 401. The virtualization logic 403 processes the descriptor information and sends an RDMA read data buffer request 421 to the server 401. One or more address ranges may be read. The server provides the data 423 to the virtualization logic 403. When the data has been received the virtualization logic 403 responds with a completion message 425.

FIG. 5 is an exchange diagram showing one technique for receiving data from a remote entity having multiple resources to allow a server to access a peripheral such as a port adapter. In one example, a server may be receiving network data received by a NIC at a remote entity. A server 501 communicates with virtualization logic 503 associated with remote peripherals. The server 501 sends a write index 511 to the virtualization logic 503. In response to the write index, the virtualization logic 503 sends an RDMA read to the server 501. The server 501 provides one or more descriptor entries to the virtualization logic 503. A wide variety of descriptor formats are available. In many embodiments, a descriptor includes a pointer or memory address. The number of bytes in the buffer or memory may also be provided. A start of memory, middle of memory, or end of memory indicator can also provided in the event that the data block is fragmented.

The virtualization logic 503 sends an optional RDMA write to provide a read index 517 to the server 501. The virtualization logic 503 processes the descriptor information and sends an RDMA write to the server with write response data 521. According to various embodiments, data is written directly into server memory at addresses specified by descriptor entries. One or more address ranges may be written. The virtualization logic 503 sends a completion message 523 when the data transfer is complete.

FIGS. 6-8 are diagrammatic representations showing examples of descriptor formats used for transmission between servers and virtualization logic. Although common fields and parameters for descriptors are shown, it should be noted that a wide range of fields and parameters can be used. According to various embodiments, descriptors are created during transfer of data between server and virtualization logic. The virtualization logic allows servers to access peripherals or resources included in a remote entity.

FIG. 6 is a diagrammatic representation showing a command descriptor. According to various embodiments, the command descriptor is used to provide an address associated with a buffer read or write. The command descriptor 601 includes a command handle 603. In one example, the command handle is 16 bits. A length 605 and a buffer key 607 are also provided. Portions of the buffer address are provided in field 611 and 613. Bits 63-32 are provided in buffer address hi field 611 and bits 31-0 are provided in buffer address lo field 613.

FIG. 7 is a diagrammatic representation showing an initialization descriptor. The initialization descriptor is used to provide base buffer addresses, ring size, and index addresses. According to various embodiments, the initialization descriptor includes an 8-bit type field, an 8-bit entry size 705, and a 16-bit descriptor ring size 707. The 32-bit fields include a read index address key 711, a base address key 713, read index address hi bits 715, read index address low bits 717, base address hit bits 719, and base address low bits 721. A reserved field 709 can be maintained for additional parameters.

FIG. 8 is a diagrammatic representation showing a completion descriptor. The completion descriptor includes a type field 801, a read index 803, a completion count 805, a reserved field 807, and multiple completion entries 809 and 811. A completion entry includes a command handle 821, packet length 823, a status field 825, and a reserved field 827. Although particular bit lengths are described above, it should be noted that a wide variation of bit lengths can be used. Additional descriptors or a subset of descriptors may also be used.

FIG. 9 is a diagrammatic representation showing one example of a remote entity. Fabric interface 921 is connected to multiple computer systems using a network such as Infiniband. Port adapters 941-945 are associated with multiple resources such as HBAs, SATAs, hardware accelerators, etc. According to various embodiments, a VNIC chip or VN chip 951 has integrated virtualization logic and port adapters in network interfaces 981-987. A VSSL card 961 also has integrated virtualization logic included in VSSL chip 967 as well as an SSL accelerator 969. Although only a VNIC and a VSSL are shown, it should be recognized that a variety of other virtualization mechanisms such as VHBAs and VXMLs can be included. The network interfaces 981-987 may be MAC interfaces associated with multiple gigabyte ports. According to various embodiments, network interfaces 981-987 include logic mechanisms conventionally found in a NIC. The server platform 911 manages interaction between the servers connected to the fabric interface 921 and various resources associated with the port adapters 941-945 and network interfaces 981-987. In one example, port adapters are stand alone entities such as HBA cards while network interfaces are interface with a VN chip 951.

The platform 911 is associated with memory 919 and a processor subsystem 913, a power subsystem 915, and a storage subsystem 917. In some embodiments, the platform 911 includes tables with information mapping various servers connected through the fabric interface 921 and various port adapter resources and network interfaces. The processor subsystem 913 is configured to manage port adapter resources as though the port adapters and network interfaces 981-887 were included in individual servers. In one example, the processor subsystem 913 is configured to initialize an IP network connection regardless of whether servers have been connected to the server platform 911.

According to various embodiments, a VSSL card 961 is coupled to translation logic 923. In some embodiments, a VSSL card 961 is separate from a VNIC or VN chip or card 951. The VSSL card includes a VSSL chip 967. In some embodiments, the VSSL chip 967 polls transmit and receive queues associated with various servers. When it sees a packet in a server transmit queue, it pulls the packet and forwards associated commands to an SSL chip 969. SSL chip 969 can be one of a number of available SSL accelerators, such as the Nitrox Chip available from Cavium Networks of Mountain View, Calif. It should be noted that although a VSSL card 961 and a VN chip 951 are shown, virtualization logic and accelerator functionality can also be implemented in a single chip. Alternatively, a VN chip 951 may be separated out into multiple components and implemented as a card. In still other implementations, all virtualization logic for the VN chip 951, the VSSL card 961, and any other components is implemented in a single device. In one embodiment, an SSL chip 969 includes general purpose processor cores and specialized cryptography cores included on a single ASIC. The cryptography cores may include SSL cores, IP Security (IPSec) cores, SSL-VPN cores, XML/web services cryptography cores, storage and fibre channel data cryptography cores, and email data cryptography cores.

IOP 963 is configured to perform context management for SSL chip 969. According to various embodiments, an SSL agent runs on IOP 963. The SSL agent manages and allocates key memory and context memory. A context is used for each SSL connection. A context is allocated before a new SSL connection is established. According to various embodiments, context can be established in several manners.

A server can send a context allocation request to the SSL transmit queue. The VSSL chip 967 pulls this request and notifies the SSL agent running on IOP 963. The SSL agent allocates a new context and passes an index back to the VSSL chip 967. The VSSL chip 967 sends the response back to the server's receive queue. An application will now have a context allocated for it and the context can be passed in for all SSL requests.

Alternatively, an SSL agent could allocate a pool of contexts per server. Whenever an application needs a new context, it could get the context locally from the VSSL driver (which keeps a cache of the contexts). Once the VSSL driver's cache runs out of contexts, it sends a request to the SSL Agent to obtain additional contexts.

According to various embodiments, a VNIC chip or VN chip 951 is also coupled to the servers. In one example, the VN chip 951 is connected to the fabric interface 921 through optional translation logic 923. The VN chip 951 also has a connection with the processor subsystem 913 and a series of network interfaces 981-987 connecting the VN chip 951 to external network entities. In other examples, the VN chip may not include NIC interfaces and instead may be connected to conventional NICs.

The VNIC chip includes classifier logic 947, a queue manager 945, and a buffer manager 943. According to various embodiments, the classifier logic 947 includes parse and lookup logic configured to identify information such as a packet destination server and priority. Classifier logic can also be used to filter incoming data or apply traffic policing policies. In some instances, classifier logic can be used to block packets in order to implement a firewall. In one embodiment, classifier logic 947 parses a packet and uses the information in the packet to identify entries in lookup tables. The data is then buffered. Buffer manager 943 manages data in memory associated with the VN chip 951. Queue manager 945 manages descriptors for data posted. A descriptor can include a reference to a memory location, a length, a source port, and a multicast count, as well as other parameters.

In one example, classifier logic 947 determines that the packet received is a high priority packet and should be placed in a high priority queue by the buffer manager 943. Parameters provided may include a pointer, a length, a source port, a multicast count, and a queue identifier. The data is then placed into memory and information referencing the data such as a pointer and a length is posted into a buffer ring or a descriptor ring. When a connected server successfully arbitrates for bus access, the server reads the buffer ring or descriptor ring and obtains the data from memory associated with the VN chip. According to various embodiments, the server reads the data directly into its own memory.

Each individual server may also include descriptor queues. As will be appreciated, the servers connected to the network switch and the remote entity arbitrate for access to the I/O Bus. When access is obtained, data can be read from memory associated with one of the server based on the information provided in the descriptor queues.

Redundancy mechanisms are also provided to allow continued operation in the event that a NIC or other resource fails or a remote entity itself fails. Redundancy mechanisms can be managed by a VNIC device or VN chip, a remote entity, or by the individual servers themselves.

FIG. 10 is a diagrammatic representation depicting an embodiment of a remote entity using virtual Direct Memory Access (DMA) engines. According to various embodiments, a virtual DMA engine and a particular resource is assigned to each application, guest operating system, system image, virtual server, or physical server to provide true hardware acceleration on demand.

A fabric interface 1021 is connected to multiple computer systems using a network such as Infiniband and translation logic 1023. Virtualization logic 1061 obtains data such as descriptor data from individual servers and provides the data to hardware resources 1071. In some examples, data transfer is performed using virtualized Direct Memory Access (DMA) mechanisms to allow minimal processor involvement during data transfers. In some embodiments, a descriptor is provided on a descriptor queue. The descriptor includes addresses, lengths, and other data parameters. A virtualized DMA engine 1063 in virtualization logic 1061 reads the descriptor and directly transfers the block of data in memory to appropriate hardware resources 1071.

In addition, although exemplary techniques and devices are described, the above-described embodiments may be implemented in a variety of manners, media, and mechanisms. For instance, instructions and data for implementing the above-described invention may be stored on a disk drive, a hard drive, a floppy disk, a server computer, or a remotely networked computer. Hardware used to implement various techniques may be embodied as racks, cards, integrated circuited devices, or portions of semiconductor chips. Accordingly, the present embodiments are to be considered as illustrative and not restrictive, and the invention is not to be limited to the details given herein, but may be modified within the scope and equivalents of the appended claims.

Claims

1. A method for transferring data between a remote entity and a plurality of servers, the method comprising: receiving, by the remote entity from a first server of the plurality of servers, a first request for access to a first peripheral of a plurality of peripherals, the remote entity comprising the plurality of peripherals, the receiving the first request comprising: receiving a first write index, of a first server descriptor ring associated with the first server, using a Remote Direct Memory Access capable network;receiving, by the remote entity from a second server of the plurality of servers, a second request for access to the first peripheral, the receiving the second request comprising: receiving a second write index, of a second server descriptor ring associated with the second server, using the remote direct memory access capable network;wherein a different server descriptor ring is associated with each server of the plurality of servers, each different server descriptor ring comprising a respective write index, a respective read index, respective instructions, and respective commands for the respective server associated with the server respective different descriptor ring;transmitting to the first server a first read data buffer request from virtualization logic, wherein the first read data buffer request is sent from the remote entity to the server in response to the remote entity receiving the first write index of the first server descriptor ring from the first server;transmitting to the second server a second read data buffer request from the virtualization logic, wherein the second read data buffer request is sent from the remote entity to the second server in response to the remote entity receiving the second write index of the second server descriptor ring from the second server;enforcing traffic policing policies, by the remote entity with the virtualization logic associated with the remote entity, to prioritize access to each of the plurality of peripherals using a virtual device driver for the peripheral, wherein the virtualization logic allows each of the plurality of peripherals to be shared by the plurality of servers based at least in part on the different server descriptor rings, wherein the plurality of peripherals comprise peripheral interfaces, peripheral components, and one or more host bus adapters, one or more network interface cards, and at least one of a cryptographic accelerator or an eXtensible Markup Language accelerator, and wherein the access to each of the plurality of peripherals comprises requests from the first server and the second server forwarded to one network interface card of the one or more network interface cards, the remote entity: receiving server traffic comprising: receiving data transferred from the first server to the remote entity in response to the first read data buffer request;receiving data transferred from the second server to the remote entity in response to the second read data buffer request;performing packet inspection to classify at least some of the server traffic with classifier logic, the performing packet inspection comprising parsing data packets of the server traffic and utilizing lookup logic to determine priorities attributed to the data packets;placing the data packets into at least one prioritized queue based at least in part on the priorities determined;placing selected data into memory associated with the first peripheral; andposting a reference to the selected data, and facilitating reading of the selected data associated with the first peripheral by the first server or the second server to obtain the selected data from the memory associated with the first peripheral.

2. The method of claim 1, further comprising, after receiving the first write index of the first server descriptor ring and before transmitting the first read data buffer request: sending a remote direct memory access read descriptor request from the virtualization logic;receiving one or more descriptor entries from the first server descriptor ring associated with the first server, the one or more descriptor entries including a pointer and a number of bytes; andsending a remote direct memory access write providing a read index, wherein transferring the data from the first server to the remote entity is based on the read index.

3. The method of claim 2, further comprising transmitting to the first server a completion message from the virtualization logic.

4. The method of claim 1, wherein the remote direct memory access capable networking technology is InfiniBand.

5. The method of claim 1, wherein the remote direct memory access capable networking technology is remote direct memory access over Ethernet.

6. The method of claim 1, wherein the remote direct memory access capable networking protocol is iSER.

7. The method of claim 1, wherein the remote direct memory access capable networking protocol is SRP.

8. A system for transferring data between a remote entity and a plurality of servers, the system comprising: a remote entity comprising a plurality of peripherals, the plurality of peripherals comprising: peripheral interfaces, peripheral components, one or more host bus adapters, one or more network interface cards, and at least one of a cryptographic accelerator or an eXtensible Markup Language accelerator;the remote entity to: receive, from a first server of the plurality of servers, a first request for access to a first peripheral of a plurality of peripherals, the receiving the first request comprising: receiving a first write index, of a first server descriptor ring associated with the first server, using a Remote Direct Memory Access capable network;receive, from a second server of the plurality of servers, a second request for access to the first peripheral, the receiving the second request comprising: receiving a second write index, of a second server descriptor ring associated with the second server, using the Remote Direct Memory Access capable network;wherein a different server descriptor ring is associated with each server of the plurality of servers, each different server descriptor ring comprising a respective write index, a respective read index, respective instructions, and respective commands for the respective server associated with the server descriptor ring;transmit to the first server a first read data buffer request from virtualization logic, wherein the first read data buffer request is sent from the remote entity to the server in response to the remote entity receiving the first write index of the first server descriptor ring from the first server;transmit to the second server a second read data buffer request from the virtualization logic, wherein the second read data buffer request is sent from the remote entity to the second server in response to the remote entity receiving the second write index of the second server descriptor ring from the second server;enforce traffic policing policies, with the virtualization logic associated with the remote entity, to prioritize access to each of the plurality of peripherals using a virtual device driver for the peripheral, wherein the virtualization logic allows each of the plurality of peripherals to be shared by the plurality of servers based at least in part on the different server descriptor rings, and wherein the access to each of the plurality of peripherals comprises requests from the first server and the second server forwarded to one network interface card of the one or more network interface cards, the remote entity:receiving server traffic comprising: receiving date transferred from the first server to the remote entity in response to the first read data buffer request;receiving data transferred from the second server to the remote entity in response to the second read data buffer request;performing packet inspection to classify at least some of the server traffic with classifier logic, the performing packet inspection comprising parsing data packets of the server traffic and utilizing lookup logic to determine priorities attributed to the data packets;placing the data packets into at least one prioritized queue based at least in part on the priorities determined;placing selected data into memory associated with the first peripheral; andposting a reference to the selected data, and facilitating reading of the selected data associated with the first peripheral by the first server or the second server to obtain the selected data from the memory associated with the first peripheral.

9. The system of claim 8, wherein, after receiving the first write index of the first server descriptor ring and before transmitting the first read data buffer request: sending a remote direct memory access read descriptor request from the virtualization logic;receiving by the first server one or more descriptor entries from the first server descriptor ring associated with the first server, the one or more descriptor entries including a pointer and a number of bytes; andsending a remote direct memory access write providing a read index, wherein transferring the data from the virtualization logic is based on the read index.

10. The system of claim 8, wherein the remote direct memory access capable networking technology is InfiniBand.

11. The system of claim 8, wherein the remote direct memory access capable networking technology is remote direct memory access over Ethernet.

12. The system of claim 8, wherein the remote direct memory access capable networking protocol is iSER.

13. The system of claim 8, wherein the remote direct memory access capable networking protocol is SRP.