WEKA networking
This page reviews the theory of operation for WEKA networking.
Overview
The WEKA system supports the following types of networking technologies:
InfiniBand (IB)
Ethernet
The networking infrastructure dictates the choice between the two. If a WEKA cluster is connected to both infrastructures, it is possible to connect WEKA clients from both networks to the same cluster.
The WEKA system networking can be configured as performance-optimized or CPU-optimized. In performance-optimized networking, the CPU cores are dedicated to WEKA, and the networking uses DPDK. In CPU-optimized networking, the CPU cores are not dedicated to WEKA, and the networking uses DPDK (when supported by the NIC drivers) or in-kernel (UDP mode).
Performance-optimized networking (DPDK)
For performance-optimized networking, the WEKA system does not use standard kernel-based TCP/IP services but a proprietary infrastructure based on the following:
Use DPDK to map the network device in the user space and use the network device without any context switches and with zero-copy access. This bypassing of the kernel stack eliminates the consumption of kernel resources for networking operations. It applies to backends and clients and lets the WEKA system saturate 200 GB links.
Implementing a proprietary WEKA protocol over UDP, i.e., the underlying network, may involve routing between subnets or any other networking infrastructure that supports UDP.
The use of DPDK delivers operations with extremely low-latency and high throughput. Low latency is achieved by bypassing the kernel and sending and receiving packages directly from the NIC. High throughput is achieved because multiple cores in the same server can work in parallel without a common bottleneck.
Before proceeding, it is important to understand several key terms used in this section, namely DPDK, and SR-IOV.
DPDK
Data Plane Development Kit (DPDK) is a set of libraries and network drivers for highly efficient, low-latency packet processing. This is achieved through several techniques, such as kernel TCP/IP bypass, NUMA locality, multi-core processing, and device access via polling to eliminate the performance overhead of interrupt processing. In addition, DPDK ensures transmission reliability, handles retransmission, and controls congestion.
DPDK implementations are available from several sources. OS vendors such as Redhat and Ubuntu provide their DPDK implementations through their distribution channels. Mellanox OpenFabrics Enterprise Distribution for Linux (Mellanox OFED), a suite of libraries, tools, and drivers supporting Mellanox NICs, offers its own DPDK implementation.
The WEKA system relies on the DPDK implementation provided by Mellanox OFED on servers equipped with Mellanox NICs. For servers equipped with Intel NICs, DPDK support is through the Intel driver for the card.
SR-IOV
Single Root I/O Virtualization (SR-IOV) extends the PCI Express (PCIe) specification that enables PCIe virtualization. It allows a PCIe device, such as a network adapter, to appear as multiple PCIe devices or functions.
There are two function categories:
Physical Function (PF): PF is a full-fledged PCIe function that can also be configured.
Virtual Function (VF): VF is a virtualized instance of the same PCIe device created by sending appropriate commands to the device PF.
Typically, there are many VFs, but only one PF per physical PCIe device. Once a new VF is created, it can be mapped by an object such as a virtual machine, container, or, in the WEKA system, by a 'compute' process.
SR-IOV technology must be supported by software and hardware to take advantage of it. The Linux kernel provides SR-IOV software support. The computer BIOS and the network adapter provide hardware support (by default, SR-IOV is disabled and must be enabled before installing WEKA).
CPU-optimized networking
For CPU-optimized networking, WEKA can yield CPU resources to other applications. That is useful when the extra CPU cores are needed for other purposes. However, the lack of CPU resources dedicated to the WEKA system comes with the expense of reduced overall performance.
DPDK without the core dedication
For CPU-optimized networking, when mounting filesystems using stateless clients, it is possible to use DPDK networking without dedicating cores. This mode is recommended when available and supported by the NIC drivers. In this mode, the DPDK networking uses RX interrupts instead of dedicating the cores.
Note: This mode is supported in most NIC drivers. Consult https://doc.dpdk.org/guides/nics/overview.html for compatibility.
AWS (ENA drivers) does not support this mode, hence, in CPU-optimized networking in AWS, use the UDP Mode.
UDP mode
WEKA can also use in-kernel processing and UDP as the transport protocol. This operation mode is commonly referred to as UDP mode.
Because the UDP mode uses in-kernel processing, it is compatible with older platforms lacking the support of kernel offloading technologies (DPDK) or virtualization (SR-IOV), as legacy hardware such as the Mellanox CX3 family of NICs.
Typical WEKA configuration
Backend servers
In a typical WEKA system configuration, the WEKA backend servers access the network function in two different methods:
Standard TCP/UDP network for management and control operations.
High-performance network for data-path traffic.
Note: To run both functions on the same physical interface, contact the Customer Success Team.
The high-performance network used to connect all the backend servers must be DPDK-based. This internal WEKA network also requires a separate IP address space (see Network Planning and Configuration of Networking). The WEKA system maintains a separate ARP database for its IP addresses and virtual functions and does not use the kernel or operating system ARP services.
Backend servers with DPDK-supporting Mellanox NICs
For backend servers equipped with DPDK-supporting Mellanox (CX-4 or newer) NICs, the following conditions must be met:
Mellanox OFED is installed and loaded.
SR-IOV enabled is not required.
The number of IPs allocated to the backend servers on the internal network must be the total number of backend servers. For example, 8 IPs for 8 backend servers.
Backend servers with DPDK-supporting Intel E810 NIC
For backend servers equipped with DPDK-supporting Intel E810 NIC using a single IP, the following condition must be met:
SR-IOV is enabled in the hardware (BIOS + NIC).
Backend servers with DPDK-supporting the other NICs
For backend servers equipped with DPDK-supporting the other NICs, the following conditions must be met:
SR-IOV is enabled in the hardware (BIOS + NIC).
The number of IPs allocated to the backend servers on the internal network must equal the total number of WEKA software processes plus the total number of backend servers. For example, a cluster of 8 backend servers running 10 WEKA processes each requires 88 (80 + 8) IPs on the internal network. The IP requirements for the WEKA clients are outlined below in the Clients section.
Clients
Unlike WEKA backend servers that must be DPDK/SR-IOV based, the WEKA clients (application servers) can use either DPDK-based or UDP modes. The DPDK mode is the natural choice for the newer, high-performing platforms that support it.
Clients with DPDK-supporting Mellanox NICs
For clients equipped with DPDK-supporting Mellanox (CX-4 or newer) NICs, the following conditions must be met:
Mellanox OFED is installed and loaded.
SR-IOV enabled is not required.
The number of IPs allocated to the clients on the internal network must equal the total number of clients. For example, 10 IPs for 10 clients.
Clients with DPDK-supporting Intel E810 NIC
For clients equipped with DPDK-supporting Intel E810 NIC using a single IP, the following condition must be met:
SR-IOV is enabled in the hardware (BIOS + NIC).
Clients with DPDK supporting the other NICs
For clients equipped with DPDK-supporting the other NICs, the following conditions must be met to use the DPDK mode:
SR-IOV is enabled in the hardware (BIOS + NIC).
The number of IPs allocated to the Intel clients on the internal network must equal the total number of WEKA system FrontEnd (FE) processes (typically no more than 2 per server) plus the total number of clients. For example, 10 clients with 1 FE process per client require 20 IPs (10 FE IPs + 10 IPs).
Clients in UDP mode
The UDP mode is available for legacy clients lacking SR-IOV or DPDK support or where there is no requirement for low latency, high throughput IO.
For clients in the UDP mode, the following conditions must be met:
The native driver is installed and loaded.
The number of IPs allocated to the clients on the internal network must equal the total number of clients. For example, 10 clients in the UDP mode require 10 IPs on the internal network.
High Availability (HA)
To support HA, the WEKA system must be configured with no single component representing a single point of failure. Multiple switches are required, and servers must have one leg on each.
HA for servers is achieved either through implementing two network interfaces on the same server or by LACP (ethernet only, modes 1 and 4). A non-LACP approach sets a redundancy that enables the WEKA software to use two interfaces for HA and bandwidth.
HA performs failover and failback for reliability and load balancing on both interfaces and is operational for Ethernet and InfiniBand. Not using LACP requires doubling the number of IPs on both the backend containers and the IO processes.
When working with HA networking, it is helpful to label the system to send data between servers through the same switch rather than using the ISL or other paths in the fabric. This can reduce the overall traffic in the network. To label the system for identifying the switch and network port, use the label parameter in the weka cluster container net add command.
Note: LACP (link aggregation, also known as bond interfaces) is currently supported between ports on a single Mellanox NIC and is not supported when using VFs (virtual functions).
RDMA and GPUDirect storage
GPUDirect Storage enables a direct data path between storage and GPU memory. GPUDirect Storage avoids extra copies through a bounce buffer in the CPU’s memory. It allows a direct memory access (DMA) engine near the NIC or storage to move data directly into or out of GPU memory without burdening the CPU or GPU.
When enabled, the WEKA system automatically utilizes the RDMA data path and GPUDirect Storage in supported environments. When the system identifies it can use RDMA, both in UDP and DPDK modes, it employs the use for workload it can benefit from RDMA (with regards to IO size: 32K+ for reads and 256K+ for writes).
Using RDMA/GPUDirect Storage, it is thus possible to get a performance gain. You can get much higher performance from a UDP client (which does not require dedicating a core to the WEKA system), get an extra boost for a DPDK client, or assign fewer cores for the WEKA system in the DPDK mode to get the same performance.
Limitations
For the RDMA/GPUDirect Storage technology to take effect, the following requirements must be met:
All the cluster servers support RDMA networking.
For a client:
GPUDirect Storage: The IB interfaces added to the Nvidia GPUDirect configuration should support RDMA.
RDMA: All the NICs used by WEKA must support RDMA networking.
Encrypted filesystems: The framework will not be used for encrypted filesystems and will fall back to work without RDMA/GPUDirect for IOs to encrypted filesystems.
A NIC is considered to support RDMA Networking if the following requirements are met:
For GPUDirect Storage only: InfiniBand network.
Mellanox ConnectX5 or ConnectX6.
OFED 4.6-1.0.1.1 or higher.
For GPUDirect Storage: install with
--upstream-libsand--dpdk.
Note: GPUDirect Storage bypasses the kernel and does not use the page cache. Standard RDMA clients still use the page cache.
Note: RDMA/GPUDirect Storage technology is not supported when working with a mixed IB and Ethernet networking cluster.
Running weka cluster processes indicates if the RDMA is used. For example:
Note: GPUDirect Storage is auto-enabled and detected by the system. Contact the Customer Success Team to enable or disable RDMA networking on the cluster or a specific client.
Last updated