Network cards

Marvell QLogic QLE2692

Product Brief Marvell® QLogic® QLE2690/2692Single-port and Dual-port Enhanced 16GFC Fibre Channel Adapters The latest and most advanced 16GFC HBA from Marvell,available in single- and dual- por




Product Brief

 

 

 

Marvell® QLogic® QLE2690/2692

Single-port and Dual-port Enhanced 16GFC Fibre Channel Adapters

 


The Marvell QLogic QLE2690 single-port and QLE2692 dual-port Enhanced 16Gb FC  (16GFC) HBAs boast industry-leading native FC performance with extremely low CPU usage and full hardware offloads.

 

Enhanced 16GFC

Compared to 16GFC, the Marvell QLogic Enhanced 16GFC solution offers higher

per-port performance (up to 650K IOPS) with low power consumption. In addition,      Marvell StorFusion technology delivers streamlined provisioning, QoS, and improved  resiliency while addressing the needs of IT organizations that require reliability, inte- grated management, and guaranteed network performance.

Enhanced 16GFC technology resolves data center complexities by enabling a storage network infrastructure that supports powerful virtualization features, applica-

tion-aware services, and simplified management. The QLE2690/2692 Adapters provide advanced storage networking features capable of supporting the most demanding

virtualized and private cloud environments, while fully leveraging the capabilities of

high-performance 16GFC and all-flash arrays (AFAs). Powerful management tools

automate and simplify SAN provisioning to help reduce cost and complexity, while the

unmatched 16GFC performance eliminates potential I/O bottlenecks in today’s powerful multiprocessor, multicore servers.

 

Superior Performance

The QLE2690/2692 Adapters can accelerate mission-critical enterprise applications by delivering up to 1.3 million IOPS for physical, virtual, and private cloud environments.

Marvell QLogic adapters deliver the best storage application performance in virtualized and non-virtualized environments with support for up to  6,400MBps of aggregate

throughput.

 

NVMe® Over Fibre Channel (FC-NVME)

The QLE2690/2692 Adapters support the FC-NVMe-2 (NVMe over Fibre Channel)

protocol. The QLE2690/2692 Adapters can simultaneously support FC-NVMe and

FCP-SCSI storage traffic. NVMe storage offers exceptionally high performance at very  low latencies. NVMe works best with a network that can provide lossless, low-latency,   and high-performing transport. The 2600 Series Enhanced 16GFC Adapters bring the best of both worlds by offering the highest performance and lowest latency access to  NVMe and SCSI storage over an FC network.


Virtualization Optimized

The QLE2690/2692 Adapters support standards-based virtualization features. Under

VMware® ESXi 6.x and later, I/O requests and responses can be tagged with the virtual     machine-ID (VM-ID) of the appropriate virtual machine, providing end-to-end visibility at the VM level. Support for N_Port ID virtualization (NPIV) enables a single FC adapter

port to provide multiple virtual ports for increased network scalability. In addition, the    16GFC line rate per physical port delivers unmatched storage performance to maximize the number of VMs per physical server.

 

Marvell StorFusion Technology

Marvell QLogic Enhanced 16GFC Adapters, powered by StorFusion technology, include advanced capabilities that are enabled when deployed with supported Brocade® and

Cisco® switches. By combining these industry-leading solutions, SAN administrators can take advantage of enhanced features that improve availability, accelerate deployment,     and increase network performance.

Marvell Universal SAN Congestion Mitigation Technology (USCM)

SAN congestion, although rare, has the potential to quickly spread and significantly

disrupt application performance leading to lost business. Modern SANs, specifically

those with flash storage and mix of Fibre Channel endpoint speeds can be prone to

congestion. SAN Congestion typically occurs when slower FC endpoints cannot accept  responses generated by flash/NVMe storage, a condition referred to as oversubscrip-     tion/over-utilization. Rarely, a misbehaving FC device can also lead to SAN congestion    when it is unable to replenish credits which are key to reliable Fibre Channel transmis-  sions, a condition referred to as slow-drain. SAN congestion when timely detected, can be isolated and decisive actions applied. A Fibre Channel standard, Fabric Performance Impact Notifications (FPINs) bring congestion awareness  to FC endpoints.

Utilizing FPINs, Marvell’s QLogic 2690 Series Adapters implement Universal SAN

Congestion Mitigation Technology (USCM). USCM uniquely works both Brocade and

Cisco SAN fabrics to keep applications running at peak performance by providing an

in-band and zero touch SAN congestion detection, notification, and avoidance system. QLogic 2690 Series HBAs interact with Cisco and Brocade SANs using both the more

reliable FC primitive signaling as well as protocol events to bring awareness of the

presence of congestion, peer congestion, link integrity, lost frames etc. QLogic 2690

Series HBAs also have the ability to enable decisive actions such as fine-grained I/O

throttling, automatic path failover, load balancing, and flow quarantining to ensure that SAN congestion is mitigated with minimum impact to application performance.

Marvell QLogic USCM technology, offered at no additional cost, brings strong awareness of SAN congestion and implements decisive actions to prevent application degradation.  It is highly recommended for modern SANs.

Improved Total Cost of Ownership and Reliability

StorFusion technology delivers advanced link diagnostics, which improve availability and

support for high-performance fabrics. Using the Diagnostics Port feature with a

Brocade switch that supports ClearLink®, administrators can quickly run a battery of automated diagnostic tests to assess the health of links and fabric components.

The QLE2690/2692 Adapters support link cable beacon (LCB), which enables adminis- trators to visually identify both ends of a physical link. In a large data center with

hundreds of ports and cables to manage, a simple command turns on port LED beacons on both ends of a link cable connection. Administrators can use LCB to quickly identify    connection peer ports without tracing the cable.

Marvell technology includes the read diagnostic parameters (RDP) feature, which

provides detailed port, media, and optics diagnostics. From any point in the fabric, an administrator can use RDP to easily discover and diagnose link-related errors and

degrading conditions on any N_Port-to-F_Port link.

The extensive suite of diagnostic tools maximizes uptime and performance, allowing organizations to address problems before they impact operations.

Rapid Server Deployment and Orchestration

StorFusion technology includes fabric pre-provisioning services that enable servers to

be quickly deployed, replaced, and moved across the SAN. By leveraging the fabric-as-

signed port worldwide name (FA-WWN) and fabric-based boot LUN discovery (F-BLD)

capabilities, the creation of zones, LUNs, and other services can be completed before the servers arrive on site—eliminating time consuming, manual tasks that typically delay

server deployment.

Performance SLA Enforcement with VM-level QoS

Network performance can be dramatically improved by implementing the indus-

try-standard class-specific control (CS_CTL)-based frame prioritization QoS, which helps alleviate network congestion. When Marvell QLogic adapters with StorFusion technol-      ogy are connected to supported SAN fabrics, traffic is classified as it arrives at the

switch, and is then processed on the basis of configured priorities. Traffic can be

prioritized for delivery or subjected to limited delivery options. As a result, mission-criti- cal workloads can be assigned a higher priority than less time-sensitive network traffic    to achieve optimized performance.

Higher Resiliency and Performance with Automatic Error Recovery

Forward error correction (FEC) improves performance and link integrity to support

higher end-to-end data rates by automatically recovering from many transmission

errors without re-sending the frames. FEC automatically detects and recovers from bit errors, which results in higher availability and performance.

Automatic buffer-to-buffer credit recovery (BB-CR) helps overcome performance   degradation, congestion, and link resets caused by buffer credit loss, especially on longer distance and high-loss fiber connections.

热门推荐

Home
Product
Company profile
Contact