Path: EDN Asia >> Design Centre >> Test & Measurement >> Exercising storage products
Test & Measurement Share print

Exercising storage products

07 Feb 2014  | John Wiedemeier

Share this page with your friends

Jammers were so effective for traditional storage applications because they grew out of a more simplistic test methodology. Early in the computer industry, storage communications between a host initiator and target devices was less complicated. When a host sends a command sequence, the target returns a status code byte indicating the command was successful, encountered an error, or indicated that the target was busy servicing other requests. It then moved onto the next command in the sequence or retransmitted until the message was finally successful. Rigid ribbon cables were used to transmit low speed data communications and lengths were limited by crosstalk noise from the cables themselves. It was during this time that the jammer became a standard tool for storage test applications. The process of setting up an error scenario (jam) was uncomplicated with few protocol safeguard mechanisms to interfere with line rate traffic modification on the test set up.


Testing storage protocols in the past
As data communication evolved into serial protocol transmissions, things changed from a data integrity standpoint. Storage eventually adopted a "point-to-point" architecture that moved away from shared channels. The shift increased data rates and reduced drive complexity and cost. Differential signalling, 8b/10b encoding, and primitives for alignment were used to reduce noise issues at the physical layer. Link layer frame information structure (FISs) packets containing control information or payload data mechanisms were created providing a higher level of link control than was afforded in previous generations of parallel ATA control systems. Once the data was processed by the link layer, the transport layer processed the FIS information and forwarded the data payload to the command layer for execution. These data structures became the focus of storage verification test plans. Jammers were ideal for testing these data structures and were used widely by the storage industry. As these protocol interfaces increased in complexity and speed, many of the jammer test features had to be scaled back, reducing the overall capability and effectiveness of the jammer.

In recent years, the storage industry has adopted PCI Express-based SSD storage protocol interfaces such as NVMe, SATA Express, and SCSI Express to address continued needs for throughput while reducing protocol latencies. New protocols using encoded data protection mechanisms that ensure reliable data transmission are required for this higher performance protocol interface. PCIe uses the following data protection mechanisms on top of existing physical layer features such as differential signalling, encoded data transmission, and spread spectrum clocking:

 • All Transaction Layer Packets (TLPs) need to be acknowledged ("Acked"). If an Ack is not received, an Ack latency timeout occurs, causing automatic "replays" of the TLP packet.
 • An incrementing sequence number is attached to every new TLP. The receiver keeps an expected next sequence number, so it knows what to expect in the next received TLP. The receiver checks the sequence number received against the expected sequence number and detects one or more lost TLPs. This receiver check ensures TLPs are not lost and that they maintain proper ordering.
 • All packet structures are automatically checked to see if they are framed correctly and that fields are properly described.
Test methodologies that developed around PCIe included protocol analysers for bus monitoring and protocol exercisers for generating controlled test stimulus to both host and device systems. The exerciser is also flexible enough to provide low level PCIe emulation as a host or a device. Powerful corner case and stress testing by can be conducted by controlling protocol traffic including speed and lane width changes, flow control, TLP and DLLP transactions, using the exerciser. In addition, various error conditions can be created through simple scripting techniques. Possible Exerciser Error Generation

Physical Layer Error List

 • Receiver Error
Data Link Layer Error List

 • Bad TLP
 • BAD DLLP
 • Bad LCRC
 • Replay Timeout
 • Replay NUM Rollover
 • Data Link Layer Protocol Error
 • Surprise Down
Transaction Layer Error List

 • Poisoned TLP Received
 • ECRC Check Failure
 • Unsupported Request(UR)
 • Completion Timeout
 • Completer Abort
 • Unexpected Completion
 • Receiver Overflow
 • Flow Control Protocol Error

There remains interest by the storage validation community in using the previous test methodology and equipment for validating PCIe storage systems. Driven by this desire, early efforts were made to develop and deploy a PCIe jammer. Because of the limitations imposed by the PCIe protocol such as protocol delivery mechanisms and encapsulated command/control structures, these efforts resulted in a device that lacked its overall test capability and effectiveness delivering a subset of its original test features when testing PCIe storage systems. Thus, many designers and interoperability test engineers have using jammers jammer and moved to more effective test tool.

 First Page Previous Page 1 • 2 • 3 Next Page Last Page


Want to more of this to be delivered to you for FREE?

Subscribe to EDN Asia alerts and receive the latest design ideas and product news in your inbox.

Got to make sure you're not a robot. Please enter the code displayed on the right.

Time to activate your subscription - it's easy!

We have sent an activate request to your registerd e-email. Simply click on the link to activate your subscription.

We're doing this to protect your privacy and ensure you successfully receive your e-mail alerts.


Add New Comment
Visitor (To avoid code verification, simply login or register with us. It is fast and free!)
*Verify code:
Tech Impact

Regional Roundup
Control this smart glass with the blink of an eye
K-Glass 2 detects users' eye movements to point the cursor to recognise computer icons or objects in the Internet, and uses winks for commands. The researchers call this interface the "i-Mouse."

GlobalFoundries extends grants to Singapore students
ARM, Tencent Games team up to improve mobile gaming


News | Products | Design Features | Regional Roundup | Tech Impact