The Storage Performance Development Kit (SPDK) provides a set of tools and libraries for writing high performance, scalable, user-mode storage applications.
Recently, in the v21.07 release, SPDK got
support for BPF tracing. It consists of a set of scripts and static probes defined within the SPDK
libraries giving the user yet another way to inspect SPDK-based applications. SPDK already had a
tracing library, called libtrace, for a long time. BPF tracing isn’t supposed to replace it, but
rather complement it. Both have slightly different capabilities - libtrace has significantly
lower overhead, but is less flexible. BPF traces, on the other hand, are more costly (they require
a kernel trap), but are more robust and can even be attached dynamically, without having to define
the probes in code.
Performance report document for SPDK NVMe Bdev 21.07 release has been published.
Continue...Performance report document for SPDK NVMe-oF TCP 21.07 release has been published.
Continue...Performance report document for SPDK NVMe-oF RDMA 21.07 release has been published.
Continue...Performance report document for SPDK Vhost 21.07 release has been published.
Continue...On behalf of the SPDK community I’m pleased to announce the release of SPDK 21.07!!
Continue...Performance report documents for SPDK 21.04 release have been published.
Continue...Thank you all for attending the SPDK, PMDK and Intel® Performance Analyzers Virtual Forum. We had a very successful Virtual Forum this year with awesome content and presenters. You can download the presentations using the links below. Please note these are static PDFs, but many of the presentations made extensive use of animations, so something may be lost in translation.
Continue...The adoption of NVM Express® SSDs in the data center is unleashing performance that was unprecedented just a few years ago. Since the inception of the SPDK project, the community has demonstrated that it’s possible to build systems capable of tens of millions of IOPS (I/O operations per second) using commodity hardware. Over the last couple of years, we have demonstrated systems capable of over 10 million IOPS (using a single CPU core!), but now we’re ready to significantly advance the state of the art. Yes, 80 Million IOPS.
Continue...