I/O

NextIO Raises $15 Million In Fourth Round Of Financing

June 8, 2009 by Robin Wauters Leave a Comment

NextIO today announced that the company has received a $15 million round of Series D funding. NextIO is the industry leader in providing intelligent, rack-optimized I/O virtualization (IOV) solutions based on high-performance PCI Express (PCIe) switching technology.

NextIO received support not only from existing VCs and strategic investors, but participation from new strategic investment sources as well. The new funding will be used to expand NextIO’s sales, marketing and support and will allow the company to accelerate the delivery of its second-generation of Express Connect products.

Through I/O virtualization, NextIO products deliver an unprecedented combination of low cost, lower power consumption, greater ease of use, and greatly reduced maintenance requirements. For example, server rollouts that previously required days or weeks can be accomplished in less than 30 minutes, and routine server maintenance time is reduced from hours to minutes.

NextIO’s second-generation products build on the success of the company’s existing Express Connect products, which are at revenue and are currently being deployed in the U.S. and Europe. NextIO solutions are currently in use in a number of different markets for applications such as simulation, adding graphic processors to commodity computing environments, expanding I/O, speeding up existing I/O, and delivering new data center technologies.

Fusion-io Partners With IBM To Boost Storage System Performance

September 3, 2008 by Robin Wauters Leave a Comment

Fusion-io, a provider of enterprise solid-state technology and high-performance I/O solutions, today announced (PDF) that the company is working with IBM to dramatically accelerate data access performance in IBM’s clustered storage systems. The two companies have adapted Fusion-io’s enterprise SSD technology for “Project Quicksilver,” demonstrating increased data access performance and reduced power consumption, while dramatically reducing latency and bottlenecks that tend to be compounded in today’s high transaction and virtualized environments.

Combining Fusion-io’s silicon-based, NAND flash storage technology with IBM’s storage technology and system and application expertise will revolutionize how enterprises access and store large amounts of data, enabling a new performance standard that addresses the growing demands from virtualization, as well as high-transaction and data-intensive I/O applications. This powerful combination will provide customers with a whole new way to architect their high performance storage needs with unparalleled performance, flexibility and scalability, while consuming far less power.

The increasing demand for a next generation solid state storage technology is driven by advancements in computer processors which, following Moore’s Law, have grown exponentially in performance. Mechanical disks, on the other hand, follow Newtonian Dynamics and experience lackluster performance improvements, introducing a performance gap. The ioMemory technology from Fusion-io creates a new tier in the memory hierarchy – one that has 100 times the capacity density and 10 times the capacity per dollar of DRAM. NAND flash-based ioMemory makes it possible to have terabytes of near-memory-speed storage within each node – bringing extremely large memory problems and I/O bound analysis to a new level of cost effectiveness.

Fusion-io

First Workshop on I/O Virtualization (WIOV’08) Announced

August 7, 2008 by Kris Buytaert Leave a Comment

Usenix has just announced it’s first Workshop on I/O Virtualization

Scheduled to take place in San Diego on 10 and 11 december this year , the event wil
l be co-located with the Usenix Symposium on Operating Systems Design and Implementa
tion (OSDI’08)

The past decade the focus was mostly on processor and memory virtualization. I/O vi
rtualization has received less attention. This workshop will provide a forum to dis
cuss the challenges of I/O virtualization that span the virtual machine monitor, gue
st operating system, processor, memory subsystem, and I/O subsystem

The workshop also opened its Call for Papers

KVM-65 Released, Supports S390 Architecture

April 7, 2008 by Robin Wauters Leave a Comment

KVM-65 was released today. The most interesting feature in this release is support for the S/390 architecture, more specifically, the System z9 line of mainframes. On x86, the most interesting change is the separation of timer and I/O completion handling into a separate thread (these used to be serviced by the same thread that executed vcpu 0). The change should result in improved responsiveness and better smp performance.

As Anthony Liguori puts it:

“The s390 is the grand-daddy of virtualization. Everything started there. In so many ways, everything we’re doing with x86 virtualization is just playing catch-up. The new exciting features like hardware virtualization support and hardware paging support have been in s390 forever.”

“s390 clearly has a very mature hypervisor. What many people may not know though is that it’s normal to run two hypervisors at any given time on s390. At the bottom level, there’s PR/SM which divides the machine into rather coarse partitions. Within a PR/SM partition, you can run z/OS or Linux. You can also run z/VM within a PR/SM partition. z/VM is another hypervisor that allows for much more sophisticated features like memory overcommit and processor overcommit. The user has the ability to decide how much hypervisor they need to maximize the efficiency of their workloads.”

These are the changes from KVM-64:

fix hotplug build for non-x86
ignore reads from the apic EOI register
fixes Linux 2.6.25-rclate bootup problems
compile fixes
fix ftruncate() on hugetlbfs use on older Linux hosts
endianness fix virtio-block
fixes virtio-blk on ppc
refactor in-kernel PIT to be a separate device
separate thread for I/O completions and timers
fix vmmouse smp
fix loading uninitialized variable into apic registers
fixes apic being disabled on smp Linux guests running X
disable kvm clock on Voyager or SGI Visual WS
s390 support
fix large pages
speedup msr processing on Intel via msr bitmap
add slab shrinker support
reduces nonswappable footprint under memory pressure
code cleanup
vm refcounting
only mark a page as accessed if it was really accessed by the guest
drop slots_lock while in guest mode
fixes long latencies with iothread
prepopulate guest pages only after write-protecting them
fixes smp race leading to guest spinning

[Source: Tales of a Code Monkey]

NextIO Raises $ 18.8 Million In Series C Funding For Virtualizing Server Communication

March 2, 2008 by Robin Wauters 1 Comment

Austin-based connectivity virtualization developer NextIO announced Friday that it had raised $ 18.8 million in a third round series as the five-year-old company prepares to launch its first products to enable distributed computing systems to connect and communicate using any of a variety of technology standards.

The deal was led by Crescendo Ventures and Adams Capital Management, and included previous investors JK&B Capital, the VentureTech Alliance arm of Taiwan Semiconductor Manufacturing Corp., and Dell, as well as two new strategic investors. The new funding brings total equity investment in NextIO to about $ 40 million and will allow the company to launch its first commercial products with original equipment manufacturers (OEM) partners in the second quarter of this year.

NextIO co-founder and CEO K.C. Murphy said the company made its first switch products available to customers in late 2007 but will launch its primary appliance product in April, in conjunction with new products introduced by OEM partners. The products are aimed at bringing the same advantages of virtualization to connectivity technology that developers including VMWare, Microsoft and the open source Xensource platform have brought to the server market.

NextIO’s product virtualizes I/O functions to bring high-performance computing capabilities to clustered systems of commodity servers. The products enable individual server connectivity using any of the several available connectivity standards with systems level technology that controls connectivity within a virtual “cloud.”

[Source: TechConfidential]

Neterion Upgrades “Virtualisation-Aware Network Adapters”

February 25, 2008 by Robin Wauters Leave a Comment

Neterion, an industry leader in 10 Gig Ethernet adapters for server and storage environments, today announced its third generation X3100 Series adapters at VMworld in Cannes, France.

The cards offload work normally done in the hypervisor – in most virtual machines, the network driver routes I/O to a virtual network within the hypervisor, which then controls the physical network interface. In contrast, Neterion’s X3100 series card behaves like multiple physical network cards, allowing each VM to have its own network interface.

“Virtualization is a key technology for reducing total cost and complexity. However, applications that required high-performance systems experienced I/O bottlenecks that limited their virtualization opportunities,” said Bob Wheeler, senior analyst at The Linley Group. “The IOV standard was designed to eliminate these bottlenecks and open up all applications to virtualization. Neterion’s SR-IOV compliant 10 GbE adapters, with support for VMware’s NetQueue technology, demonstrated near 10 Gbps line rate throughput under VMware ESX Server 3.5. This kind of performance makes it possible to virtualize even the most I/O intensive applications in enterprise data centers.”

“This product announcement is a watershed event for the industry,” said Dave Zabrowski, president & CEO of Neterion. “With X3100 Series Adapters supporting full IOV compliance with a silicon-based I/O path architecture, all applications become candidates to run on a virtualized server. This means data centers can realize new cost reduction opportunities while maintaining performance levels comparable to non- virtualized environments. Our new adapter series provides the missing piece of virtualization to data centers that will undoubtedly accelerate the adoption of server virtualization.”

The devices are compatible with VMware’s ESX Server 3.5 and its NetQueue I/O virtualisation (IOV) spec, and can reduce CPU utilization by up to 50 % via a new Large Receive Offload feature, CEO Dave Zabrowski added.He claimed they are also the first of their kind to support the PCI-SIG’s new SR-IOV (single-root IOV) standard, which allows one adapter to behave like multiple physical adapters in a virtualized server.

“This product announcement is a watershed event for the industry,” said Dave Zabrowski, president & CEO of Neterion. “With X3100 Series Adapters supporting full IOV compliance with a silicon-based I/O path architecture, all applications become candidates to run on a virtualized server. This means data centers can realize new cost reduction opportunities while maintaining performance levels comparable to non- virtualized environments. Our new adapter series provides the missing piece of virtualization to data centers that will undoubtedly accelerate the adoption of server virtualization.”

[Source: The Register]