Software-based techniques are transforming networking. Commercial off-the-shelf hardware is finding a place in several networking use cases. However, high-performance hardware is also an important part of a successful software-defined networking (SDN). As you optimize your networks using SDN tools and complementary technologies such as network function virtualization (NFV), an important step is to strategically assess your hardware needs based on the functions and performance requirements. These need to be aligned with your intended business outcome for individual applications and services.
Two Categories of High Performance Hardware
- Network hardware that utilizes purpose-built designs. These often involve specialized Application-Specific Integrated Circuit (ASIC)s to achieve significantly higher performance than what is possible or economically feasible using commercial off-the-shelf servers that are based on state of the art, x86-based, general purpose processors.
- Network hardware that uses standard x86 servers that is enhanced to provide high performance and predictable operation for example, via special software techniques that bypass hypervisors, virtualization environments, and operating systems.
Where to Deploy Network Functions
Can virtualized network functions be deployed like cloud-based applications? No. There is a big difference between deploying network functions as software modules on x86 general purpose servers and using a common cloud computing model to implement network virtualization. Simply migrating existing network functions to general purpose servers without due regard to all the network requirements leads to dramatically uneven and unpredictable performance. This unpredictability is mainly due to data plane workloads being often I/O bound and/or memory bound and software layers containing important configuration details that may impact performance.
These issues are not specifically about hardware but how the software handles the whole environment. Operating systems, hypervisors, and other infrastructure that is not integrated into best practices for data plane applications will continue to contribute to unpredictable performance.
Bandwidth and CPU Needs
A good way to begin to assess hardware requirements is to examine network functions in two dimensions: I/O bandwidth or throughput needs, and computational power needs. In considering which network function to virtualize and where to virtualize it, CPU load required and bandwidth load required throughout different layers of the network can help determine that some but not all network functions are suitable for virtualization.
Applications with lower I/O bandwidth and low-to-high CPU requirements may be most appropriate for virtualized deployment on optimized x86 servers. Applications with higher I/O bandwidth and low-to-high CPU requirements may be best deployed on specialized high-performance hardware with specialized silicon. Many other factors may play a role in determining what hardware to use for which applications, including cost, user experience, latency, networking performance, network predictability, and architectural preferences.
Service-Network Abstraction is Key
Additionally, you might not need high performance hardware for certain functions initially. But as such a particular function scales, it might require a high performance platform to meet its performance specifications, or it might be more economical on a purpose-built platform. So you might start out with commercial off-the-shelf hardware and then transfer the workload to the high performance hardware later. If you have focused on establishing a clean abstraction of the services from the underlying hardware infrastructure using SDN principles, the network deployment can be more easily changed or evolved independently of the upper services and applications. This is the true promise of SDN.
Read more about how to assess hardware performance requirements in your SDN in the Cisco® white paper “High-Performance Hardware: Enhance Its Use in Software-Defined Networking.” You can find it here: “Do You Know your Hardware Needs?” along with other useful information.
Do you have questions or comments? Tweet us at @CiscoSP360
Tags: abstraction, ASIC, cpu, High Performance Hardware, network function virtualization, Networking Performance, SDN, software defined network, VNF
Today, I am pleased to announce Cisco’s intent to acquire privately held Memoir Systems, a company that develops semiconductor memory intellectual property (IP) and tools that enable ASIC vendors to build programmable network switches with increasing speeds. This acquisition will enable the proliferation of affordable, fast memory for existing Cisco switch ASICs and will help advance Cisco’s ASIC innovations necessary to meet next-generation IT requirements.
Currently in the data center switching market, denser infrastructure and data-intensive workloads are driving demand for higher port density (feeds) and greater bitrates (speeds). At the same time, the accelerating growth of scale-out (non-virtualized) Big Data applications like Hadoop are driving increasing East-West data traffic – furthering the need for greater data center network density. Unfortunately, the physical memory in typical ASIC switch chips cannot cope with the design requirements for these more intense needs and as a result, can become the bottleneck that limits the density and performance of future data center switches.
To help solve the ASIC memory issue, Memoir currently licenses soft-logic IP, which speeds up memory access by up to 10 times. It also reduces the overall footprint this memory takes up in typical switch ASICs. As a result, this technology allows the development of switch and router ASICs with speeds, feeds, and costs typically not possible with traditional physical memory design techniques. This differentiation is critically important as port densities and port speeds move from 10G to 40/100G.
The acquisition of Memoir Systems is expected to close in the first quarter of Cisco’s fiscal year 2015. The Memoir team will report into Cisco’s Insieme Business Unit, under Senior Vice President, Mario Mazzola.
I look forward to seeing Memoir’s technology used across Cisco’s future ASIC projects. Memoir’s technology and strong team will allow Cisco to continue to innovate at the chip level and advance our ASIC and overall networking strategies.
Tags: ASIC, Big Data, Cisco acquisitions, datacenter, Hilton Romanski, Insieme, intellectual property, Mario Mazzola, Memoir Systems, Mergers and Acquisitions, programmable network switch, semiconductor memory
It’s always interesting and often entertaining to observe how competitors promote their products and what they choose to focus on—and more importantly, what they choose not to focus on and what they hope people won’t ask questions about.
Consider yet again how a competitor chooses to position their “purpose built” AP vs. the Cisco Aironet 3700 802.11ac Access Point Series.
This competitor frequently (and somewhat obsessively) points out that its 802.11ac AP has dual “active” 800 MHz cores while the Cisco AP3700 has only one “active” 800 MHz core. This is not completely true since it completely overlooks the fact that the Cisco AP3700 also has a dedicated CPU core and DSP for each radio subsystem.
Furthermore, it also overlooks that the dual “active” cores in the competitor’s AP share 512 MB of DRAM. The single “active” core of the AP3700 has dedicated 512 MB of DRAM. Also each radio subsystem has a dedicated 128 MB DRAM (for 768 MB total DRAM in the AP3700).
Why is all of this important? Read More »
Tags: 11ac, 802.11ac, access point, AP, architecture, ASIC, client, compute, Computing, design, device, DRAM, efficiency, hardware, HD, HDX, high density, Industry, infrastructure, LAN, MB, memory, memory management, Mhz, mobile, mobility, network, networking, performance, rf, system, wi-fi, wifi, wireless, wlan
As a Product Manager there is some anxiety but more of an excitement around introducing a platform to the market. Today I am proud to be part of Cisco team that is bringing to market the Cisco Aironet 2700 Series Access Point. What it offers is a tremendous amount of power at a very attractive price point.
We all know Wi-Fi is here to stay and is expanding all around us rapidly. That need for speed is exciting. But what does that mean? Not everyone feels comfortable being on the cutting edge. Many of our customers are not as concerned about chasing the future and have more limited budgets that they hesitate to put down for the best AP knowing there are lower priced options. At the same time, everyone is aware technology moves ahead with or without you, so they don’t want to give up lot of the new capabilities by going totally to the other extreme of not upgrading at all. What they want is something that’s going to last for a while that gives them the advantages available today, but not have to invest a lot to get it. I equate this to buying something like a car. A year ago when I was in the market to buy a new car I didn’t want to sacrifice whole lot of options but if there was one or two options that I could give up in order to save a bit of money, I was okay with that.
This is similar to what Cisco is offering with Aironet 2700 Series. Customers have to choose something that they can utilize in their network that is better than any of the competitive solutions out there, truly built-for-purpose, sleek design on the outside yet tough on the inside and very powerful. Read More »
Tags: 11ac, 11n, 2.4 GHz, 2700, 802.11, 802.11ac, 802.11n, access point, aggregate throughput, AP, application, ASIC, built-for-purpose, chipset, Cisco, client, ClientLink, collision, data rate, GHz, HDX, infrastructure, latency, maximum, mbps, memory, memory contention, network, network processor, offboard, onboard, Packet, packet processing, performance, purpose-built, radio, RAM, rf, scale, silicon, smartphone, tech, technology, throughput, wi-fi, wifi, wireless
As more and more 802.11ac devices come to the market this year, businesses need to make sure the best possible 802.11ac wireless infrastructure gets deployed to make sure those 802.11ac end points are performing at both the best possible data rates and application throughputs to maximize the move to 802.11ac.
Cisco’s Aironet 3700 with HDX Technology does just that. If you’re thinking that the 3700 is just another 802.11ac AP, think again: not all 802.11ac AP’s are created equal.
To demonstrate this, let’s take a Cisco 3700 access point..
When you open a Cisco AP, you will see dedicated memory (RAM) on the radio chipset itself (one on the 2.4 GHz radio, another on the 5 Ghz radio) to ensure the RF packets get processed “onboard” each radio instead of “offboard” in order to reduce latency and any packet processing collision from memory contention on the AP. Additional packet processing can be handled on the “offboard” memory that is part of the network processor portion of the AP platform as well. This unique, innovative ASIC-based Wi-Fi chipset by Cisco exemplifies the built-for-Purpose design, and is the hallmark of Cisco’s 3700 Series AP.
Contrast this with the competitive landscape that claims to be Purpose-Built, but in reality is leveraging off-the-shelf merchant silicon-based 802.11ac WiFi chipsets. Read More »
Tags: 11ac, 11n, 2.4 GHz, 802.11, 802.11ac, 802.11n, access point, aggregate throughput, AP, application, ASIC, built-for-purpose, chipset, Cisco, client, ClientLink, collision, data rate, GHz, HDX, infrastructure, latency, maximum, mbps, memory, memory contention, network, network processor, offboard, onboard, Packet, packet processing, performance, purpose-built, radio, RAM, rf, scale, silicon, smartphone, tech, technology, throughput, wi-fi, wifi, wireless