This will probably date me, but when I started in telecommunications a few years back, 10-Mbps thin-net Ethernet was the cool new technology. I used to think, “Who would ever need that much bandwidth?” Since then, IT technology has changed dramatically, with applications continually demanding more and more bandwidth. Ethernet switching capacity has advanced by leaps and bounds to keep pace with demand, ratcheting up connectivity speeds from 10 Mbps to 100 Gbps. For most data centers today, 1- and 10-Gbps connectivity is commonplace. But now 100 Gbps is quickly gaining traction in the vertical markets that require the highest performance. If history is any indication, 100 Gbps will be commonplace in most data centers in the not too distant future.
To meet the high-performance demands of today’s service providers, research labs, and large enterprises, Cisco started shipping a 12-port 100-Gbps module for the Cisco Nexus® 7700 platform switches about a year ago.
Nexus 7700 100G, 12 port module
The module is based on the Cisco® F3 chip, which offers the industry’s most comprehensive data center feature set for the core and the data center interconnect, including multicast, Multiprotocol Label Switching (MPLS), Virtual Extensible LAN (VXLAN), Cisco Overlay Transport Virtualization (OTV), and Cisco Locator/ID Separation Protocol (LISP). The module was designed to deliver line-rate performance with a total switching capacity of 1.2 terabits per second (Tbps). So, theoretically, a fully loaded Cisco Nexus 7700 18-Slot Switch chassis with 192 100-Gbps ports could deliver up to 38 Tbps of bidirectional throughput. No matter how you slice and dice it, that’s a lot of throughput—enough to meet the demands of any network.
As a matter of fact, 192 100-Gbps ports with 38-Tbps throughput would make the Cisco Nexus 7700 18-Slot Switch the industry’s highest-density 100-Gbps switch, with the Industry’s highest throughput rate. To verify this industry leadership, we put the switch to the test.
Cisco commissioned Miercom to conduct an independent performance test on a fully loaded Cisco Nexus 7700 18-Slot Switch with 192 100-Gbps ports. One of the first challenges Miercom faced was how to generate 38 Tbps of traffic to test line-rate performance. Miercom called upon Ixia’s world-renowned labs to help. The solution called for multiple Ixia Xcellon modules in Ixia’s iSimCity lab to conduct the tests (Figure 2).
Figure 2. Ixia test lab set up
Along with the raw throughput testing, Miercom also tested critical Cisco Nexus 7700 platform features to see how they would perform under full load. Tested features included MPLS, IPv4/IPv6 multicast, and hitless In Service Software Upgrade (ISSU).
After the testing was complete, the Cisco Nexus 7700 18-Slot Switch proved that it offers the industry’s highest 100-Gbps density and performance with line-rate services and exceptional availability. This level of scale provides customers with many years of investment protection as they transition from 1 and 10 Gigabit Ethernet to 40 and 100 Gigabit Ethernet architectures in the future.
Robert Smithers, CEO of Miercom, summed up the test results nicely: “Miercom independently exercised and evaluated Cisco Systems Nexus 7718 and was frankly stunned by the incredible power and throughput of this system, coupled with consistent low latency and latency variation, as well as solid MPLS support and high-availability features. The first switch we have independently tested with 192 x 100GE ports, the Cisco Nexus 7718 is awarded Miercom Performance Verified in our ongoing Data-Center-Class 100GE Switch Study.”
For the full details of the test, check out the comprehensive Miercom test report and accompanying test/results video. Also, here is what Miercom and Ixia had to say in their press releases.
Here’s a quick summary of the Miercom test results:
Testing found the Cisco Nexus 7718 can forward at line rate on all 192 of its 100GE ports – delivering over 38 Terabits/s of bidirectional traffic
Testing confirmed the Cisco 7718 can distribute real-world IPv4 and IPv6 multicast traffic at wire speed, with each of 191 receiver ports handling 1,250 IGMPv2 groups
The Cisco 7718 can process real-world MPLS traffic at line rate on all 192 of its 100GE ports, with no loss and low latency
Testing confirmed that an active Supervisor module or fabric module can be replaced with no packet loss, with IPv4 and v6 traffic running at high capacity on all of its 100GE ports
The Cisco 7718 executed an in-service software upgrade, with IPv4 and v6 traffic running at high capacity on all 192 of its 100GE ports, with no loss
It is nearly five years since Cisco, EMC and VMware came together to set up VCE and introduce one of the world’s best integrated infrastructure solutions with the Vblocks. The promise was to deliver “dramatic efficiencies” to customers promising significant reduction in capital and operating expenses giving customers flexibility and choice. Customers appreciated the operational simplicity of the model and Vblock sales took off with multi-billion dollar annual run rates.
Much has changed in the industry since then. The social-mobile-cloud-big data revolution has accelerated posing new requirements for IT and increasing the relevance of data centers and private cloud deployments. SDN has moved from being just a buzzword with several use-cases. Server virtualization has continued to drive efficiencies and hybrid clouds have become the new norm. Amidst all this, customers continue to crave operational simplicity and consumable infrastructure for their data center and private cloud deployments making the VCE approach as relevant as ever.
So, today, we’re very happy to share the success and celebrate the joint innovations as VCE rolls out its next generation Vblock systems that drive new levels of convergence. With Cisco continuing to refresh its portfolio with new Nexus products and industry leading SDN with the Cisco Application Centric Infrastructure (ACI) approach, and with Cisco UCS introducing next-generation products, it is natural that these innovations be reflected in the VCE Vblock integrated solutions.
Cisco is helping bring in new innovations to the party. The Nexus 9000 forms a key element with a very compelling form factor and industry leading price-performance. For customers interested in venturing into Software Defined Networking (SDN) and making their infrastructure application centric, the Application Policy Infrastructure Controller (APIC) provides a central point of management and policy application. The result is a simplified operational model and lower TCO across a variety of deployment scenarios.
As VCE introduces Vblock Systems 240, 540 and 740 today, they provide the flexibility of consuming the network elements as standalone switches or SDN deployments in an ACI mode. Vblocks can therefore operate in a standalone mode with current automation mechanisms or in an ACI ready mode subscribing to the APIC policy-driven model. Customers adopting the new Vblock systems get the operational flexibility to choose.
Five years ago, VCE was created with the goal of providing a simple, efficient solution to deploy and run IT infrastructure. VCE’s Vblock Systems have enabled customers to focus on business innovation instead of integrating, validating, and managing IT infrastructure. It would be an understatement to say VCE has been successful. Last year, Vblock Systems, built on Cisco UCS integrated infrastructure, surpassed their 2013 goal of $1 billion in annual sales and was recognized as a leader inthe integrated infrastructure market. In fact, in Gartner’s inaugural Magic Quadrant for integrated systems, VCE Vblock Systems is rated in the Leaders Quadrant, based on the tight integration of industry and market leading technologies from Cisco and EMC.
Today, VCE announced a major update and expansion to their Vblock Systems portfolio using the latest Cisco UCS servers and Cisco ACI-Ready switches. The new Cisco M4 model servers recently celebrated four world-records benchmarks, offering performance improvements up to 145 percent since the last processor generation. Customers can be confident that Cisco UCS servers will deliver outstanding application performance as part of a Vblock System. IT leaders want to accelerate infrastructure and application deployment and these new ACI-Ready Vblock Systems are an extension of Cisco’s application-centric data center strategy. We feel our application-centric approach, where the automated configuration of IT infrastructure in sync with the needs of the application, is essential to keeping pace with todays dynamic business priorities.
VCE also announced a cloud management solution with Cisco UCS Director. VCE’s Integrated Solution for Cloud Management with Cisco pre-integrates UCS Director with a Vblock System, providing the capability to quickly instantiate an initial private cloud foundation for customer environments. UCS Director enables the automation and provisioning of compute, network, and storage resources, both physical and virtual. This automation of integrated infrastructure can further expedite the deployment of application-ready infrastructure.
Cisco is excited that our new products and technologies have been integrated into the Vblock portfolio and congratulate the VCE team on today’s announcement. We believe these new Vblock Systems and solutions will make it easier for customers to deliver the performance, agility, and availability for the most demanding applications.
Are you looking for a reasonably priced, yet powerful, flexible SAN solution?
Cisco MDS 9148S Multilayer Fabric Switch, a new 16G Fibre Channel SAN solution for small to medium businesses. This switch is powerful and flexible, with up to 48 autosensing line-rate 16G Fibre Channel ports and comprehensive enterprise-class features in a compact one–rack unit form factor. Plus, with an affordable price, the Cisco MDS 9148S brings the power of 16G Fibre Channel to a new level of value.
Join our next webcast(8-Oct-2014 08:00 AM PST) and learn more about the technical capabilities, design considerations, and best practices of implementing small SANs. You will also learn how to grow your SAN transparently. See use cases, including designing small fabric, core-edge design, and migrating from 8G to 16G.
Watch this video as our experts demonstrate plug-and-play features and simple setup of MDS Fabric Switches with Device Manager
That is the approximate number of cloud services that Ken Hankoff, Manager of Cisco IT Risk Management’s Cloud and Application Service Provider Remediation (CASPR) Program believes Cisco’s 70,000 employees use. For the last 14 years, this program has assessed and remediated risks associated with using a cloud-hosted service.
An assessment process for new cloud services is a vital step toward reducing the risk of using externally hosted services. Many customers I speak with struggle to rapidly assess cloud services and integrate them into their IT organization. As part of my blog series on governing cloud service adoption, I asked Ken to share some of his ‘lessons learned’ in assessing the risks of cloud services and bringing them into Cisco IT’s fold.
How do you ensure that teams wanting to use new cloud services work with your team?
Our team is not in the business of sourcing cloud vendors. That responsibility lies with the individual business units and their architecture teams who are seeking to use the service, often in partnership with IT. Once a vendor is selected, there are two primary ways in which my team gets engaged. First, through the Global Contracts team as they have made Cloud Service Provider assessment a part of the contracting process, and second when a new service is being integrated within IT.
How do you evaluate whether a new cloud service is risky to the business?
We look at seven risk factors to create a formula for risk—business criticality, financial viability, security, resiliency, architectural alignment, regulatory compliance, and assessment status.
We establish the business criticality of the service to determine how Cisco would be impacted or disrupted in the event the capability provided by the vendor would go away, and whether we could react or compensate.
We then look at the financial viability of the vendor to give us comfort that they will remain in business. To evaluate vendors we leverage Dunn & Bradstreet’s Predictive Scores & Ratings. We rely heavily on Cisco’s Information Security (InfoSec) organization to provide us with a Security Composite Risk score. Depending on the parameters of the cloud provider engagement, InfoSec will look at the vendor’s application development process, infrastructure, data handling security, system-to-system interoperability, and other areas. For resiliency we focus on how they meet our standards around business continuity and disaster recovery to ensure that our business data will be there when needed, regardless of what happens.
We also need to ensure that we stay compliant with regulations. A vendor that has to comply with HIPAA, SOX, or other regulatory/privacy requirements poses a higher risk than one that doesn’t. For this reason, we look into whether regulatory compliance is a factor, and if so, that it is addressed appropriately.
Finally, we also assess if the vendor aligns to the broader architecture that Cisco IT is investing in to support the business. Vendors are deemed higher investment risk if they do not align to the business and operational roadmap that Cisco is pursuing.
We re-asses vendors on a periodic basis according to their overall risk score. If a service is overdue for a reassessment, that in itself increases the risk of doing business with the provider, so we factor it in.
In your opinion, what are the three most important things to manage the business risks of cloud services?
First, I would suggest establishing ownership and governance of cloud services via a centralized PMO at enterprise level, not just within IT. This ownership needs to go beyond just assessing vendors for security risk, and focus on establishing company-wide policies for overseeing cloud services at the enterprise level.
Second, provide visibility into existing services and how they are being used. This helps enable a catalog of assessed and approved vendors for people to access. If you can have fewer vendors being used, you can reduce your risk.
Third, continually monitor services across the board to know what risks we might be facing, and ensure that the service providers are meeting their SLAs. Additionally, this helps to ensure that investments aren’t being wasted. There is a natural CSP application lifecycle – selection, implementation, adoption, and eventually that service usage might decline and you may end up supporting something that has very few users if you don’t have a lifecycle approach to phasing out services.
What is your biggest lesson learned in assessing new cloud services?
I wish the program had collected more metrics earlier. What we are finding is that there are a significant number of services being contracted all over the company. By collecting really good metrics we might have been more effective in showing executives what services are being used, who is using them, and how. We are making good progress on this now, but I wish we started earlier.
How are you monitoring cloud services and gathering this intelligence?
Our professional service team has helped us a great deal. With the Cisco Cloud Consumption Services, we have begun to capture an enterprise view of what cloud services are being used, by whom and have a great dashboard of metrics we can now use to inform Cisco executives. I never imagined before we were using the software that we had nearly 2,000 cloud services in use, but with Cisco Cloud Consumption we now know and can monitor activity.