Welcome!

Virtualization Authors: Elizabeth White, Pat Romanski, JP Morgenthal, Sebastian Stadil, Adine Deford

Related Topics: Virtualization, SOA & WOA, Cloud Expo

Virtualization: Article

Managing Cloud-Scale Network Virtualization

Integrating Flexible, Standards-Based Switch Technology Increases Long-Term Network Value

Network infrastructure design has never faced more complex real-world demands. Fast, reliable connectivity is at the heart of business today, whether it is powering global systems, internal operations or customer-facing communications and services. Private and public cloud applications, complex usage models and scale requirements that may span literally billions of transactions have sparked a growing focus on scalable, high performance networks.

Private cloud networks installed in large enterprises are greatly influenced by legacy use models, whereas public cloud networks are characteristically greenfield designs, built from the ground up without legacy equipment or usage model restrictions. Public cloud network designs are generally chosen for their cost-savings, equipment vendor-agnostic, commodity-like scaling capabilities. The resulting network is designed for the future, with off-the- shelf, easily replaceable, and cost-effective network equipment. While the same performance demands pertain to private clouds, a public cloud’s focus on massive scale and support for multitenancy demands vendor-agnostic equipment and standards-based design approaches.

Virtualization is critical to meeting these cloud-scale demands overall – ensuring datacenter efficiency and keeping technology investments poised to evolve based on future demands. At the same time, virtualization is not a free pass for creating workable network infrastructures; the same efficient IT policies that apply to use of virtual machines in physical servers are relevant to network infrastructures as well. This is especially true of an enterprise evolving from a singular focus on operational efficiencies to a broader view of growth, flexibility and return on IT investments. Technology choices should address not only performance, but also footprint, power and long-term agility in transforming seamlessly to what’s next. Design flexibility is more important than ever, and greater value must be placed on flexible switch technologies that enable multitenancy as well as dramatic levels of scaling in support of virtual machines (VMs).

New Essentials of VM-Aware Switching

Historically, physical servers have been tiered for north-south data flow from access, distribution, and core layers to the wide area network and back again. Today’s network traffic moves in an east-west pattern – inherent in modern distributed systems and applications with data that travels across racks and pods. This methodology has resulted in rapid adoption of fast, fat, and flat network topologies and network virtualization in which deployments rely on sophisticated switching technologies to deliver maximum scale and performance.

VM-aware switching is ideal in these environments, which increasingly require native OS-based server level functionality without the performance hits sustained by virtualized servers running multiple VMs. Several standards-based switching technologies including SR-IOV (Single Root I/O Virtualization by PCI Special Interest Group) and VEPA/EVB (Virtual Ethernet Port Aggregator/Edge Virtual Bridging by IEEE 802.1) have emerged to address this issue, and ensure improved performance and scalability of applications that run in VMs. With intelligent switching solutions, such as Smart-NV technology enabled Broadcom switch solutions, virtual switch ports support link aggregation, queuing, ACL, statistics, and mirroring services similar to how those services are readily available for physical ports.

Maximizing Fast, Fat and Flat Using L2oL3 Techniques

Today’s fast, fat, and flat networks implement high-bandwidth, high-density, fixed-configuration aggregation and access layer switches connected in a spine-leaf model. Often interpreted as an L2 network (either physical or virtual), flat topologies span multiple pods or sites within a data center or even across data centers. TRILL (Transparent Interconnection of Lots of Links) or SPB (Shortest Path Bridging) technologies may be deployed to create scalable and large flat physical L2 networks with no multipath constraints.

When the underlying network is L3, a flat virtual L2 network can be attained using L2oL3 (layer 2 over layer 3) network virtualization technologies, such as VXLAN (Virtual Extended LAN) or NVGRE (Network Virtualization using Generic Routing Encapsulation). L2oL3 overlay techniques are well-suited to large-scale data centers that rely on L3’s proven scalability of L3 addressing and multipathing technologies. These techniques further extend the benefits of fast, fat, and flat architectures, enabling network virtualization at cloud-scale and proving essential for combining L2 and L3 network topologies. For network designers, L2oL3 technologies – and the sophisticated network switches that support the transit switch and gateway features needed for such overlay technologies – are crucial to today’s cloud infrastructures and the design flexibility that is required.

L2oL3-based network virtualization technologies eliminate the VLAN-based scaling challenges (limited up to 4K VLAN IDs) that exacerbate scaling in multitenant networks. They also promise to detach network virtualization- related configuration from physical switches, enabling software-defined networks across multivendor equipment — an attractive benefit for public and hybrid cloud deployments.

Technologies such as Smart-NV technology enabled Broadcom switch solutions support new and innovative L2oL3 overlay network technologies to enable network virtualization at cloud-scale. This meets cloud-scale requirements – including extending the scale of virtual LANs, as well as providing VM scale, network partitioning, and hybrid cloud enablement for multitenancy support. This type of deployment also allows efficient VM-based workload placement through live VM migration across pods or sites in a single data center or across data centers.

Keeping Networks Poised for What’s Next

Global data traffic is increasing exponentially – today’s levels are expected to rise 26-fold within the next three years. In the year 2015, millions of minutes of video will cross the network every second and the number of connected devices is expected to reach twice the global population. With this unprecedented growth, the highly scalable, virtualized network deployments expected to manage an incredible amount of traffic must rely on innovative switching technologies to enable the full potential of network performance-sensitive applications.

Where tiered and over-subscribed network architectures have successfully served the tiered server environments of the past, today’s deployments instead require virtualization and clustered applications in servers. The significant increase in east-west data center traffic patterns make it clear that network architectures must transform to enable the required performance demands of these applications. Network infrastructure in turn can be used to gain a competitive advantage – today and in the future – as more and more data is generated by public, private and hybrid cloud environments. Virtualization and switching technologies optimized for cloud computing have the potential to future-proof the network for years to come.

Network virtualization complements server virtualization in private and public cloud data centers, helping higher ROI and business performance through dynamic resource allocation. Network switches designed for such data centers must support multiple virtualization technologies. Driven by both legacy and new use models, comprehensive network virtualization technologies are needed in enterprise private clouds. Further, public cloud data centers are designed and engineered for multitenancy, scale, and cost-effectiveness, requiring a subset of such virtualization technologies. Ideal solutions will effectively meet all network infrastructure virtualization requirements – offering flexible comprehensive performance and scale for current and next-generation private, public, and hybrid cloud networks.

More Stories By Sujal Das

Sujal Das serves as Director of Product Marketing for Broadcom Corporation’s Infrastructure and Networking Group (ING). In this role, he is responsible for driving Broadcom’s Ethernet switch business in the data center and enterprise LAN market segments, and the development of product, ecosystem and strategy based on technology trends and application workloads. Das has extensive experience in semiconductors, networking software, data center network architectures, virtualization, and server systems. Prior to Broadcom, he served in senior product development and marketing roles at AMD, Marvell Semiconductors and Mellanox Technologies.

Das earned a BS EEE degree from the Birla Institute of Technology and Science in Pilani, India and an MBA from Santa Clara University. He has published and presented on numerous data center networking and virtualization related topics in well-known industry publications and events. He has been active in the open source community, driving multiple initiatives that have helped proliferate the adoption of high-performance networking in data center applications.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.