Today’s traditional network infrastructure resembles a tree. In the data center, network infrastructures generally include an access layer, an aggregation (distribution) layer, a core layer, and an edge services layer. Connectivity fans out like branches from the core, through these multiple tiers of switching, to connect the myriad devices in the data center. This tree structure topology exists in almost every data center in the world. Where did this architecture come from?
It originated with the local area network.
The Ethernet switch was originally designed to solve the LAN problem. If you’ve ever worked with coaxial cables in a local area network, you’ll remember what a nightmare it was. The Ethernet switch was a major breakthrough, greatly simplifying network deployments and becoming the basic building block of all subsequent networks. The tree structure was created to provide the level of fan-out required to support and connect all the clients in the local area. At the top of the tree was the workgroup server; almost all network traffic moved “north and south” between the clients at the bottom of the tree and the workgroup server at the top. As the workgroup grew, the tree got bigger.
Eventually, as networks grew to incorporate remote locations and evolved from local to wide area networks, the workgroup servers were moved back into the data center where they were easier to manage. At the same time, Ethernet became reliable enough, fast enough, and cheap enough to displace alternate technologies such as SNA, DECnet and Token Ring in the data center network itself. When this happened, it became possible to take the same topology that had proved so successful in the local area network and apply it to the data center in a multi-tier tree structure. For the most part, this architecture worked fine in a client/server world since most traffic ran north and south, between the server and the client.
However, with the advent of Service Oriented Architectures, or SOA-based applications, there was a fundamental change in the traffic patterns of the modern data center. With the advent of the web browser, a piece of the client-side processing was pulled back into the data center while at the same time the server-side application was disaggregated. As a result, what was once a fairly monolithic application became a set of federated services interconnected by the network, enabling greater application scalability and flexibility. Not only did this dramatically increase the number of servers, it also fundamentally changed data flow patterns in the data center. Traffic that was once expressed via an internal IPC mechanism communicating through the memory of the server was now exposed as network traffic between servers. This was further exacerbated when the storage became virtualized and traffic that was once contained on an internal SCSI bus was now expressed over a network.
Thus, whereas 95% of network traffic within the client/server data center was north-south, today as much as 80% of network traffic is now east-west. And while traffic out to a client interfacing with a human can tolerate a certain amount of latency, this is not true of east-west traffic. In fact, traffic between servers and between servers and storage is extremely sensitive to latency, which has a direct impact on the delivered behavior of applications. Add the fact that there has also been an exponential increase in data traffic on data center networks and it becomes clear that it is time to rethink the legacy tree structures.
At Juniper, we believe that a fabric, not a tree structure, is the ideal topology for the data center. Check out this animation to learn more about how Juniper’s approach can allow you to build a more efficient data center and drive your business model better:
I am sure many of you have either read or know of the concepts outlined by Thomas L. Friedman in his book The World Is Flat. It seems more and more that Friedman’s point of view is a reality of modern life. However what is even more fascinating to me is that just as technology enabled an ever flatter world, this changed world is now creating new pressures and requirements on technology solutions. One such example is a renaissance in the world of Networking: “Flat is back!” Read more...
Late last year, I had the opportunity to participate in an executive roundtable with Phil Sayer of Forrester Research and Andy Bach of New York Stock Exchange (NYSE) Euronext. The conversation included an interesting mix of industry-wide observations from Phil on cloud computing, a description of Juniper’s vision for building cloud-ready data center networks from me, and Andy’s experience of designing and building the next generation exchange network.
Phil outlined how cloud computing and server virtualization will drive the future of data center networking. My takeaway was that regardless of company plans for public cloud services, evolving their data centers into a private cloud will deliver significant business benefits like greater agility and lower costs.
Andy spoke about the industry-wide mandate to reduce or even eliminate network latency – which represents a shift from the previous focus on primarily network availability. NYSE Euronext has nearly 3,500 listed companies with a market capitalization of more than $12 trillion. Today, a trade transacts in microseconds, and there may be hundreds of quotes per trade. In 2005, the value of the trades transacted exceeded the market capitalization of the listed companies; this meant a big increase in the turnover rate of shares. In 2008, the value of trades transacted nearly tripled the market capitalization. This turnover increase is driving over 120 trades per millisecond! Obviously, the exchange has a critical need to protect itself from malicious attacks and to record and store all transactions for seven years to comply with federal and international regulations.
One logical question is “What do my business and technical needs have in common with NYSE Euronext, and how are they related to cloud computing?” The answer is “a lot,” in terms of network availability, latency, performance and compliance.
At Juniper, we offer a pragmatic process to make your data center network cloud-ready: simplify, share and secure. Our simplification strategy flattens the network to remove latency, improve server-to-server performance and lower costs. Our process of sharing extends Layer 2 domains between global data centers to enable mobility with live migration, improving disaster recovery and application availability. And our security process evolves policy enforcement from traditional physical separation to identity- and application-based enforcement, with a single tool for administering policy and monitoring compliance.
Anyone who says “If it’s good enough for NYSE Euronext, then it must be good enough for me” is more right than they know. The high-availability, low-latency and compliance benefits NYSE Euronext will achieve with a simplified two-layer data center network design can help you build a cloud-ready data center network too.
Below are some videos of the event. Let me know your thoughts by leaving a comment and I’ll be sure to get back to you.
VIDEO ONE: Introduction
Phil Sayer presents his view of the current marketplace and how the recession is affecting budgets and staff. Andy Bach considers the ever increasing pressure on IT to deliver results, even in the face of decreasing budgets and staff. I lay out how Juniper is helping customers address these financial challenges with high-performance networks.
VIDEO TWO: Virtualization
Phil explains why he thinks that Virtualization is now. Andy Bach explains NYSE’s adoption of virtualization in the context of the real-time data environment and how virtualization has helped NYSE Automate back office processes I discuss the new requirements virtualization has placed on the Data Centre and how it is changing the fundamental design and architecture.
VIDEO THREE: Ethernet
Is Ethernet the future for storage? Phil says yes, because it simplifies the switching architecture and reduces cost but, he explains how legacy Fiber Channel Storage Networks will be around for a while longer. I explain how Convergence Enhanced Ethernet and Fiber Channel over Ethernet standards are helping the deliver Ethernet Storage Networks and I introduce iSCSI. Andy Bach discusses the cost advantages of delivering a unified fabric within the data center using 10gE.
VIDEO FOUR: The Cloud and Data Centre Architecture
Phil sets out the landscape for Cloud Networking and discusses how developments such as Infrastructure as a Service (IaaS) and Dynamic and Shared Services is impacting the Architecture and the Security of the Data Centre. Andy explains how Internal Cloud Computing is reducing time-to-market in Financial Services. I talk about the adoption of the Cloud within both Service-Providers and Enterprises.
VIDEO FIVE: Business Best Practices and Dual Vendor Strategy
Is a dual vendor strategy a good initiative? Phil outlines the case; based on competitiveness and innovation. Andy Bach discusses the NYSE approach to the second vendor strategy to ensure their network diversity and innovation. I discuss how Juniper can mitigate the risk for an organization adopting a dual vendor strategy by ensuring Interoperability between vendors.
Exploring the vision for the networking industry and the issues shaping its future.
Vice President, Business Strategy and Marketing
Software Solutions Division