Server Virtualization 3.0 And Load Balancing
April 21, 2010
Server Virtualization 1.0 started when VMware first appeared on the radar of data center managers. Back then, the software had to interoperate with existing infrastructure or it didn't get deployed. As the popularity of VMware increased, priorities changed and we moved to Server Virtualization 2.0. VMware, Hyper-V and Xen server virtualization moved to the center of data center architectures. Data center managers suddenly required other infrastructure to work with the server virtualization software or they wouldn't deploy it. No wonder every vendor on the planet now claims to work great with VMware and spends a ton of money exhibiting at VMworld events around the world.
The first two phases of server virtualization helped IT organizations increase the utilization of their server hardware. Server Virtualization 3.0 has just arrived and it's defined by an entire ecosystem of server, storage, networking and software vendors delivering infrastructure that's integrated with server virtualization platforms to optimize the reliability, performance and security of applications running on virtual machines. By the end of the Server Virtualization 3.0 phase, IT will manage the capabilities of all data center infrastructure on a per virtual machine basis. What makes tight integration between virtual server software and other infrastructure possible is well-documented application programmer's interfaces and software developer kits. For example, VMware provides APIs and SDKs that address host configuration, virtualization management and performance monitoring; server hardware health monitoring and storage management; extending the vSphere Client GUI; access and manipulation of virtual storage; obtaining statistics from the guest operating system of a virtual machine and scripting and automating common administrative tasks.
If you don't have hands-on experience, you might find it difficult to sort out what infrastructure will most affect your server virtualization deployment because even printer-ink suppliers are claiming to make your virtual environment work better. However, there are some products that, with tighter integration, can make a big difference. Best-in-class load balancing systems from Cisco, Coyote Point and F5 are great examples.
For IT organizations that support HTTP and secure HTTP applications, FTP, Email, DNS, or other TCP or UDP based protocols; layer 4 through 7 application load balancing and acceleration systems are designed to help IT provide a faster, more reliable experience to their customers and employees.
In secure application environments such as e-commerce, banking and medical, top-notch load balancers include an SSL offload acceleration chipset that can support over 10,000 HTTPS transactions per second. You can also expect best-in-class load balancers to include match rules for content-based load balancing, cookie persistence, URL rewriting, redirects from HTTP to HTTPS or from one URL to another, SSL offload and certificate management, as well as a compression engine to reduce bandwidth utilization for web based applications.
In virtual server environments, keeping up with the real time changes happening on virtual servers requires that the load be distributed across servers in an efficient manner. Furthermore, it must be distributed based on server latency. As server performance diminishes, the load must dynamically be distributed to other servers in a transparent manner. With load balancers, managing this change is automatic. Once servers are identified and configured, load balancers automatically distribute network traffic to a pool of virtual servers. If the pool capacity is uneven due to some servers having to share their capacity with other busy virtual machines, the load balancer automatically detects the uneven latency and sends the new traffic to more responsive servers.
The most advanced load balancing solutions include unique technology that allows VMware servers to provide proactive load balancing metrics. These load-balancers use the well-defined VMware API to query VMware's Virtual Center to gather CPU load, memory utilization and other status information for virtual servers running under VMware. They combine this information with their own network performance statistics and they use it to make intelligent decisions about server availability and load. For instance, you can automate responses once a certain network event or threshold is reached. Or you can power up or power down a VMware server based on network traffic. Being able to control VMware servers based on network traffic allows you save on power, cooling and management cost of running the data center.
Basic load-balancing is needed in the data center to provide a faster, more reliable World Wide Web and Application experience. Advanced load-balancing is a key component of an energy efficient, dynamically provisioned virtualized environment.