Making Web applications faster is big business for big businesses-like carriers. Mid-market enterprises can also benefit from web acceleration, which is where Networking vendor Coyote Point is concentrating its efforts.
Coyote Point is now rolling out a new Equalizer GX line of Web Acceleration appliances that offer the promise of improved Web application speeds as well as the ability to directly handle VMware virtual servers to improve scalability and efficiency.
The new Coyote Point release comes as the application acceleration business itself continues to heat up with Citrix, Cisco, Blue Coat, F5, Juniper and others all vying for a piece of the pie that could be worth over a $1 billion dollars in 2010.
While Coyote Point includes its share of proprietary development and features into its Equalizer GX platform, the core platform sits on top of an open source FreeBSD operating system.
“We are using a modification of FreeBSD version 6 which provides for us the basic scaffold we need to build the appliance,”Bill Kish, CEO and CTO of Coyote Point, told InternetNews.com. “FreeBSD gives us the file system, an I/O subsystem and device drivers, Web server for our management interface and it gives us all sort of great open source tools and we use them to the fullest.”
FreeBSD 6 first came out in 2005 and has recently been superseded by FreeBSD 7, which hit version 7.1 earlier this month. Kish noted that Coyote Point will likely trail FreeBSD development by a year in order to make sure that the operating system is fully stabilized.
Kish added that Coyote also has contributed back to the FreeBSD project, specifically in the device driver area.
Though FreeBSD is at the core of the Coyote acceleration appliance, Coyote Point adds its own secret sauce to the mix as well. “When a packet actually comes into the device and it is destined for application acceleration or load balancing at that point it is picked up entirely into our code,” Kish explained.
“So we didn’t have to put effort into developing the other bits and pieces we rely on the FreeBSD community to do that for us. When the actual traffic management is involved, we optimize that and that’s where our core intellectual property is in understanding the application flows and how the protocols work,” he added.
FreeBSD is no stranger to the networking operation system world. Juniper Networks’ JUNOS was originally based on FreeBSD.
Kish noted that Coyote Point has been around since 1999, focusing on application traffic management, load balancing and acceleration. With the Equalizer GX platform they’re also getting into integration virtualization into the mix.
Coyote Point has developed something called Smart Controls for virtualization, which it includes with the Equalization GX. It allows an appliance to dynamically power up or down VMware virtual servers in order to meet Web application demand. The Smart Control additionally provides the capability to users to set specific rules for how the VMware servers are engaged.
As such, a user can specify a minimum or maximum number of always on virtual machines as well as the traffic conditions under which a server will be added or removed from the server pool. Citrix has a similar type of functionality enabled with its Netscaler appliance using Xen-based virtualization.
At the root though of Coyote Point’s existence and the web acceleration business in general is that fact that Web servers and applications aren’t as optimized as they could be.
“Even the TCP/IP stack can be a bottleneck on certain servers,” Kish argued. “If we can remove the need for a server to do compression or SSL
As opposed to other vendors in the Web acceleration business, Coyote Point is not placing a specific focus on particular application awareness. Blue Coat, for example, recently bought out application awareness vendor Packeteer specifically to have an offering that could identify and optimize different types of traffic.
Cisco also claims to be application-aware and has specific optimizations to improve virtual machine traffic among other types of traffic.
“We think of application in our world as data delivered over HTTP protocols,” Kish commented. “Our depth of understanding and manipulation is all based on HTTP. We don’t look into the guts of a stream of SQL