Intel next year will plant itself square in the middle of the budding market for systems that speed network traffic by rolling out something called I/OAT.
I/OAT stands for I/O Acceleration Technology and it will be previewed for the first time at next month's Intel Developer Forum (IDF) in San Francisco. Intel remains cagey about what exactly I/OAT is, but it dangled a few details today in front of reporters ahead of the IDF event.
"This will solve some of the problems we have heard from a lot of enterprise server customers," said an Intel spokesman.
The problems in question relate to network slowdowns that occur as servers try to crunch through the TCP/IP stack. Customers often find that their servers spend an inordinate amount of time dealing with network traffic when they should be hammering away on application data. To that end, a number of companies such as Adaptec and Alacritech have developed TOE cards or TCP/IP Offload Engines. These cards plug into servers or storage systems and handle much of the network traffic.
Intel plans to sidestep the need for separate TOE cards by building this technology into its server processor package - the chip itself, chipset and network controller. This should reduce some of the time a processor typically spends waiting for memory to feed back information and improve overall application processing speeds.
Intel expects the technology to debut as a standard part of its server products in 2006. It will also have a tight storage component to the I/OAT project but won't say word one about what that storage component may be. Companies have typically tried to handle the TCP/IP stack problem with networked storage systems.
Thus far, TOE cards haven't proved terribly popular in part because of their relatively high-cost and lack of widespread support. Companies such as Sun Microsystems have talked about developing their own TOE card-style technology and about dedicating parts of multicore processors to handling TCP/IP and security traffic.
Intel is increasingly pointing to new inventions such as I/OAT to pull attention away from its failings in other areas of the chip-making game. The company has decided that GHz are now less important than the "platform" or total package of processor and accompanying technology such as wireless support or partitioning tools.
Intel, however, has seen all of its major rivals race ahead with more crucial parts of the processor "platform" like dual and multicore chips and 64-bit support.
Intel vowed to talk up multicore chips and 64-bit processors extensively at IDF along with new items like I/OAT. Better late than never. ®