x
Optical/IP

Cisco Set to Detail XML Strategy

Competitors and at least one analyst say Cisco Systems Inc.'s (Nasdaq: CSCO) much-anticipated Application-Oriented Networking (AON) technology is scheduled to touch down tomorrow.

The announcement, said to be slated for Cisco's user conference in Las Vegas this week, represents the routing giant's plunge into XML processing, a Layer 7 technology that's already attracted a handful of startups and established XML players (see Telecom Startups Play in XML).

The rumored launch of Cisco's application-aware networking strategy has been circling among the XML-networking crowd, who say that's what they expect Cisco to focus on tomorrow.

Cisco officials have hinted that something's up, but they aren't revealing details. A media alert issued today says Cisco plans to "unveil new technology that adds intelligence to the network to support better business decision-making and increase productivity."

So what would the details entail? For one, there is a question of how Cisco will put XML routing into product form.

In a note published this morning, analyst Steve Kamman of CIBC World Markets says Cisco's announcement will center around an XML routing blade for the Catalyst 6500, Cisco's flagship switch, and software partnerships with the likes of IBM Corp. (NYSE: IBM) and SAP AG (NYSE/Frankfurt: SAP).

Separately, Kamman expects Cisco to announce a wireless and RFID module for the ISR line of service-creating routers (see Cisco Takes Apps on Board).

Previous reports note AON includes lots of home-grown technology but also incorporates an XML acceleration chip from Tarari Inc.

Better known as a data center technology, XML could become more important in telecom circles as the use of the markup language increases. This year, companies such as DataPower Technology Inc. and Solace Systems Inc. have announced XML gear targeted at the telecom space (see DataPower Flexes XML and Solace Unveils XML Message Router).

The goal behind XML switching is to handle entire messages at a time, without first distilling traffic down to IP form. Most players in this space are offering XML accelerators, which speed the processing of traffic in XML format.

Cisco's XML push is similar in philosophy to the Juniper Networks Inc. (Nasdaq: JNPR) acquisitions of Peribit Networks Inc. and Redline Networks Inc. (see Juniper Takes Two: Peribit & Redline). In both cases, the companies are pushing into higher layers of traffic processing, trying to improve the efficiency and effectiveness of data routing.

Cisco's AON will create a "choppier market environment" for the Juniper-Peribit-Redline combo and could have an effect on Layers 4 through 7 players, too, Kamman notes. In the latter camp, he names F5 Networks Inc. (Nasdaq: FFIV), Foundry Networks Inc. (Nasdaq: FDRY), Packeteer Inc. (Nasdaq: PKTR), and Radware Ltd. (Nasdaq: RDWR).

— Craig Matsumoto, Senior Editor, Light Reading

quick_or_dead 12/5/2012 | 3:10:07 AM
re: Cisco Set to Detail XML Strategy Oh, that's almost as bad as the end-to-end / bellhead argument in the other thread. XML squeezing IP is a little like saying that IP is squeezing Ethernet, they are not in conflict with each other -- they do different things.

As far as I can tell, all of the established xml networking products have IP addresses & ethernet ports.
dljvjbsl 12/5/2012 | 3:10:07 AM
re: Cisco Set to Detail XML Strategy With this announcment from Cisco, there is confrimnation that IP is being squeezed from both sides. XML at the edge and MPLS at the core are compressing the zone in which IP is useful.

We can look to see that IP will be squeezed out of exisitence jsut like ATM.
mr zippy 12/5/2012 | 3:10:06 AM
re: Cisco Set to Detail XML Strategy I agree.

MPLS may not rely on the IP header for forwarding, however the whole control plane of MPLS relies on it. Next hops are identified using IP addresses, IP is used to carry both the Fowarding Equivalence Class (commonly the IP subnet) and the label distribution protocols, and IP is also usually used to identify the end-nodes so that they can be catagorised into a matching FEC, which is then used to place the IP packets onto the corresponding MPLS Label Switched Path.

"XML Routing" won't get rid of IP, and neither will MPLS. Even if MPLS is used to switch IPv6 packets after the first hop and before the last hope, without looking at the IPv6 header, the end-nodes have to still be indentified, and there's still going to be more and more of them, hence, IPv6's large address space will still be useful, as will IPv6 oriented versions of OSPF, BGP etc, as they will be still necessary to distribute the FECs around the MPLS network.

There seems to be quite a bit of misunderstanding about how things will have to work around this "XML routing" space.
dljvjbsl 12/5/2012 | 3:10:01 AM
re: Cisco Set to Detail XML Strategy
Oh, that's almost as bad as the end-to-end / bellhead argument in the other thread. XML squeezing IP is a little like saying that IP is squeezing Ethernet, they are not in conflict with each other -- they do different things.


Ethernet addressing is a way of carrying packets from the host to the router. It has little other significance. IP squeezed out Ethernet in significance and therefore profit.

The IP squeezing argument follows the same path. IP will be used to carry packets between XML routers. and given the capabilities of MPLS in the core, it will be doing this only locally at each end) Teh significant addressing will be done in XML space. IPV6 won't be needed because the exponential increase in addressable things will be happening in XML space.

It will make little sense to give someone's toaster a network routable IPV6 address. It will be controlled locally by an XML address and so the IPV6 address space will just be a liability since it opens up internal devices to hackers. This is the major failing of the end to end argument
materialgirl 12/5/2012 | 3:09:43 AM
re: Cisco Set to Detail XML Strategy "Ethernet addressing is a way of carrying packets from the host to the router. It has little other significance. IP squeezed out Ethernet in significance and therefore profit. "

You already see this happening. EXTR and FDRY may pass packets really fast, but they cannot make money any more. All the action is in the higher levels, ala FFIV and the whole JNPR thing.

I question how CSCO can just add this type of processing to their boxes that already seem to strain under the load of new features.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE