Will Cisco’s 'Big Switch' Provoke a New Technology War?

In January Cisco Systems (CSCO) announced Nexus 7000, a major “rethink” of the corporate network switch. To be honest, it’s better not to think of the Nexus 7000 as a switch; a switch can be something you turn a light on with. Instead, think of it as a cage full of network blades that just happens to be the corporate network in a box. And by the way, it’s a bewilderingly fast network too. If you have a wide enough pipe into “the cloud”, it could download and distribute the whole Wikipedia in a hundredth of a second, or 100,000 movies in just over 6 minutes or the whole of the searchable Internet in 7.5 minutes.

Technically (skip this paragraph if you don’t speak Geek), the switch is designed for 10 Gigabit Ethernet networks, but allows for the future availability of 40Gb and 100Gb Ethernet and unified fabric I/O modules, and it scales beyond traffic rates of 15 terabits per second. And, get this; it’s powered by an OS (or perhaps we should call it a NOS) the Cisco NX-OS, which is there primarily to ensure continuous system availability and network flexibility.

In summary: The network is a computer.

And if you think that’s just a smart-ass bit of word play: it’s not.

A World Turned Upside Down

Cisco’s vision, which can become reality with the Nexus, is of a data center that is no longer defined by computer architecture, but by network architecture. This makes sense on many levels. Let’s list them in the hope of making it easier to understand.

1. Networks have become so fast that in many instances it is practical to send the the data to the program, or to send the program to the data, or to send both the program and the data somewhere else to execute. Software architecture has been about keeping data and process together to satisfy performance constraints. Well Moore’s Law reduced the performance issue and Metcalfe’s Law opened up the network. All the constraints of software architecture reduced and they continue to reduce. Distributing both software and data becomes easier by the year.
2. Software is increasingly being delivered as a service that you connect to. And if it cannot deliver the right performance characteristics in the place where it lives, you move it to a place where it can.
3. Increasingly there is more and more intelligence being placed on the switch or on the wire. Of course Cisco has been adding intelligence to the switch for years. Those Cisco firewalls and VPNs were exactly that. But also, in the last 5 years, agentless sotware (for example some Intrusion Detection products) has become prominent. Such applications simply listen to the network and initiate action if they “don’t like what they hear”. The point is that applications don’t have to live in server blade cabinets. You can put them on switches or you could put them onto server boards that sit in a big switch cabinet. They’re very portable.
4. The network needs an OS (or NOS). Whether Cisco has the right OS is a point for debate, but the network definitely needs an OS and the OS needs to perform the functions that Cisco’s NX-OS carries out. It also needs to do other things to like optimize and load balance all the resources in a way that corresponds to the service level needs of the important business transactions and processes it supports. Personally, I do not see how that OS can do anything but span the whole network - including the switches.

From this perspective, the most important feature of the Nexus 7000 is its “fabric” which is designed to orchestrate data center resources (not just the network). That means; servers, storage resources and applications. And when I use the term data center here, I’m not talking about a single building or single site, because the Nexus can thread multiple sites that are thousands of miles apart together and, in theory at least, run them as a unit. It’s not prohibitively expensive either. The Nexus and its software start at $75,000 and Cisco puts the cost to a typical company at about $200,000.

Cisco is either very lucky or it has excellent timing, because the Nexus has been launched at precisely the time that virtualization has become the “trend du jour” and the virtual resource spaces that virtualization creates are in serious need of global management. With the Nexus, companies should be able to configure given combinations of bandwidth, processing, storage and software.

As far as I know, no-one else can do this - because they cannot allocate/virtualize bandwidth. (and if you think that bandwidth allocation is not a factor in performance, you are wrong).

The Big Switch

Of course Cisco is not the only network player in the game. Juniper Networks and Brocade will undoubtedly be introducing products that are competitive to the Nexus soon. But the big question is: How does this affect IBM (IBM), HP (HPQ), Dell (DELL) and Sun (JAVA)?

In the grand scheme of things, if these vendors cannot produce switch technology, then in the long run they are reduced to the role of component providers to the data center. Because even a big bad blue tyrannosaurus mainframe is just a component as far as the Nexus is concerned - and a huge SAN is “just another storage resource”.

And that’s the whole point. Cisco’s switch is game changing and it will surely provoke a new technology war between some of the giants of the industry. For a while, of course, it will be a phony war, until the early adopters of the “big switch” have implemented it and it becomes clear what difference it actually makes.

At the beginning of January, I published some forecasts including one for the communications sector: Forecasts for 2008: #5 Communications Convergence. As far as Cisco is concerned, I’m already wrong. I was suggesting that the main communications vendors would need to “go soft”. I’m now of the opinion that Cisco does not need to go soft in the way I suggested. It has already gone soft by developing a NOS and a fabric, and that is enough of a strategic differentiator to allow to continue to pump iron for years.

Disclosure: none

Comments

[rhammel:]

Have you been paying attention to what's going on in the market? BRCD has announced and is actually shipping a much more impressive switch than the Nexus while all CISCO is showing is a cardboard mock up.

2008 Feb 26 07:20 AM

[User 156022:]

Please would you clarify, this switch is designed for 10 Gigabit Ethernet, but QLogic Corp. InfiniBand offers up to 20Gb of bandwidth switch that can be used to eliminate the need for separate physical server connections to storage and network resources, which I believe is the one that uses IBM and others
Thak you,
Blas

2008 Feb 26 08:50 AM

[Tony Rybczy...:]

You're right. High speed networking creates new opportunities for how and where processing is done; this in turn drives the need for ultra-reliable networking. Nexus falls short because it takes a box rather than network view of reliability. I conclude that the Nexus is no Lexus. See:
blog.tmcnet.com/the-hy...

2008 Feb 26 11:41 AM

[User 156258:]

Nice unabashed Cisco praise. You fail to mention the fact that the box is 8-to-1 oversubscribed and isn't event available. Foundry has a box that crushes the Nexus 7000. It has been out for awhile and is actually shipping today. Also, what does the release of the Nexus and its new OS mean for Cat 6500 customers? Can't be good.

2008 Feb 26 12:58 PM

[Robin Bloor:]

I'm impressed with all the feedback concerning competitive product. The article was never supposed to offer any kind of comparison between Cisco and its competitors (and it doesn't). Rumor has it that Cisco brought its Nexus announcement forward to steal thunder from Brocade. It may well turn out that some of Cisco's competitors outsmart and outmarket the giant. The whole point of my article/blog posting was to focus on the fact that the networking market is colliding with the server market to the advantage of the networking vendors. This is probably positive for all the networking vendors who can hold their own. It is not positive for the sever vendor.

2008 Feb 26 02:15 PM

[Douglas Gou...:]

I hope I addressed the performance questions that people are bringing up here: blogs.cisco.com/datace...

I do take some offense to those who would say that it is a 'cardboard cutout'. That is rather disparaging to the hundreds of engineers who have worked for years to build this product. If you went to NANOG you would have seend a working chassis, I will post some pics for everyone over the next couple of days though :) I know Network World got a few of my shots from the external customer launch.

If someone still really doubts me on this- shoot me an email... come on over to Building 7 and see a few hundred in test. That's what it takes in the Data Center. :)

dg

2008 Feb 26 05:54 PM

[Dadofamunky:]

The Nexus 7000 is the Cat 6500 successor. For now I think it will be sold as such; it has the faster re-architected fabric and a more extensible software architecture. Now, the Cat 6500 is one of the most successful products in networking history. Penetrating that market and getting people to re-focus on network virtualization will probably take longer than the sanguine estimates in this article. It remains to be seen how well users will accept having to learn yet another new command line. With enough value-add benefits, like some of the kinds of things discussed in this article, Nexus should have a good 10-15-year run. No one ever got fired for buying Cisco, Juniper and Foundry not excepted.

2008 Feb 27 06:20 PM

[Phil Dewey:]

The bottleneck always has been and always will be software. For all the lightning fast speeds and feeds one inefficient PL/SQL in a DB will render this gleaming hunk of titanium a frustrated, power burning noise maker. The network guys never look at a system holistically and hence will never solve the problem.

2008 Feb 28 02:18 AM

[Douglas Gou...:]

Just one thought as well, we modeled the CLI very very closely after our existing IOS CLIs, but made some changes when and where it seemed appropriate and there was substantial end-user input that we needed to make a change: i.e. moving to neighbor centric BGP configurations or adding configuration rollback capabilities. But if you are familiar with a Catalyst 6500 you should be able to sit down at a console on the Nexus 7000 and be productive in a few minutes. The areas of training and learning would be around new features and architectural capabilities.

Also- we posted the other weeks TechWise TV episode to YouTube and here is a link to the episode index. (ignore my section if you will, I know it will probably invite some fun commentary here :)

www.mytechwisetv.com/p...

dg

2008 Feb 28 11:27 AM