We’ve watched TCP/IP spread beyond anyone’s wildest dreams. Cisco (CSCO) has predicted that the network will connect 14 billion devices by 2010. This level of connectivity promises to make the network even more strategic to business operations in coming years, as Cisco's Chambers has suggested.
Yet there is a problem. While systems and endpoints have become more powerful, more dynamic and more mobile, network innovation has been focused on speeds and feeds, and layers of functionality requiring ever-increasing amounts of manual labor. Today’s network is essentially as static as it was decades ago, managed by ever growing ranks of “IT laborers” (responsible for configuring layers of appliances added on by various vendors to address larger and larger pain points, from security to application delivery) as the network grows.
This ironic world of manual labor within IT is about to evolve. It has to.
The Cobbler’s Kids Story and the Network Team
Ironically, the network itself may be the last bastion of IT to be automated and rendered dynamic. CIOs are learning that as innovation is focused on particular pockets within IT, the manual labor burden is merely shifted to other IT departments where automation has been neglected. As more devices are assigned IP addresses and networks become more complex, many enterprise networks have already reached a point of exhaustion and are being kept available by rising tides of increasingly costly manual labor.
The smoking gun was a recent Computerworld study of core network services, including IPAM, which found that management costs per IP address escalate as networks grow. Economists recognize rising per unit costs accompanying volume increases as a diseconomy of scale, a point where systems are breaking down. The Computerworld study also found that 2/3 of respondents were still using spreadsheets and manual labor to gain visibility into the status of their increasingly large and complex networks. This makes CIO visibility into the state of the network look a lot like CFO visibility into enterprise finances before the rise of Oracle and dynamic enterprise databases.
VirtSec was a Harbinger
Virtualization security was a warning shot across the industry’s bow, a signal that the strategic network had become to brittle, too static to support VMotion, one of the core powers enabled by the hypervisor. Few if any of the VMsafe security solution partners have managed to announce any new solutions to address the VMotion challenge, and that has throttled back the virtualization business case.
And the escalation of change enabled by increased flexibility only added more distance between what the CIO could see and what was actually happening real-time. That has driven pockets of complementary initiatives around change management and VLANs, which make virtual infrastructure less flexible, which is a kind of back to the future scenario.
Similarly, RFID/supply chain and wireless/VoIP initiatives have put more demands on increasingly inflexible and costly networks. As the gap between dynamic systems and static networks grow it sets up networking vendors for perhaps the most significant boom opportunity since the rise of TCP/IP.
The CIO “Shell Game” Sets the Stage
As IT vendors ply their wares they encourage “silo-think” or the over-marketing of new technologies that often merely shift costs and resources between systems and the network. As virtualization entered the data center, for example, the promise of VMs traversing hypervisors (and making systems more flexible and economical) turned into virtualization-lite, a throttled back flavor of VLAN spaghetti whereby the new network looked a lot like the old. As I mentioned, the network was too static to handle the power of the mobility enabled by the hypervisor.
It is easier to create a server and move it with virtualization, yet the network required the same manual configuration tasks to keep up. The lesson: system automation needs to be accompanied by network automation; or dynamic systems require dynamic networks. That is essentially the case for Infrastructure 2.0.
Without dynamic infrastructure, dynamic systems will produce less than stellar results and will increase pressure on the network and increase anxiety within network teams “doing more with less.” Because of this conundrum, network vendors and CIOs have a massive incentive to deliver on the promise of dynamic infrastructure, and the signs are encouraging.
I addressed this recently on a pod cast with Cisco’s James Urquhart and GigaSpace’s Geva Perry about the impacts of cloud computing on networks, yet the points raised could apply to any IT initiative automating systems (or making them more flexible or mobile). That may be why network pros get labeled as being against these innovations; they realize that automation in peer areas will simply drive more manual work in their direction. You could call it the “IT hustle.”
This prevalent need for network automation and dynamic connectivity intelligence will drive new markets and opportunities for those who led us to where we are today, including Cisco, F5 Networks (FFIV) and Microsoft (MSFT); as well as those investing heavily in or delivering new virtualization and cloud computing capabilities, including Google (GOOG), Amazon (AMZN), VMware (VMW) and Citrix (CTXS).
The dynamic infrastructure will be a strategic part of new investments as more enterprises hit pain thresholds that either increase operational costs or network downtime. It has to be, even if some of the cloud computing pundits caught up in the “dumb bus” illusion haven’t figured it out yet. If they don’t understand the strategic importance of network integrity to cloud computing there is a good chance for rain on their parade.
What’s in it for the CIO/CEO/CFO
Teams of networking pros now involved with manually tracking and/or configuring new devices and IP addresses - as well as manually configuring and updating the racks of network appliances - can focus resources on other more strategic activities, including the establishment of dynamic data-driven policies (like risk modeling and management and business case analysis) capable of driving sizable reductions in the total cost to operate a network while enabling more responsiveness free of manual configuration and tracking costs. The results: higher availability (even when a network is expanding) and reduced network TCO.
The CIO will have the same visibility of the status of the network as the CFO has on the status of the business. Spreadsheets will be replaced by dynamic core network services systems (and other innovations) that will do for the CIO what Oracle did for the CFO.
The business case for virtualization and enhanced network, application and endpoint intelligence will mushroom thanks to innovation in "connectivity intelligence". IT jobs will become less about manual labor and more strategic because the network is keeping up with changes on its own and is applying policies directly to endpoints and applications regardless of location and without ongoing and error-prone manual configuration challenges.
Network Vendors Stand to Benefit
As the business case improves for new network-enabled initiatives (thanks to the emergence on dynamic infrastructure) and the network continues to grow, a reduction in TCO could fuel massive innovation and adoption for new solutions from Cisco, F5 Networks, Microsoft, Infoblox, Juniper Networks (JNPR), Riverbed (RVBD) and other network players.
The network will reach into more nooks and crannies around the world and within the enterprise by automating even more processes with more tools and solutions. Cisco’s green data center initiatives could be given a sizable boost from dynamic infrastructure, because system automation would bring bigger payoffs.
As Cisco planned for this week’s C-Scape Global Forum, I had a chance to talk to Chris Wiborg on the Cisco enterprise infrastructure team and learn about Cisco’s very promising Service Oriented Architecture project aimed at designing more advanced services into the network. You can watch Chris' interview with Forrester's Robert Whiteley for more information. It’s pretty interesting.
F5 has similarly been blogging about the ability to further optimize application delivery by aligning dynamic networks with applications and the business; this is very much in sync with the shift to dynamic infrastructure and the rise of network automation. That is no doubt why they’ve joined the discussion of Infrastructure 2.0.
Infoblox (my employer) has been growing within the confines of this recession by automating DNS, DHCP and IP address management. It now claims close to 25% of the Fortune 500 as customers thanks to its ability to deliver real-time visibility into the state of the network and attached devices, as well as dynamic software updates (without downtime); and turn static networks into dynamic meshes that boost availability and reduce TCO, among other things.
Note: Cisco, F5 and Infoblox executives will be on a Dynamic Infrastructure panel at the May 19-22 Future in Review (FIRE) conference produced by the Strategic News Service, Mark Anderson's highly influential newsletter. All three companies are in talks regarding a January 15 Dynamic Infrastructure event at The Fairmont Hotel in San Jose. Contact me if you’re interested in attending, sponsoring or participating in either event and I’ll get you in contact with the planners.
Others will benefit from Infrastructure 2.0 as Well
Virtualization players like VMware and Citrix (and now Microsoft) as well as an emerging ecosystem of cloud computing players who are throttled back by the static network also stand to benefit as the network becomes more capable of simply keeping up with system and endpoint automation now delivered in pockets within the enterprise.
The dynamic network will be able to keep up with the higher velocity of change coming from increasing populations of mobile, dynamic systems and endpoints blending with a host of new initiatives. The network can play offense as well as defense and become a new source of innovation.
For example, a dynamic infrastructure would allow VMs to keep a security profile even if they are moved from one hypervisor VLAN to another. That would generate even more powerful economics for data center virtualization and grow the market considerably. More flexibility would mean more cost savings and less electricity consumption.
In addition to private company Infoblox, on the core network service automation side there are some other interesting earlier stage private companies involved with VMotion, including 3 Leaf Systems and Xsigo Systems, who have what I call “I/O front ends” which treat pools of commodity servers like a virtualized environment. The automation of core network services and the delivery of VMotion both represent aspects of dynamic infrastructure that are appearing in various stages and delivering new IT economies enabled by new levels of cohesion.
History Repeats Itself
IT has tackled cohesion-related issues before. We watched the flat database world give way to relational databases, driven in part by rising complexity and the operational impacts of duplication. Oracle (ORCL) became a technology powerhouse because it understood the power of the relational database to the enterprise before anyone else and seized the data. The result was a surging market cap enabled by unleashing the power of enterprise database cohesion and keeping CFOs informed of material changes in the business.
As TCP/IP beat out ATM as the network of choice, Cisco became a leader by delivering on the promise of connectivity and collaboration which enabled organizations to act more cohesively than ever before, despite physical distance. The result was an incredible boost in productivity, again enabled by cohesion.
When complex enterprise applications began being distributed over the WAN, F5 Networks was among the first to move to address the complexity and operational impacts of chatty protocols traveling over diverse links to ever widening assortments of devices and branch offices. As worlds collided, they understood the new demands and addressed them as competitors acquired rivals. They rode the branch office boom to new heights and potentials.
Oracle, Cisco and F5 have generated sizable market caps by generating unprecedented cohesion for their customers and addressing strategic IT challenges as they arose. As we look forward, Cisco, F5, Microsoft and little known Infoblox are well-positioned to address the challenges of more network-attached appliances and higher rates of change by ushering in the age of dynamic infrastructure. They will also raise the fortunes of virtualization and cloud computing players be making aligning networks with the requirements for transporting, protecting and managing the traffic between more dynamic systems and endpoints.
Summary: As We Look Forward
New enterprise initiatives like RFID, wireless/mobility and virtualization are now placing new demands on network infrastructure that has been static for close to three decades. Complexity is only part of the story; some of these initiatives erode cohesion because only portions of IT are automated, essentially shifting manual labor demands from one department to another. If virtualization enables the automation of server creation and movement, it requires similar gains in network automation. Otherwise, the time savings of system automation are eroded by the configuration and tuning demands inherent with a static, brittle network.
When static components require more manual labor, they erode the initiative business case for CIOs who own both sides of the equation. Large enterprises already experiencing diseconomies of scale for mundane tasks like IPAM can erode the business case for wireless/mobility and/or virtualization, etc. by deploying piecemeal automation, and shifting burdens around in a kind of shell game that hides operations burdens within already stressed network teams.
That is why I think the next tech boom will involve the emergence of dynamic infrastructure (Infrastructure 2.0) which will then lead to an explosion in network, endpoint and application intelligence. Dynamic infrastructure leverages the automation of classic core network services, such as DNS, DHCP and even IP address management (or IPAM) to establish connectivity intelligence (feedback loops combined with dynamic databases) between networks, endpoints and applications. Those feedback loops establish the preconditions for an explosion of new functions and capabilities, placing CIOs (and network teams) in more control of their destiny than at any time previously.
These trends toward selective automation driven by hype cycles will no doubt continue to erode the business case for technology innovation; and tighter network operations budgets are likely as we watch the global economy wobble. That puts the networking industry at a strategic point of leverage. Those who get it will prosper while those who continue to focus on “speeds and feeds” will become irrelevant.
Disclosure: Long Cisco.