The Tipping Point for Network Automation

Virtualization could easily be called the most critical enabler of cloud computing. After all, it decouples applications from hardware, freeing up companies from being tied to specialized hardware and enabling new mobility potentials for processing power. Those potentials could allow servers to be spun up and moved on demand from one data center to another, dramatically reducing the electricity requirements for IT services.

This potential has no doubt fueled interest in companies like VMware (VMW), Microsoft (MSFT) and Citrix (CTXS), just as cloud computing has generated renewed interest in Google (GOOG) and Amazon (AMZN). Cisco (CSCO), Juniper (JNPR), IBM and even Sun (JAVA) are all racing to form "cloud clubs" that will position them at an advantage over competitors trying to sell piecemeal into the coming unified fabrics of processors.

The Rise of Core Network Services

Yet, underneath the headline-grabbing buzz about virtualization and cloud computing another perhaps more mundane - yet almost equally strategic - shift is taking place in how networks are managed.

While the trade press for the most part has covered this in the form of various IPAM (IP address management) and other related product announcements (including other core network services, from DNS/DHCP, NTP to RADIUS, etc.) they do not get the new network reality that parallels the systems reality: the automation of networks, especially core network services is today a strategic imperative.

In the same way that virtualization automated the set up and movement of servers, network automation is enabling higher levels of availability and lower costs for enterprise networks, especially those used in manufacturing, supply chain, retailing and finance/banking. The recent Grainger case study insert, recently included in Network World, sheds light on what an impact the automation of core network services can deliver.

Reducing the sheer amount of manual labor required to keep a network up-to-date and available (given the explosion of endpoints and now the movement enabled by virtualization) is now a strategic consideration in the same way that CIOs are seeking virtualization to save on systems expenses.

A recent blog at Infrastructure 2.0 talks about network automation relative to cloud computing; another explains how "The Three Horsemen of the coming Network Revolution" will place increasing economic pressures on manual labor networks.

The common theme is that the core network service freeware delivered with many services has now reached a point of exhaustion in larger enterprise-class, networks as Cisco’s Gourlay recently commented (via YouTube) when he called them convenience features. You can get a similar perspective about the need for automated, integrated core network services from Cisco’s James Urquhart.

A study last year conducted by Computerworld found the IPAM (IP address management) smoking gun; with rising per unit IP address management costs incurred as networks grow. Virtualization increases the motion of these tired networks, and pushes more manual labor into areas already slowed down by manual configuration, committees, checklists and rising network infrastructure expenses. This CIO shell game is being played in perhaps half of the world’s largest enterprise networks.

CIOs are temporarily blinded to the impact of rising manual labor and inflexibility in the network as the penalties and tolls are extracted across multiple departments, from ops, to networks to help desks. Network administrators are sometimes tempted into thinking that manual labor is a form of job protection… until outages increase.

Yet the onset of virtualization and the emerging cloud business case promises to force those clinging to outdated network practices into even harder career, budget and people choices. Thus we are likely to see a new category of specialized network appliances appear just as existing categories are being commoditized. They are the first generation of Infrastructure 2.0 appliances, automating the critical core network services (from IPAM to DNS/DHCP, etc) that all networks rely upon.

This new category will tackle a host of emerging technology challenges, from DNSSEC to IF-MAP and IPv6, laying the foundation for tomorrow’s dynamic infrastructure.

Disclosure: long CSCO

Comments

[john potranca:]

I appreciate that you take the time to share you thoughts, but its hard to see them as anything more than marketing material the companies you own stock in...

Mar 30 09:07 PM

[GNESS:]

User:

Our household portfolio has stock in Cisco as disclosed. We may have some VMW. That's about it as I recall. I do tend to invest and write about stocks I like but I'm not offering any investment advice, mere thoughts about bigger picture tech trends and who could win or lose.

G

Mar 31 12:23 PM

[Lightway:]

The marketing backers of virtualization in any IT infrastructure area try to make it sound as though you just plug in the virtualization appliance and it does all of the work for you. Sometimes these virtualization applicance can be more painful and intrusive to setup than the tried an true technologies are.

This article also seems biased against network administrators, as though they are so bored they are looking for things to do. Every network admin I know is constantly busy, so they would welcome any opportunity to automate their jobs, because trust me, there is never a shortage of tasks. I find that claim in this article a little outrageous.

In the network layer, we have already had automation of IP addressing for years now--it's called DHCP.

Cloud computing to me is increasing becoming best practices scalability repackaged into a sexier new set of terminology. I can see why the industry pushes this, since new technology is what gets investors excited. Companies like Cisco are becoming established, infrastructure companies, similar to utilities, so there is a need to get the investment community excited over otherwise boring, but very mission-critical, equipment.

Mar 31 01:54 PM

[GNESS:]

Lightway:

I'm extremely sympathetic to the situation thet network administrators are in, hence the emphasis on automation. Yet there are admins who prefer to keep their spreadsheets because it has been what they did for years. Every market has late adopters.

The issue is about DNS/DHCP, IPAM, NTP, RADIUS and other services as well as manual configuration of various types uf network gear. I think cloud is more than a repackaging of the mainframe etc but rather a new approach to IT service delivery... scale, efficiency, flexibility, cost etc are all elements.

Thaks for an excellent cooment.

Greg

Mar 31 05:30 PM

[Phil Dewey:]

What this means of course is that automation at the network layer is done before it gets started as the cloud provider will do that with his labor costs baked into the price of a VM. The next big thing will be automating the care an feeding of a vast sea of VMs. Fortunately the technology lends itself to this. It will be the small virtual data center and data center automation vendors who take the day here.

Apr 01 01:39 AM

[Jeff Noel A...:]

With IBM coming into play in the new buzzword as of today, they will effectively utilize what they have in their possession to gain a stronghold in the game, unleashing their sales and Business IT hoardes, and possibly swallowing up smaller techs (Zoho?) for more leverage.

How much longer before much-awaited karmic Koala comes out and anihilate everyONE?

Apr 01 10:41 AM

[User 240577:]

Wasn't IBM 5 years ago talking about utility computing and what was the marketing term - ah yes - "Autonomic Computing" ....and what do they have to show for it some disconnected technologies put together in one nice umbrella...it will be the same with cloud computing; Amazon already has IBM beat at delivering something meaningful in this space. There are lots of blogs on why disruptive innovation is so difficult in big companies...IBM, CSCO imo are doomed to inorganically "acquire" innovation

Apr 01 11:14 AM

[GNESS:]

User 240577:

If you go back far enough plenty of the ideas/concepts in virtualization and cloud can be traced back to mainframes. I think who you see as havoing better offerings will depend on your needs, apps, etc. If you're at home or in an SMB chances are you're thrilled about cloud services; and if you're in the enterprise for the most part much less thrilled.

As soon as Amazon and Google start showing material revenue, sharing availability and security information and deliver robust apps suitable for the enterprise I'll take them more seriously from an enterprise standpoint.

They are, indeed, creating a huge amount of interest. At the end of the day they are impacting everyone's expectations. Just how those expectations are fulfilled (as a business) will shape how fast cloud is adopted.

Thanks for the comment.

Greg

Apr 01 01:03 PM

[GNESS:]

Phil:

Excellent comment! Yet many of the networking vendors have been slow to embrace network automation. DNS, for example is offered as freeware that needs manual spreadhseet updates. The CIO needs for hte network what hte CFO has for the business: real time status of everything connected, etc.

Thanks,
Greg

Apr 01 01:06 PM

[GNESS:]

Jeff:

Its amazing what IBM has accomplished. I remember the WSJ article from the 80s that said that they were on a one way track to an unglamorous end... and look how far they've come given the sea change in computing. Amazing.

Thanks for the comments
Greg

Apr 01 01:08 PM

[Jeff Noel A...:]

can VMware and IBM be even put in the same phrase?

Apr 02 08:53 AM