Network Cloud and Virtual Readiness

The widespread adoption of cloud and virtualized computing and the various ways that enterprises will change to fully exploit it will have a profound effect on the network and on IT Network Operations Groups. 

Customer expectations are changing.  New requirements are emerging, service levels are becoming more stringent and some time tested strategies for managing costs and ensuring adequate service levels are being invalidated.  Trying to use the same old network and network management strategies and tools without Local Management will cause virtualization and cloud initiatives to fail or to incur runaway costs. 

Cloud computing will increase network utilization, criticality and visibility in a number of ways while also increasing the number of network devices that require management.  High impact changes include:

Increasing Remoteness of Applications

In Client/Server application architectures, a significant portion of the processing typically takes place local to application users.  This shields users to a large extent from intermittent WAN outages and congestion.

An Environment Built for Client Server Applications with PC Clients,
Local Servers and a Centralized Database Back End

Increasingly enterprises, for the purpose of simplifying management and to take advantage of lower cost more scalable compute resources in virtual datacenters and the cloud, are moving application functionality away from users. 

An Environment Built for Centralized Applications with Thin Clients (e.g. SaaS Web Applications)

This has a number of consequences for the network.

  • User experience becomes extremely sensitive to network performance and availability. 
  • In most cases network utilization will increase a great deal. 
  • Monitoring application and network performance from a user perspective becomes considerably more complex as an application can be an array of application services hosted and delivered from multiple sources over different network segments that are never aggregated other than at the user’s location.  

Physical Disassociation of Applications from Infrastructure

In a traditional datacenter, application components from a single business application are installed on specific servers connected by specific network devices.  This makes it is possible to ensure that a set of properly configured network devices are used by the application and that the network devices used can be selected to be appropriate for the application they support in terms of performance and reliability (including HA) vs. cost.

One of the primary benefits of virtualization is the fluidity of server resource allocation in response to business and application needs.  Unfortunately, from a networking perspective this means that any business application component in a given cluster can be on any server in the cluster using any network devices used in that cluster and even any number of them.  Further, VMotion (and its other hypervisor equivalents), in order to deliver its benefits, requires that the system is constantly changing.  As a result of this, every network device in the cluster must be treated as a component of the most critical application in the cluster. 

Management of Virtual Switches

Every single virtual host (the individual server with the hypervisor that the various virtual machines run on) requires a virtual switch to manage network communications between the VMs on it and between these VMs and the rest of the network.  Each virtual switch requires almost the same configuration and management as a physical switch in the network and issues that impact physical switches can have the same impact on hosted VMs. 

Changing Internal Customer Expectations

Cloud Services Providers have made it easier than ever for Line of Business users to rapidly and inexpensively source computing infrastructure, storage, hosted development platforms and enterprise class business applications directly over the Internet.  This can be started with a very low monthly subscription fee, no commitment and a credit card. 

This reality is redefining the way that business users expect to interact with their technology providers, and this directly impacts internal IT.  Line of Business leaders are acting more like customers trying to get the most for their money rather than colleagues trying to figure out the best answer for the organization, and they have options beyond internal IT.

At the same time, none of these services can be accessed without a robust network.  Even as internal customers turn more to third party service providers for a portion of the services that they once looked exclusively to internal IT for, their dependence on internal IT to enable them to access these services increases.   This shifts IT’s focus to network performance, security and availability. 

Uplogix Local Management and Cloud and Virtual Computing

The only cost effective way to confidently meet the pervasive higher service levels that are mandated for networks by virtual and cloud computing is to implement Local Management to reduce the likelihood of problems occurring and to achieve big decreases in Mean-Time-To-Recovery (MTTR) when issues do occur.  In this manner, Local Management provides a way to meet the elevated availability requirements of virtual and cloud computing that is cost effective enough to be implemented everywhere.

Uplogix Local Management includes the following:

  • A Local Management device or server co-located with and connected directly to the network gear to be managed (not using the network).  This can be an Uplogix Device or a virtual Local manager using your hardware.
  • Local Monitoring that takes advantage of direct connectivity to more rapidly identify failures and issues (typically the console port can support 30s intervals vs. 5-15 minutes with centralized network dependent monitoring systems using SNMP).
  • Point-to-point network monitoring between Local Managers installed in proximity to both users and application service points of origination that can measure the performance of specific network segments.  This makes it possible to accurately assess performance as experienced by users of centralized applications.
  • Automation capable of performing a sequence of intelligent recovery actions in response to problems comparable to what would be done by an on-site technician up to and including power cycling devices.
  • Configuration Management that automates complex configuration tasks to improve efficiency, reduce errors and allow for configuration changes to be included in automated problem resolution routines. 
  • Locally enforced Role Based administrative access and detailed logging to prevent malicious outages or outages that result from non-compliance with standard administrative procedures and to protect against unauthorized data access.
  • The ability to establish an out-of-band connection for when human intervention is required and to flow information to centralized network management tools when the network is down.
  • A flexible rules engine to allow for a wide range of additional automated actions as required by your unique goals and environment.




In our research, the more sophisticated IT executives who run 'aggressive' IT organizations consistently place 'automated problem/resolution' as a key item on their wish lists. They want tools that replace manual processes for maintenance and recovery with automation. We expect a growing number of IT staffs to look for these capabilities.”

— Robin Gareiss,
Executive Vice President, Nemertes Research