Enterprise Strategy Group analyst Colm Keegan says that Web scalability is the answer to what he calls "silo-ization syndrome," when virtualized applications are separated into compute, storage and network resources. While hyper convergence is one way to get rid of silos, Keegan says that that hyper converged infrastructures can still have a silo problem -- even if those silos are more efficient. What does he suggest? Scale-out architecture, such as that created by Nutanix, can help reduce hyper converged silos. When it comes to silos in hyper converged infrastructures, the main issue is that compute and network resources have a lot of space for data, but storage runs out of room. By including compute, network and storage resources into the same scale-out nodal design, Keegan says you can achieve Web scalability, which ultimately gives IT planners a way to deploy resources across all of the available virtualized workloads.
Read more about why Keegan thinks Nutanix has the ultimate Web scalability product.
The future of network design and architecture is in software and freedom
Ethereal Mind blogger and network engineer Greg Ferro says that up until this point, network design and architecture has been planned with whatever resources are available. Yet future network design and architecture needs to be based on software and the freedom to mix and match in order to be predictable and safe. What does predictable mean? For Ferro, a few things come to mind:
- Configuration changes can be made at any time without risk.
- Devices can be added, removed and replaced at any time with certainty of operation.
- Resilient performance under failure, outage or exceptional events.
Hardware and software work together, but Ferro says the focus needs to intensify onto software. This means several things, but chief among them this piece of advice: Virtual devices should be preferred over hardware and companies should participate properly in growing open application performance interfaces (APIs) instead of building proprietary APIs. When it comes to the freedom to mix and match components and technologies, the bottom line is that the network can and will be part of data center orchestration.
Read more about what Ferro says about the need to change network design and architecture.
Troubleshooting network problems, tips and suggestions
Cisco team lead Denise Fishburne blogs on Packet Pushers that even after 30 years of successfully troubleshooting network problems, her process has remained the same. Her six tips to follow no matter what your network issue is are:
- Be methodical
- Know what is normal
- Get to the "crime scene" as fast as you can
- Have "crime scene maps" that help and don't hinder
- Let the clues and evidence guide you
- Learn and improve
Of the six tips, she says being methodical is the toughest challenge because of the added stress and pressure created by trying to solve the problem. By following a process -- no matter how bad the situation is -- you can keep a clear head and work through the steps more efficiently.
Read Fishburne's explanation of each step and find out which troubleshooting tactic you should focus on the most.
Gartner announces "cool" vendors in enterprise networking 2015
Gartner blogger Andrew Lerner says that Gartner has recently published its 2015 version of Cool Vendors in Enterprise Networking. The list reveals what the firm perceives to be its five highest rated vendors: Appcito, Big Switch Networks, CloudGenix, OpenDaylight and Viptela. Lerner says for a company to be considered "cool," it must have an innovative approach to troubleshooting customer challenges. One of the coolest trends: companies that simplify branch networking with hybrid WAN platforms. While the top five vendors have the most innovative approaches, Lerner also mentions Nuage and VeloCloud as cool vendors from last year that are still considered emerging technology providers with products equipped with promising features.
Read more about what it takes to become a "cool" vendor in enterprise networking.