Cloud Hosting

Remember pets and cattle? CERN's 2012 metaphor to describe on-premises servers you name and care for lavishly versus virtualized cloud servers you never name, run in a herd and snuff out without a second's thought?

Well the metaphor's evolved: VMware and Pivotal are now talking about McMansions and hotels to explain how they will bring virtual networks into the world of DevOps.

The companies feel that most DevOps work is taking place either with cloud-native applications or on the margins of a big organisation's software flet. Core applications remain largely untouched in the push to continuous deployment, largely because even small changes to older code in the heart of a business require detailed security and compliance oversight. That slows things down because large organisations have silos to take care of those things. Between politics and governance, that makes it hard to get close to continuous delivery.

That state of affairs got the two companies thinking about the surprise breakout use case for NSX. VMware first imagined NSX as a control plane for networks comprised of different vendors' hardware. It's turned out to be more immediately useful for microsegmentation the practice of creating virtual networks to link a small set of resources, often tied to a specific workload. Because these virtual networks are only required to do certain things, they are defined by policies that don't let them do anything else. If behaviour not defined in policy is detected, microsegmented networks either isolate themselves or make red lights and klaxons go off down in Ops.

At Dell EMC World the companies will explain how NSX will be integrated with Pivotal Cloud Foundry so that when developers work on stuff that touches compliance-and-security-sensitive applications, they'll do so inside virtual networks that reflect all the worries security and audit teams want taken into account. Instead of creating compliance-friendly new development environments a McMansion with a room for everyone and every household activity - they'll check into a hotel with just the services needed for a short stay, but a policy-enforced empty minibar.

These development environments will have their own IP and MAC address spaces and, while they may run on shared hardware, will be logically discrete from production environments and from other testbeds.

The two companies think this approach will be especially attractive to developers building containerised systems on top of core applications, because spawning containers, chaining them and they destroying both the containers and the connections between them sets off alarms among compliance pros. Those folks are accustomed to being able to trace transactions with great granularity, a task that possible with containers but made harder by the fact containers are treated as even more disposable than cattle. Showing governance types that all of this whacky work happens within virtual environments that adhere to policy makes for greater comfort.

We're not sure at this stage exactly what Pivotal and VMware will announced, but The Register understands this is a day two announcement. Michael Dell, David Goulden and Intel's Diane Bryant are the day one speakers at the show.

Read the original here:
Servers as pets or cattle was 2012. Now it's McMansions or Hotels ... - The Register

The biggest fear early cloud adopters overcame was the fear of disruption. How would they move everything from their existing data center to a cloud service without disrupting business operations? The answer was, they didn't.

Smart companies realized that the safest way to transition from on-premises systems to the cloud would be to first assess all their applications, workloads, and other data assets to determine which would be most amenable to the cloud transition. They began with low-criticality workloads and slowly worked their way up the list. Some are still working through that process.

For those organizations, a new concern emerged. How would they manage an environment that basically consisted of two network cores; one still on-premises and the other now in a cloud facility? Would they need to maintain two separate environments? Two management consoles? Two sets of protocols? What would happen when they needed to combine resources from both environments to accomplish something?

The alignment of Active Directory for Windows Server with Active Directory for Azure meant that data center managers could create a unified, seamless system that combined both on-premises and cloud-resident systems under one AD forest. This enables IT to use one set of protocols, one set of user credentials, one set of security standards, and one unified data environment.

It also facilitated the smooth movement of more applications, workloads, and other data assets from the on-prem network to the cloud. Once the two are aligned, it becomes almost effortless to provision new virtual machines and new resource stacks, and move them across the network.

Gartner recently predicted that, by 2020, the "cloud shift" will affect more than $1 trillion in IT spending. That means that more organizations will be migrating more of their IT operations to the cloud.Some will still need to port, modify, or rewrite some applications before they can be migrated. The recent embrace of Open Systems by Microsoft clearly signals Redmond's commitment to assist in this challenge. One way in which the company is facilitating cloud migration is by developing more and more enhancements to Azure services that will accommodate a broader variety of platforms and environments.

Many resources combine to provide a complete Azure service for your applications. These include virtual machines (VM), your storage account, web apps, databases, and your virtual network (VLAN). Originally, when you managed an Azure cloud environment, you created, deployed, and managed each of these separately. When adjustments needed to be made, you went to the required consoles manually. When you wanted to replicate your solution in another Azure environment, you had to recreate everything.

The introduction of Azure Resource Management (ARM) in 2014 changed all that, mainly by providing resource groups, which are defined by Microsoft as "a container that holds related resources for an Azure solution. The resource group can include all the resources for the solution, or only those resources that you want to manage as a group. You decide how you want to allocate resources to resource groups based on what makes the most sense for your organization." Now all the required resources can be managed as a group, which saves tremendous time and effort.

As the cloud revolution marched on, it became clear that cloud customers enjoy better IT services at a lower cost. It made little sense to own and operate your own servers when a cloud provider could deliver a professionally operated data service, while you pay only for what you use.

Now, all of the tools required to ease the path from on-premises to cloud facilities are emerging very quickly. Companies can move from on-premises to robust co-existence and then finally complete the journey to the cloud on their own schedule with virtually no disruption whatsoever.

Read more:
Transforming the datacenter - ZDNet