I can say I have been done the right choice!

I’d like to share my experience of planning and developing an IT project. My customer’s need was to realize a centralized infrastructure with a disaster and recovery site in a different region and connect about 80 little remote locations.

The customer environment consisted in a mix of operating system that hosted various services as: Active Directory, DNS, Databases and third party applications. These had to be always available 24x7x365. Besides all sensitive data transmitted from the branch office to the central site had to be available for offline browsing and stored on a tape library for a long time to be compliant with the country law where the customer was located.

Considering the big amount of servers present in the central site, the choice was inevitable: to propose the customer a virtualization solution able to offer one centralized point of management, high availability of the services, disaster and recovery site, and at the same time to explain that with this choice it would save money.  

Two solutions were available: VMware and Hyper-V. In a very short time with my team we implemented in our lab two different systems with both solutions to compare them and help us in the choice.

After few weeks of testing, below shortly what I can assert:

• VMware ESXi inserts a robust virtualization layer directly on the server hardware for near-native virtual machine performance, reliability and scalability. VMware ESXi is a compact hypervisor.
• Hyper-V requires an operating system to be installed in the "Controlling Layer" before the "Hypervisor Layer" can operate.
• Vmware in breadth of operating system supported is the clear winner.
• Storage is another critical shared resource in a virtual environment, and the choices are a bit more limited with Hyper-V.
• File System:  VMFS (Virtual Machine File System) is an incredibly powerful and scalable file system. VMFS delivers a number of features and capabilities that are not found in the Microsoft’s competing VHD (Virtual Hard Disk) format.
• Migration of virtual machine: VMware’s storage vMotion technique is a no downtime affair meaning that is fully transparent to the end user.  Quick Storage Migration lies in the fact that it’s not fully transparent to the end user. At the end of the migration process, the VM is placed into a save-state for up to one minute.
• Upgrade simpler in Vmware. Windows Server 2012 with Hyper-V is subject to unrelated patching.

Without any doubt I proposed Vmware to our customer and I worked immediately. Below in summary what I realized:​​

Primary Site: A Farm VMware composed by 4 Host ESXi connected to a SAN. The infrastructure was protected with a backup system composed by a physical server connected in fiber channel to the tape library .The production environment  was later replicated on the secondary site, using the NAS replica functionality and integrating VMware Site recovery manager to provide the central point of management.

Secondary Site: Same Farm VMware
 

I would remark that I installed the vCenter Server, created the Virtual Datacenter, configured both networking and storages, configured the resource pools and enabled the functionality of HA and DRS. All this was achieved from a unique interface, vSphere, in a very simple and intuitive manner.

Regarding the replication between Primary and Secondary site, through the Site Recovery Manager, that is a vSphere component, I first created the array manager to make the replicated storage visible, and then I configured the respectively protection group and recovery plan. The important thing to highlight is that I also tested the functionality failover and failback and in both cases I fully respected their SLA to be on line in case of disaster of primary site.

In conclusion I am very excited of this experience, especially when I met the customer after one year and he said: “You made a good work; the virtualization solution proposed fitted perfectly our goals. VMware is really a synonymous of safety, reliability and stability. Thank you!”

Architects of What's Next (playlist)

Hyper-V and VMware vSphere - Architecture and hypervisor footprint differences

Some of the most obvious differences between VMware and Hyper-V lie in the product architecture and footprint.  Both Microsoft Hyper-V and VMware are using a virtualization architecture named TYPE 1VM to implement their server virtualization technologies, but the Hypervisor design implemented is completely different.  Before to go in deep on this topic, considering that the classification of specific hypervisor implementations as Type 1 or Type 2 is not always clear cut I will before explain the differences:

• Type 1 (or native, bare metal) hypervisors run directly on the host's hardware to control the hardware and to manage guest operating systems.  VMware Vsphere and Hyper-V are example of Type 1 hypervisor. (On the architecture front, Hyper-V is often mistaken for what’s known as a Type 2 hypervisor when in reality is TYPE 1)

• Type 2 (or hosted) hypervisors run within a conventional operating system environment. The main difference to consider is that with the hypervisor layer need to be considered as a second software level; guest operating systems run at the third level above the hardware. VMware Workstation and VirtualBox are examples of Type 2 hypervisors.

To summarize, Type 1 hypervisor sits directly atop the hardware and runs directly on it; a Type 2 hypervisor runs on another operating system, such as Linux, or Windows.

We will focalize on Type 1VM and for that reason we can analyse the two subcategories, where the differences between the two products lie: Monolithic Hypervisor Design and Microkernelized Hypervisor Design.  

• Monolithic Hypervisor Design (used by VMware vSphere):  should be considered as one big program where process and memory management, interrupt handling and I/O communication, run in kernel space.  The advantage of being monolithic is it deals with resource management rapidly but it require more time during initialization.

• Microkernelized Hypervisor Design (used by Microsoft Hyper-V):  the kernel take care only of the basic process and I/O control, the other system services has been moved in the user space in form of normal process (as so called servers).  Less initialization time is required but Microkernelized Hypervisor Design requires an operating system to be installed in the "Controlling Layer" before the "Hypervisor Layer" can operate.  

Both the designs have pros and cons and, but in my opinion, at present, vSphere 5 holds a considerable lead in features and performances. One of the main reason is that VMware ESXi inserts a robust virtualization layer directly on the server hardware for near-native virtual machine performance, reliability and scalability. VMware ESXi is a compact, hypervisor.  It is a fraction of the size of a general purpose operating system for unparalleled security and reliability.

Microkernel  Hypervisor is a new approach used  to implement an hypervisor (It should provide much more than Virtual machine) Monolithic Hypervisor is specifically designed for that scope.  It is a native solution that not requires operating system for controlling all the components of the virtualization product.  

 

This is also the primary reason for this significant space difference of the two hypervisor solutions.VMware vSphere Hypervisor installation requires only 144 MB of disk space; a Hyper-V installation requires a minimum of 3 GB for a Server Core installation.  Hyper-V relies on the installation of the general purpose Windows Server 2008 R2 operating system (when the Hyper-V role is enabled, the existing Windows instance becomes a virtualized root partition) while vSphere does not have such weight attached to it.