Load Balancing in Azure with LoadMaster

The Microsoft Azure Infrastructure as a Service (IaaS) platform enables applications to be easily provisioned in Microsoft’s cloud. Existing on premise applications can also be seamlessly transitioned into Azure, allowing technology decision makers to benefit from the scalability, elasticity and shift of capital expenses to operational ones. When Microsoft Azure is incorporated into an organization’s hybrid cloud strategy, many exciting benefits are realized!

  • Azure makes higher agility possible for your business
  • Azure makes Development Operations easier to achieve
  • Azure makes integration with on premise applications and data simple

High availability and disaster recovery are of highest importance for application infrastructures since other important factors such as capacity and performance are negated if service either is not available or cannot quickly be recovered in the event of an instance failure or degradation. While the Azure platform provides native, basic application load balancing capabilities, KEMP’s LoadMaster greatly enhances and complements these capabilities by delivering true layer 7 Application Delivery. When LoadMaster for Azure is combined with KEMP’s GSLB functionality known as GEO, high availability across on premise and cloud based application pools and intelligent global traffic distribution across multiple private cloud infrastructures and Azure becomes possible to support highly resilient, scalable and smart application deployments.

KEMP’s close relationship to Microsoft and commitment to enhanced integration with Microsoft platforms such as Azure and their suite of Enterprise productivity application results in optimal end-to-end user experience for mutual customers.

The LoadMaster Load Balancer/ADC for Azure is easy to provision, configure and manage right within Azure

  • Easily provision from VM Depot in the Azure management portal
  • LoadMaster assumes Public/Private Endpoint functionality for services
  • Connect Application Server instances to LoadMaster for Azure
  • Start load balancing your application with LoadMaster's app delivery capabilities

Compare the Azure built-in Load Balancer and KEMP LoadMaster for Azure

  Azure Load Balancer KEMP LoadMaster-for-Azure
Network Level L4 load balancing    
Application-Aware L7 load balancing    
High availability  
using Traffic Manager
Hybrid Traffic Distribution    
(with advanced traffic distribution)
Load balance across Azure VNET regions    
(Internal Azure and external network)
Scheduling methods Round Robin Only L4/L7
Server Persistence  
(L4 only)
L4/L7(Cookie and more)
SSL Termination/Offload    
Content Caching/Compression    
Least Connection Scheduling    
VM Resource Availability Awareness    
Web application Firewall Protection    
Support Basic Tier VM (20% Savings)    
Support Standard Tier VM    
Single Sign On    
SmartCard(CAC) / Certificate Authentication    
LDAP Authentication    
Radius Authentication    
Kerberos Constrained Delegation Support    

Introducing Load Balancing in Microsoft Azure

Before we create a LoadMaster Virtual Machine (VM) in Azure, it is important to understand the traffic flow so that VMs in Microsoft Azure can be configured appropriately.

Microsoft Azure IaaS deployments accept traffic only on published endpoints. Any request to access Microsoft Azure workloads passes through the default load balancing layer of the Microsoft Azure platform. Figure 1-1depicts the default deployment without the use of KEMP LoadMaster in Azure.

Figure 1-1: Microsoft Azure Load Balancer

Any workload being published consists of a cloud service, which represents a single Virtual Machine (VM) or multiple VMs. When a VM is created, if a cloud service exists, you have an option to connect the VM to an existing cloud service. As more VMs are connected to an existing VM (and thus to an existing Cloud Service), built-in Microsoft Azure load balancer distributes connections when creating a load balanced endpoint.

If you wish to use LoadMaster for Azure for your deployment, we must complete the following steps:

  1. The LoadMaster for Azure needs to be deployed first, creating the required cloud service.
  2. All the virtual machines that need to be load balanced using LoadMaster can then be created and must be connected to the existing LoadMaster VM to create the required grouping.
  3. Finally, when creating endpoints, we cannot use Load-Balance traffic on an existing endpoint option as we do not want to use the Microsoft Azure Load Balancer to load balance incoming connections. Figure 1-2 depicts the flow when LoadMaster for Azure is deployed:

Figure 1-2: LoadMaster for Azure

Notice that VM1, VM2 and VM3 in this example are grouped into a single cloud service and the endpoint for published virtual services will be created only on the LoadMaster VM. By doing this, we receive all load balanced traffic on the LoadMaster VM and the logic of load balancing incoming connections will be applied as per the configured virtual service on LoadMaster for a given workload.

Also notice that VM1, VM2 and VM2 will not have any endpoints as they are not going to be published directly to the internet.