Global Traffic Management for Cloud, Data Centers, and CDNs
Twenty years ago it would be common to have multiple data centers load balanced with specialized devices. The applications in multiple data centers would use the Domain Name System (DNS) to look up the application destination. The DNS system would be set up to with multiple data center locations. This would "round robin" the results between the data centers. If you have four data centers, the customers would get "round-robined" between each data center. Round Robin with DNS lookups was not really load balancing though. Architects needed tools that would balance traffic between multiple physical locations and move traffic away from overloaded data centers. They needed tools that would geographically load balance to get people to the closest data center. They needed resiliency options so that if a data center failed, the traffic would reroute. Global Server Load Balancers (GSLB) were created to add enhancements to DNS to move behind round-robin techniques. GSLBs would be primary DNS servers for the application's zone. The GSLB would normally be in one of the data centers, monitor the load with other data centers, and pass each customer query to the most appropriate, closest, and under-loaded data center. The GSLB would route around failed or overloaded data centers. GSLB would achieve resiliency with a primary and secondary GSLB in a different data center.
The traditional "GSLB in a data center" worked well in the "all data center" world. That has all changed in today's hybrid cloud and the edge-compute world.
Everything you know about global load balancing has changed!
Today's IT team must deploy and maintain applications all over the Internet. Applications continue to be in data centers. New applications get pushed into the cloud. Many of the applications are distributed on CDNs. In some cases, the organization will use multiple cloud operators. Other deployments will have a hybrid of data center and cloud. Other applications will use multiple CDNs.
Through all of this, the IT team must work with their peers to maintain a 99.999% customer SLAs. If they lose instances on cloud deployments, they have to ensure that it will fail over to the new instances. If there is an outage in a data center, they need to ensure that traffic swaps to the new data center. When they have a hybrid of data center and cloud deployments, they need to manage the load between the instances.
Load Balancing, Geographic Balancing, and Performance monitoring are all still required in this "modern Internet complexity." Today's requirements are more complicated. The traditional GSLB model inside of data center is stressed to match today's modern Internet complexity. It is only logical for the GSLB to move out of the data center into the cloud.
In the cloud, the GSLB function is decoupled from the physical data centers. In fact, an independent GSLB can be in an "overwatch" role separate from all the "hosted" locations of the applications. A cloud GSLB would be able to monitor the resiliency of the applications and services from outside in. Outside in better simulates the customer experience. An independent cloud GSLB can manage multiple cloud deployments. This allows the organization to pick the best cloud service for each unique application element (compute, storage, analytics, distribution, etc). A cloud GSLB allows for easier hybrid architectures using the best elements in various data centers and from various cloud providers.
Why would the GSLB Function be in the Cloud?
GSLBs in the cloud is the solution needed for today's complicated Internet topology. For example, an IT Team might have the following requirements:
- Seven data centers/cloud in Europe, US East / West, LatAm, and Asia to provide global coverage. Asia and LatAm are cloud deployments;
- Geographic Mapping - Constraining a geographic region's traffic to data centers within that region;
- Measure data center/cloud performance with monitoring network and user feedback transactions;
- Accept policy data changes programmatically with API transactions;
- Shape traffic across each of the data centers/cloud - manage the traffic load; and
- Establish a maintenance mode to allow for one of the seven data centers/clouds to be taken offline for maintenance. The GSLB would route traffic to the other 6 data centers/clouds.
This type of geographic deployment is normal for today's Internet. The GSLB would be asked to support multiple protocols and services to support all the functions of the organization. Trying to run this GSLB function from inside the seven data centers places limits on the design. A cloud-based, globally deployed GSLB function provides the IT Team greater architectural flexibility.
Two Cloud-Based GSLM Solutions from Akamai
Today Akamai has two solid, scalable, and robust GSLM services used by thousands of customer throughout the world. Most people do not think of Akamai as a global load balancing provider. Yet, the essence of Akamai's business is scaling the Internet. Moving the GSLM function into the cloud is a critical element for Akamai to help our customers scale their operations. With this, Akamai offers our customers two GSLB tools Global Traffic Management (GTM) for any traffic to any data center (e.g., API traffic shaping) and Application Load balancer (ALB) to optimize origin traffic flows and session affinity (e.g., original switching). GTM is a flexible multi-use, multi-protocol tool used by our customers to manage banking tools, databases, data centers, multiple-CDNs, and a surprising range of other options. ALB is an Akamai Cloudlet option that leverages our massive intelligent edge to move traffic management to layer 7. Both GTM and ALB follow Akamai's practice of maximized reliability, availability, and scale in all conditions independent of any underlying "stressors."
In future blogs, we'll provide a range of uses cases for GTM and ALB. The following provides a summary.
Akamai's Global Traffic Management (GTM)
Akamai's GTM is a modern, highly reliable, dev-ops optimized, GSLB service that operates at the edge of the network across a diverse, distributed network of networks. Akamai's leading Information Security program governs with world-class controls and compliance and its Network Operations Command Center is on 24/7.
GTM allows for a range of load balancing, traffic routing, performance monitoring, geo-mapping, and fault monitoring. A global GTM deployment is an integral part of Akamai's Intelligent Edge Platform and provides organizations options to move services to optimal geographic locations. GTM services such as geo-mapping allow operations to deploy applications on multiple clouds and route user application connections to the geographically or topologically closest edge servers. Akamai's performance monitoring is an outside-in approach, setting up monitors worldwide. Akamai's GTM performance monitoring leverages the same monitoring used for Akamai's media, web, cloud, and security products that some the largest application, media, and network providers depend on every minute of every day.
Akamai's Application Load Balancer (ALB)
ALB, Akamai's latest load balancing solution leverages the power of GTM, and additionally introduces layer 7 visibility and control when traffic is delivered through the Akamai platform, enhancing origin traffic routing options, and reducing risk to the user experience when expected outages occur. ALB introduces the following additional capabilities:
- Layer 7 routing enables routing decisions based on header information such as URL, device type, or content characteristics;
- Session affinity protects the user experience by binding users to an individual origin during normal runtime operation maintaining critical data such as login, shopping cart, or other user unique data; and
- Instant failover reduces downtime risk when unexpected outages occur by rerouting user requests to backup origins with no disruption to the user experience.
ALB Plus GTM
There will be many architectural situations where an organization will have GTM deployed to cover the resiliency for some functions and ALB deployed to cover specific applications. The combination of ALB and GTM provide devops, network architects, and system engineers with the flexibility to choose the optimal cloud-based GSLB solution. Some applications architects might need cloudlet ALB functionality to manage in-line experiences that Akamai delivers. At the same time, network architects for that same company might need to shape traffic between multiple cloud deployments and providers. With GTM and ALB, system engineers and application developers have full control - the power of AND.
What does the Future Hold?
GTM and ALB are an integral part of Akamai's Intelligent Edge Platform. Stay tuned to Akamai's blogs and subscribe to Akamai's Community to follow new updates, articles, and presentations as our solutions continually evolve to meet our customer's needs.
Find out More about Akamai's GTM and ALB
- Global Traffic Management (GTM) - Ensure fast and reliable user experiences by balancing traffic across all your data sources - both cloud and on-premise.
- Application Load Balancer Cloudlet (ALB) - Maximize application performance and availability with fast, flexible, and reliable load balancing when using Akamai's Intelligent Platform to deliver content.
- Critical Blog: Protecting your Domain Names: Taking the First Steps - Critical advice to protect your domain names from being hijacked.
- Designing DNS for Availability and Resilience against DDoS Attacks (Akamai White Paper).