eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Google is introducing new networking and storage capabilities to its global cloud platform that officials said will make it easier for organizations to move workloads between data centers and to scale and run their operations.
The Web giant announced the new capabilities—HTTP load balancing and persistent flash storage—June 16, about a week before Google kicks off its Google I/O user conference in San Francisco June 25. Google competes closely with the likes of Amazon Web Services and Microsoft in the crowded cloud computing space, and the HTTP load balancing feature in particular is a key differentiator for the company.
It’s also an example of what Google can do with its wide use of its software-defined networking (SDN) technologies and the worldwide reach of its cloud offering, according to Tom Kershaw, product management lead for Google Cloud Platform.
“The key is really flexible software and how SDN is applied to this, and also having a global footprint,” Kershaw told eWEEK. “The global network is really an asset. … It really is unparalleled in terms of its size and reach and performance.”
The HTTP load balancing capabilities will offer organizations and developers a range of benefits, he said. The key benefit is being able to load balance across different regions without disrupting the workload. The compute resources can automatically be moved from data center to data center based upon availability and proximity of the workload and demand to the data center. If a service is in demand in Asia, the workload doesn’t have to run in the United States; it can be automatically shifted to data centers in Asia.
The capability will help reduce latency in the service and enable organizations to optimize and scale their compute resources in particular regions when needed.
Kershaw also said that with the HTTP load balancing, Google’s cloud platform can scale up to more than 1 million requests per second without having to warm up and will support content-based routing.
In offering the HTTP load balancing, Google is leveraging its homegrown Andromeda SDN infrastructure, which is based on the OpenFlow protocol the company built to give it a faster, more flexible and more easily programmable network to handle that massive demand on the infrastructure. The search giant also was able to take advantage of software for managing the network, including finding bottlenecks and maneuvering around them.
In addition, whereas most clouds use multiple IP addresses, which Kershaw said can lead to silos, the Google Cloud Platform now uses a single IP address for the entire global environment.
Google’s solid-state disk (SSD)-backed persistent storage feature is aimed at developers who need high input/output-operations-per-second (IOPS) capabilities and want a better rate than what is offered at Amazon Web Services and elsewhere, Kershaw said. It’s priced at $0.325 per gigabyte per month, and supports up to 30 IOPS per GB. The standard, non-SSD backed regular persistent storage comes in at 4 cents, and features speeds of 0.3 to 1.5 IOPS. In addition, unlike Amazon, Google Cloud Platform doesn’t charge extra for each IOPS request, making it cheaper than Amazon, according to Google officials.
Both the HTTP load balancing and the SSD persistent storage are in limited preview, Kershaw said. Those developers who want to test the SSD persistent storage can put in a request on the Google Website.