Hybrid Clouds: Understanding available options, hazards, and opportunities
Unless you’ve been following daily cloud computing news, chances are your understanding of the technology is outdated. This is especially true over the last 18 months, in which telecom, media, and technology companies (among others) have dramatically increased the type and number of applications they are able to deliver over public clouds.
No reason to delay the move to cloud any longer
What happened? Two things: Major public clouds from Amazon, Google, and Microsoft have become even more sophisticated, if not specialized, in how they codify and virtualize applications. And secondly, bare-metal server technology makes it possible for legacy applications such as Oracle Database to be more easily distributed in a cloud-like way, without the costly migration headaches like before.
In other words, we used to say that most large companies could successfully move 20-30% of their enterprise applications to the cloud, for better performance, customer experience, and cost gains. Thanks to the above new technologies, we now estimate that most large companies can migrate upwards of 60-75% of their application with a mixture of modern clouds and bare-metal servers.
There is one notable exception to this: if your business still relies on a large amount of mainframe computing, you likely won’t be able to achieve that promising number. Nevertheless, most of us don’t have to perform open heart surgery anymore to migrate to the cloud. Unless you depend on mainframes, there’s no good reason to avoid or delay cloud migration anymore. Whereas it used to be incredibly hard, these new options make it more accessible than ever before.
More options, more speed
And the options are seemingly endless, even for regulated or accredited companies. If you already have a mix of physical and virtual machines, all of that can be moved to public clouds. If you have your own dedicated infrastructure, data centers, VMware stacks, Oracle middleware (or anything else made by Oracle), that can be moved too.
This can be done with a mix of Google Cloud for native, new apps, in Microsoft Azure for .net and java applications, and Amazon Web Service for everything else, especially media transcoding. With a public cloud provider plus bare metal, you can ultimately move the majority of your digital real estate now.
Take our work with the Olympic Games, for example, where 100% of critical applications are run in the cloud. Following progressive deployment from data centers to a fully cloud-based environment, Atos oversees overall IT management and technology integration while maintaining their global performance with a much smaller agile team. But their reach, accessibility, and availability is actually greater than ever before, thanks to their full migration to the cloud.
Not every organization can achieve that kind of success, of course. RS Components, one of the largest technology companies in Europe, recently retained Atos to help them migrate the majority of their applications and data to the cloud.
Unless you depend on mainframes, there’s no good reason to avoid or delay cloud migration anymore.
Recently finished, the project took over 18 months to migrate one critical multi-national application. Still, the company was happy with the success given the constraints of what they were working under and can now benefit from cloud on tap services including data analytics, machine learning and artificial intelligence.
The good news for everyone else is that the options, speed, and cost gains have since improved. If a company like theirs is happy, chances are your satisfaction will be even better with the increased number of options available.
Benefits for telco and media
Of course, there are still several hazards when migrating to the cloud. Although easier, cloud migrations are certainly not a slam dunk. Telecom companies, for example, already have to manage their own network operation data centers to keep their telephony needs in order. This, of course, is tempting for them to host their customer facing applications on the same data center, even though a public cloud can make for a better, faster, and more affordable customer experience.
Media companies, meanwhile, have to consider the cost of large amounts of data, and how to best avoid egress charges as they move that data to and from end customers on different devices.
For example, Atos spends millions of dollars per year with Amazon alone to support one of Europe’s largest broadcasters. The data is expensive, but with the right configuration it’s still cheaper and more flexible than hosting it with no capacity restrictions. It also provides a better experience across all customer devices.
Over the last 18 months, we’ve learned a thing or two about successful modern cloud migrations. They are as follows:
1. People are still the biggest hazard.
Very few people can build native cloud applications. A high proportion of resources are still not ready and are tied up managing legacy applications on dedicated infrastructure. To learn this stuff, you really have to be native cloud learners who are both interested and invested. In the DevOps world, cloud developers have to be proficient in application development and need to be aware of infrastructure and network operations. That’s a tall order for anyone.
2. You can do more with less.
The counterpoint to the above is that once you find the right people, you can manage an entire cloud-based organization for a factor of 10 to one. That is, cloud computing is so highly automated that it can usually be managed by a factor of one to 10 what it used to cost with an on-premises data center. Very few people believe this can be done so they are hesitant. That was certainly the case with many of our projects including our Major Events team, which now run sporting applications in the cloud. Better performance, cost and availability managed by a team that’s twenty times smaller than before. That’s big.