I recently had the privilege of spending two days with 30 CIO’s from various South East Asian countries in the Chinese entertainment mecca, Macau. I have to say I found the rapidly evolving perspectives toward Cloud were just as impressive as Macau’s changing skyline!
Only one year ago, many of these CIO’s were using public cloud services exclusively for application development and testing purposes. Now a good number of them are running select applications with production workloads. So why the change?
The Mobile Infringement Notices trial run by the New South Wales Police in Australia provides a hint. This project demonstrates that even for applications that deal with highly sensitive personal information, like vehicle registration and license details, public cloud can be useful in parts of the workload. In this instance delivering middleware processing with scale, without compromising government regulations.
Although the majority of enterprise cloud workloads today are done in either on-premise or managed private cloud, things are changing quickly. The nexus forces of greater competition amongst the public cloud (as seen by recent Cloud Storage Wars) together with cheaper WAN bandwidth in metropolitan areas are making Hybrid Cloud a more cost effective way to platform applications. Even our very own Chief Economist, David Merill recently blogged we have reached a cross-over point where the economics just make sense.
But another macro trend propelling this rise in Hybrid Cloud adoption is mobility. Although many people associate mobility with app delivery through smartphones and tablets, Hu Yoshida describes this more broadly as a transformative way in which we work. Beyond endpoint devices, mobility encompasses the data within datacenters - on premise, off premise, across metropolitan boundaries or even continental borders.
Hybrid Cloud is gaining momentum, with some surveys suggesting that 70% of organizations are either using or evaluating it and analysts predicting it will represent as much as 30% of workloads in the next 4 years. To truly reap the benefits of combining the best of Private and Public Cloud, it's vital to tackle a couple of important attributes.
A recent blog by Colm Keegan at Storage Swiss highlighted four obstacles to Cloud Storage adoption which really echo what I have heard from the CIO’s. So in this blog, I would talk about how Hitachi Data Systems and our recent enhancements to our Content and Mobility portfolio helps address them;
- Integration. The fact is most legacy applications don’t talk the cloud language (like REST and S3) making most public cloud services difficult to consume. As Colm rightly points out, the process of re-coding legacy applications is not for the faint hearted, it can be resource consuming not to mention expensive proposition. Hitachi Content Platform (HCP) Version 7 now supports adaptive tiering to a number of public cloud services like Google, Amazon, Azure. By presenting standard CIFS and NFS interfaces, HCP acts as a seamless gateway between legacy applications and your cloud based repository.
- Performance. Although the instant provisioning and bottomless capacity attributes of public cloud services remove a big operational burden for administrators, they come at the expense of high latency. The performance penalty for reads and writes renders it unusable for applications which demand high performance. However many of these applications only need a small subset of data for short periods of time. Therefore technologies that employ clever caching algorithms which shuffle data between a local device, like Hitachi’s Data Ingestor (HDI) and a cloud storage repository like HCP over a WAN gives you the best of both worlds.
- Security. Despite overcoming many waves of resistance over the years, one that still remains in pure public clouds is the question of security. With horror stories like those experienced by Code Spaces a few weeks ago, CIO’s quite rightly question the risks of putting certain workloads into open environments. With HCP used in a hybrid cloud topology, sensitive data sets can remain on-site while less important data can be archived to the public cloud, all whilst being managed by a consistent set of tools and policies. Another benefit is having full auditability of access, something that is not offered by public cloud storage services. Furthermore, encryption can be managed in-house within HCP instead of the public cloud service provider, giving you an extra level of protection. And when it comes to file sync and share, HCP Anywhere delivers a 100% on-premise solution with the latest version 1.2 including additional security features like ICAP support, MDM, single sign-on, link management and de-registration. But don’t take my word for it, in February Gartner scored HCP #1 in the Security category of its Critical Capabilities.
- Cloud Lock In. In the world of traditional IT procurement, organisations protected themselves against lock in by purchasing products and solutions which worked to standards and created an environment for vendors to compete for their business. It's only natural to want the same when investing in cloud services. Another genuine concern amongst CIO’s is once I put my data in, how can I get it out? The collapse of cloud storage specialist Nirvanix sent shockwaves through the industry last year giving customers only two weeks to move out their data. Although this is an extreme case, the fact is you may want to move data because of lots of different reasons – cost, latency, regulation, sovereignty or just better service from another supplier. With the latest version of HCP, you can use adaptive tiering to seamlessly migrate data objects between service providers all without disruption and maintaining the balance of choice.
Here in Asia Pacific, the public cloud landscape is becoming increasingly competitive many players established up in major locations around the region. With vendors investing heavily into integration of their private cloud solutions, the popularity of Hybrid Cloud is set to jump.
So what plans do you have to use Hybrid Cloud, and what workloads do you think are best suited to this environment?