What do Virtualization and Cloud executives think about 2011? Find out in this VMblog.com series exclusive.
Contributed Article By Anand Babu (AB) Periasamy, CTO and co-founder of Gluster
Prediction 2011 – The Future of Virtualization
Cloud to Redefine Operating Systems
In 2011, the operating system (OS) as we currently know it will be categorized as a legacy system.
Virtualization and the cloud will cause fundamental changes in the design and composition of the OS, once the innovation reaches a mature stage. We will see the first evidence of this evolved OS in 2011, witnessing its ultimate realization over the next five years. Red Hat's acquisitions of JBoss and Makara, VMWare's acquisition of SpringSource, and Salesforce's acquisition of Heroku are good signs of what is to come.
Desktop OS: Android and iOS replacing Win 7
Microsoft's desktop domination will start fading. Fat, desktop-centric applications will transition into the cloud at a much faster rate. Tablets and netbooks will outsell workstations, desktops and even laptops. Everything from silly games to serious enterprise applications run on a browser already. Google's Linux-based Android and Apple's iOS will put the final nail in the coffin. It is much easier to write applications natively on a tablet OS than on a browser platform. Tablets, netbooks and iPads will access their applications from the cloud, and that is how all future services will be delivered. Native applications will eventually go away and the majority of them will become browser-based and hosted on the cloud. Cloud enables powerful software services to be hosted on a scalable infrustructure with virtually unlimitted computing and storage resources. Without the Internet and cloud-based applications, tablets and smartphones will be relatively useless. Did you watch Google destroying a Chrome OS netbook just to prove the point that data is safe on the cloud?
After years of battle, finally the cloud will make it possible to regain control from Microsoft. Microsoft's Win 7 for tablets does not have an app store. It is simply a desktop OS with touch screen support. Red Hat and Canonical's Ubuntu are not even in this game.
Server OS: Enterprise OS to be commoditized and become a battleground
Enterprise server operating systems will be commoditized and redesigned for the cloud. Red Hat Enterprise Linux and Microsoft Windows Server will join the legacy systems club. They were primarily designed for monolithic scale-up server platforms, whereas modern cloud and virtual environments are based on scale-out commodity platforms.
The operating system kernel will eventually break into hypervisor and thin guest kernels. The hypervisor architecture will make the dream of micro-kernels a reality. Guest operating systems will be stripped down to the JeOS model (Just Enough Operating System: pronounced "juice"). Operating systems will run as collectives across multiple servers with shared services over RESTful APIs. The definition of OS will broaden to include middle tier platform components (such as JBoss and Heroku) and rich services (payment gateway, CDN, document management).
Linux/KVM will rise above Citrix Xen and pose a clear long-term threat to VMWare. Red Hat Enterprise Linux will no longer be relevant in the cloud and virtual environments, though. With everyone from Amazon and Rackspace to RightScale pushing their own GNU/Linux distributions, it is time for Red Hat to recognize that the OS is becoming a commodity in the cloud. Red Hat should embrace the cloud and take a lead by introducing a thin free Red Hat Linux for the cloud. There is enough money to be made higher up in the stack; it is important to gain market share at this stage. Red Hat should not forget its old Red Hat 9 days.
Microsoft recognized the need for a new cloud OS: Its Azure is modeled after Amazon AWS and has plans to support on-premises environments in future. Most of the services can be accessed via RESTful APIs. However, Microsoft is going to struggle against GNU/Linux dominance in the cloud.
Virtualization primarily grew in the enterprise data centers. Cloud grew on the public hosted environments. Both public and on-premise data centers will move towards a unified cloud-based model. It will be virtualized, commodity-based, multi-tenant, self-serviced and mostly free / open source.
Storage will be a software problem
EMC and NetApp have ruled the storage industry so far, much like the mainframe business. 2011 will mark a defining moment for the storage industry. On average, applications demand 10 times more capacity, 10 times faster performance and a tenth of the cost. Monolithic enterprise storage arrays will be ruled out of this equation. Only scale-out software-based solutions can deliver this 1000x improvement. Free / open source scale-out NAS filesystems such as Gluster will enable cloud applications to seamlessly manage petabytes of application data and VM images on top of commodity drives. SAN disks will be stored on NAS as virtual images. Similarly, Object Storage and NoSQL solutions will gain momentum under the Bigdata brand. Once the crown jewels of the storage industry, traditional RDBMS products like Oracle will begin their journey into a niche market. Without addressing the storage bottlenecks, cloud / virtualization progress will be futile.
Google's competitive advantage erodes
Ironically, the cloud spells bad news for some cloud pioneers. Companies like Google and Facebook, which had the advantage of being the first to create their own infrastructures and software stacks from the ground up, will watch their assets become liabilities over the next five to ten years.
Only a few years ago, it seemed impossible to compete with Google. In the coming years, Google will face more competition from a lot of smaller companies that are cooler, faster and do not have the baggage of a custom-built infrastructure. Already there are opportunities for new and existing companies to deliver Google-like services better and cheaper than Google because they do not have to build the infrastructure themselves.
In 2011, the cloud will gain real steam!
About the Author
As CTO and Co-founder, AB sets the vision and strategy for the Gluster product platform. Prior to Gluster, AB served as CTO at California Digital Corporation, where his work led to the scaling of the commodity cluster computing to supercomputing class performance. He drove the adoption of cluster computing and GNU/Linux at enterprise data centers and helped close strategic accounts at CDC. In 2004, AB led the development of the world's second fastest Supercomputer "Thunder", for Lawrence Livermore National Laboratory. AB also serves on the board of "Free Software Foundation - India". He is the author / contributor of various other Free Software projects like GNU FreeIPMI (Intelligent Platform Management Interface), GNU Garp (Gratuitous ARP Daemon), bios-config (edit/replicate CMOS parameters), librpci/hdb (RPC interpose for GNU Hurd) and Hymn/PlayFair (iTunes ripper), GNU Freetalk (Scheme extensible messenger for Jabber, Google talk), and Freehoo (Scheme extensible messenger for YahooIM). He holds a Computer Science Engineering degree from Annamalai University, Tamil Nadu, India.