Over the past two weeks Xpand in conjunction with Australasian Architecture Network has hosted two Q&A panels in Sydney and Melbourne around the topic of Predicting the Future of Cloud.
Here at Xpand we fully believe in having a deep passion for our markets and we were delighted to welcome our amazing panel and guests for bothÂ nights.
- Andy Pattinson, Senior Sales Director for Cloud Sherpas
- Bruce Haefele, Chief Architect for Healthdirect Australia
- Nicholas Bellamy APJ General Manager, OCTO Innovation & Advisory Services HP Enterprise Services
- Nigel Watson Technology Partner Lead, AWS ANZ
- Nicholas Bellamy APJ General Manager, OCTO Innovation & Advisory Services HP Enterprise Services (Back for more)
- Saul Caganoff, CTO of Sixtree
- Shiva Narayanaswamy Solution Architect Amazon Web Services.
- Dan Kregor Microsoft Cloud and Datacenter Management MVP, Senior System Centre Consultant and owner of the Inside Podcast Network (ipn.tv)
A huge thanks to Nicholas Bellamy (below left), APJ General Manager, OCTO InnovationÂ &Â Advisory Services HP Enterprise Services and Saul Cagnoff (below right), CTO of SixtreeÂ for annotating some of the outcomes and discussion points on the night;
Â Â Â Â Â Â
1. What defines a cloud service?
Nick: A catalogue of easily consumable resources, encompassing IaaS through to XaaS, be they classed as Public, Private or Hybrid Cloud.
Saul:Â Well I could be boring and refer to the NIST definition of "cloud services" and enumerate the attributes: on-demand, pay as you go, elastic and multi-tenant. But my litmus test for cloud involves the credit card. Something that has revolutionised the way I do IT is the ability to buy infrastructure on demand for short-lived activities - development, testing, prototyping. If I can't swipe my credit card and get instant access to something that I can later shut down, then it ain't cloud.
2. Cloud Broker or Cloud Provider?
Nick:Â We see the evolution of the Cloud Provider to be a Cloud Broker and niche players becoming Cloud Brokers all the way up the stack and across the environments and platforms of enterprises. Agile players with robust Delivery models will be the most successful Cloud Brokers.
Saul:Â We're in an early phase of cloud were the bulk of the players are "providers". They provide IaaS, PaaS or SaaS, mostly in a standalone capability. But as the cloud matures and more mainstream users come on board, I think we'll see the rise of the cloud brokers. Brokers will package up cloud services and make them work together, perhaps aimed at particular verticals or work processes. About 20 years ago a small company started doing this with open source software - picking the best applications, making sure they worked well together and packaging them up for use by mainstream businesses. That company was Red Hat and they were the original open-source software broker. I think that will happen with cloud and particularly with SaaS
3. Is there such a thing as a private cloud?
Nick: Yes, these instances will be few due to the higher cost, generally driven by regulatory policy and compliance certification i.e. Australian Signals Directorate (ASD)
Saul:Â There is such as thing as having on-premise virtualization and automation so that your infrastructure can be on-demand, pay as you go etc. But you are not really multi-tenant, and importantly you are not really elastic. Or there is a limit to your elasticity. So strictly speaking I don't think private clouds are really clouds. But ultimately I don't think this question matters too much.
4. Where does Hybrid Cloud sit in an Enterprises Evolution
Nick:Â Critical, a Hybrid Cloud environment is the most common â be that hybrid to ensure vendor contestability or the ability to agilely deliver business solutions quickly from Sandpit, Development & Production. The business can once again truly utilize IT and collaboratively drive better business outcomes.
Saul:Â Right in the "here and now". All enterprises have on-premise infrastructure. Most enterprises are now in the cloud at some level and inevitably there is the need for on-premise and cloud assets to interoperate. So hybrid is incredibly important at the moment. At Sixtree we embarked on helping businesses do hybrid cloud a couple of years ago, and we are now incredibly busy. So I know it's true.
5. Is your App really Cloud enabled?
Nick: Whether you are re-hosting, re-architecting an application the workloads of the future will need to be State aware to take full advantage of what the Cloud can provide and therefore enable fundamental shifts in how business and users can drive meaningful change to the way we have historically conducted business.
Saul: New platforms often start out looking like old platforms. Early automobiles looked like buggies without the horses - they were even manufactured the same way. Early TV sets looked like radiograms with a tiny little CRT built into the wood panelling. I think the same thing is happening with cloud. IaaS feels a lot like my on-premise data centre. Many of the cloud CRM and ERP systems feel a lot like my on-premise CRM/ERP, but just in someone else's data centre. Practically, a legacy architecture can limit an Apps ability to be elastic or to be truly multi-tenant. There are new Apps emerging which have a cloud native architecture. Netflix is a great example and Adrian Cockroft, the ex-Chief Architect at Netflix tells a brilliant tale about how cloud-native is based around microservices running on top of "ephemeral" infrastructure. Cloud-native architectures provide true cloud characteristics. Right now we are seeing many App vendors working at splitting up their monoliths into microservices heading toward being cloud-native, elastic and reliable.
6. Do cloud services mean the end of IT as we know it?
Nick: No, but it does drive a new way we deliver IT â be that from the economics of an âaaSâ model to CMOâs now wanting to work with the business and IT to agilely deliver seamless user experience and interactions.
Saul:Â IT is already changing very rapidly. I dimly recall the last IT paradigm change in the '80s when we migrated off Mainframes into client-server architectures and desktop computers. Back then, "shadow IT" was bringing your IBM AT into the office from home. This change drove a restructuring of the way we did IT. New application architectures emerged - two-tier, three-tier and ultimately N-tier architectures, and in-turn we changed the structure of development teams. Those architectures were influenced by scarcity - scarcity of compute power, scarcity of storage and scarcity of bandwidth. We are now in an era of abundance and not just because of cloud. Data storage comes in a variety of shapes and sizes, bandwidth is pretty good and compute power seems endless. This era of abundance makes possible new ways of building systems and new ways of organizing teams. We're witnessing fundamental shifts in the way people do IT and in the way that IT supports business.
7. What is the future for desktop applications?
Nick: The use of the term âdesktop applicationâ is really a legacy term for the way we now deliver applications, via the browser to multiple devices â laptops, slates, smart phones etc.
Saul:Â I'm not really sure. The browser is becoming the new operating system and I'm attracted to the idea behind ChromeOS but I think we've got a way to go in terms of operating web applications in a disconnected mode. The desktop is already augmented by the cloud and I think it will become more-so.
8. What happens in the event of data loss?
Nick: The same as with Legacy but with a new increased resilience depending on the type of service you choose i.e. Business Continuity, Disaster Recovery or true High Availability.
Saul:Â Risk management has always been a factor in how we architect systems. What happens if your disk crashes? What happens if the network partitions and what happens if your data centre is hit by a tornado? We already design our systems with such concerns in mind and that doesn't go away with the cloud. Ultimately we are responsible for the integrity of our own data and that doesn't go away. Perhaps we need to be aware of additional factors such as our data being locked up in a proprietary SaaS system, or what happens when a provider goes bankrupt? Caveat Emptor still applies.
9. How easy is it to migrate to a different provider?
Nick: The Transition & Transformation of data is still the crucial step on how easy is it to migrate âthis is ties in as a financial and risk consideration.
Saul:Â Very difficult at this stage. I'm sure that affects quite a few cloud or on-premise decisions.
10. Are we relying too much on SLAâs?
Nick:Â No â SLAâs are becoming easier to digest and are now able to be âboltedâ onto existing MSAâs.
Saul: My answer here is the same as the one around data loss. We always have the responsibility to architect our systems to be resilient against failure. The good news is that we're getting pretty good at that. Netflix has its Simian Army which randomly kills processes, hosts or even entire AWS availability zones - and in production, no less! New application architectures (such as microservices) and new ways of managing systems help us develop anti-fragile systems which get stronger under duress.
11. Does the cloud really enable anything new?
Nick: Yes, a new culture that drivers greater collaboration and the ability to âfail fastâ with new revenue generating concepts.
Saul:Â Absolutely! cloud takes away a lot of the barriers to entry. It gives us velocity - the ability to change rapidly and to fail faster, with lower cost. That ability drives innovation which - when used right - enables everything.
12. What are the biggest challenges you'll be facing in the next 5 years?Â
Nick: The ability to stay agile in an increasingly niche market â our channel and partner environments will be key to driving some of this success.
Saul:Â As usual, the technology is the easy bit. The biggest challenges we face are skills and culture. The way we build systems is changing rapidly and people are comparatively slow to change. There is a learning curve - even the most enthusiastic of us have to learn and adopt new approaches. But there are also many people in mainstream IT who may be reluctant to change. There are many who are threatened by change and uncertainty. Fundamental change often takes a generation as those who "grew up" in the old ways need to move on to allow change to happen. That might happen faster than we think because companies that change sooner will displace those that are slower to change.
13. What's your prediction for the future of cloud services?
Nick: I think we will see more cross industry pollination, with a continued development and uptake of consumable platform services i.e. Analytics, IoT.
Saul:Â How much longer do we have tonight? I think we are in the early days of a huge paradigm shift and the future of cloud services will look nothing like we imagine. We still have mainstream adoption ahead of us and the mainstream are consumers of cloud, not providers. So the future will be less about IaaS and more about SaaS and a little PaaS. Much of the way we use cloud today is in isolated silos. Cloud providers don't interoperate very well, or at all. Some providers like Salesforce.com are starting to stitch services together, but piecemeal, point-to-point stitching isn't going to scale well. The growth of programmable APIs and cloud integration platforms allow cloud services to be stitched together to support end to end business processes. It's early days yet but these platforms will become more sophisticated and specialized until individual businesses can leverage and integrate commodity SaaS capabilities to express their own particular business value. Jonathan Murray, CTO of Warner Music Group calls this concept "the composable enterprise" and there are some organizations such as News Corp. Australia who are already heading down this road. Something we often forget as cloud providers is that the mainstream view of the cloud will be more about consuming services than providing services.
Our next Q&A panel events are on Wearable Devices and the Internet of Things (IoT) and DevOps register below