Cloud computing conference report
What would a Enterprise, in particular one using virtualisation, take away from a Cloud computing event? What if the speakers were from Cisco, Yahoo, Google, Microsoft, Baker & McKenzie and Deloitte Digital? Well two weeks ago I went to such an event, took lots of notes and engaged in some interesting discussions. Here is what some of what occurred and my updated thoughts on the Cloud space.
The event was held by Key Forums and was help in Sydney, Australia on the 3rd of December. Billed as a one day, comprehensive conference, the expectation was
This conference will kick start your Cloud strategy and will get you up to speed on what the main players, critics and users think about the potential of Cloud Computing. It will give you the opportunity to discuss your issues surrounding Cloud Computing with your peers and our expert panel of speakers.The speakers were local heavy hitters and an impressive list. Some of their presentations are available online and worth a look. The Google one was originally public but has been pulled at Google’s request. Of course the presentation from the legal firm was never public, go figure.
- The scope and potential of Cloud Computing, presented by Kevin Bloch, CTO, Cisco
- Living on a Cloud: Gaining leverage without the costs, presented by Keith Marlow, Chief Architect, APAC Region, Yahoo!
- Why Cloud Computing has moved from a risky proposition to a viable business opportunity, presented by Jason Senn, Google
- Thinking strategically about Cloud Computing , presented by Anna Liu, Group Manager - Emerging Technologies, Microsoft.
- The legal liability of Cloud Computing: Privacy, data retention and compliance, cross border transfers, contracting, audits, security and disaster recovery, presented by Allison Manvell, Associate, Baker & McKenzie
- Innovating in the Clouds: Driving new business models and innovation, presented by Peter Williams, CEO, Deloitte Digital
What were the highlights then, as there is way too much to report on in detail in this forum and format?
- All of the vendors see a strong growth in the various forms of Cloud computing. To paraphrase Anna Liu from Microsoft, “You either embrace and anticipate these changes or get left behind”. For example Cisco are one of the biggest users of Salesforce.com, the NY times used Amazon and Hadoop to convert 4TB of images to text, in just 24 hours, Gmail has over 10 million paid accounts and Cisco now own Webex (a SaaS based app) and its in their list of top return on investment acquisitions.
- There was a reasonably common view of what cloud computing is and is not, yet each company had a particular slant to their area of the cloud (which you would expect). The breakdown was typically your SaaS, PaaS and IaaS. From the IaaS site Cisco were certainly the most strong as you would expect, although it was termed “hosting” which was a little strange. In terms of alignment with VMware Cisco was very close, even mentioning the recent visit my Paul Maritz and discussing hybrid cloud of on and off premise.
- Both Cisco and Yahoo spoke about virtualisation and the changing space of the server world. Keith Marlow from Yahoo! made the obvious but insightful comment that the overhead of virtualisation was a constant and due to the increasing power of processors this constant was becoming irrelevant. Keith showed a picture of the Yahoo! datacenter with 20,000 nodes. With all of that processing power sitting there, it makes economic sense to sell some of the capacity. Kevin Bloch from Cisco spent some time talking about virtualisation and increasing core counts and how this was changing the game in data centers.
- Cisco made a good observation that WAN acceleration technologies are going to be important as clouds are built and federated with more and more data flowing around. Of course WAAS was mentioned.
- Security was mentioned multiple times as issues being presented as a concern by the market place. Whilst everyone acknowledged and respected this, the responses were generally aligned. The view was that currently people trust their banking and credit cards to the Internet and forms of cloud, so we are already starting to see acceptance. Comparison was made to the level of skills, quantity of people and attention to the problem that the providers give compared to the usual very small set of security staff within an organization. (However as one college reminded my today, they have much larger attack surface to cover too).
- The concept of access from any where at any time from any device came up a number of times. This links in very well with the VMware vision for View and access to your desktops, applications and data from anywhere at any time.
- Openness was certainly a theme from Yahoo! and Google. Open APIs, or Open Standards, being able to embrace and extended were seen as important.
- Certainly the presentation from Microsoft was the best, from my view anyway. Anna was not only a good speaker but had some good insights into the space. I recommend flicking through the presentation and I would love to hear her speak again on the topic. Some great comments such as just because you have control of an SLA, that is, its in house, does not mean that the SLA is any better than it might be if you don’t control it, such as in the cloud. A good contrast was made between control and economies of scale. A car gives you lots of flexibility for transporting things anywhere you want but does not have as good economy of scale. A freight train has great economy of scale but comes with a constraint of controlling flexibility. This balance of control vs scale was compared to build vs buy and on premise vs in the cloud. The Microsoft way was to move to services from the cloud (SaaS) or rewrite applications (PaaS). Around PaaS the view is strongly aligned with .net and hence Azure. Anna indicated that moving is “non trivial” and that there is going to be on premise and cloud models for quite a while.
So if this conference was to help one understand and think move about the Cloud computing space, what influence did it have? Did it change or enhance my thinking on Cloud? In some ways yes, and I recognize that I too come at cloud from a very specific angle.
Here is my current positioning.
SaaS is going to be a massive market. However in the main this adoption is going to be more in growing new services that are tactical rather than the main game of organizations core business processes. Yes Cisco may move to Salesforce.com but for most enterprises core elements are going to stay in house. It’s the new systems and non core that are going to see the most growth in SaaS.If you are in Australia and interested in cloud computing like I am. I would recommend you put March 25th, 2009 in your calendar. IDC are holding a Cloud Computing Conference on this day in Sydney. See, aren’t you glad that you read all the way to the end of this post! You can register for free as an early bird attendee.
When that project team needs a new intranet site, rather than waiting two weeks for IT to not deliver, they will put their credit card number into a website and be up and running in 10 minutes.
The challenge for VMware here as well as the opportunity is to capture some of this space with Virtual Appliances. If a mature set of Virtual Appliances can be available through the market place; if these can be downloaded from within the management environment painlessly; and if they can not only describe their service levels, backup and disaster recovery requirements (vApps) but also implement them automatically (AppSpeed, Data Recovery Appliance, SRM) VMware could be onto a good slice of this pie. Put your credit card in, download the service and run it locally in your own security model, owning the data. You can even still have the provider maintain the application as part of a maintenance agreement. Most of the benefits of SaaS without many of the current concerns and limits. VMware need to lead by example here.
PaaS is the challenging space. Microsoft is going to push .Net real hard, and that means Azure. The challenge for all of the players is to support open languages such as Hadoop and Ruby on Rails. This strategy is good for customers in the enterprise. If they develop their own applications in these open languages they can execute them internally on their own clouds. After all a key element of VDC-OS is running the workloads of today and tomorrow. With VDC-OS you can run mixes of work loads and change them on the fly, one day 10% of your cluster might be running Hadoop nodes and tomorrow, to scale up for a specific project work load, it may be 30%. Even better because you have written to an open standard you can go to a provider in the open market to buy capacity for short or medium term, if you really need to scale up. Even if you run that application out in the external cloud, if there is a problem, or for testing, or for DR, you can always run it in house on your own cloud if needed. VMware need to work hard here to not let the ISV market get away from them. Maybe that’s why Paul Maritz can’t stop saying Ruby on Rails.
IaaS is VMware’s sweet spot. The enterprises know they need to move to the benefits that cloud and utility based computing can bring. They want to run like Google and Yahoo! The challenge is how to do this in today’s environment and that’s where VDC-OS, vCloud and vApp come into play. Running internal clouds, federating them, running the workloads of today and tomorrow, its not just a dream, its like Christmas, you know its coming and its not far away. The problem for VMware is to not been seen like Azure and be another closed shop. That’s why it’s good to see vApp being based on OVF. We know VMware is the best system for running workloads, if we can keep the portability of workloads it means the best technology wins. If it all goes closed, the best marketing company wins, and that is not VMware. Also VMware need to keep tight (like they already are) with complementing technologies. Networking is going to play a huge role in enablement of the cloud, from things like split VLANs to WAN acceleration.
If you have some thoughts on the Cloud space, leave a comment.