OpenStack Summit 2014 - Day 2

Atlanta, Georgia: neighbor to Gainsville where it’s illegal to eat chicken with a fork, where the official state fish is the largemouth bass, and home of the 2014 Openstack Summit.

Day 2 was opened by Mark Collier, OpenStack Foundation COO, who talked about the importance of speed. After making some humorous references to the hit tv series Breaking Bad, Mark talked about speed, not just in application run time but, in terms of time to market for new services and time to react against change in infrastructure requirements.

Collier went on to say that operational agility has made competition more aggressive, "At the current churn rate, 75% of the S&P will be replaced by 2027.". Insisting that if you’re working for a big company now you need to be implementing OpenStack to increase agility and remain competitive. He then introduced Toby Ford, SVP of IT Operations Strategic Realization of AT&T.

When Ford joined AT&T open-source was pretty much forbidden, which was surprising given their long history and involvement with Unix. However, seeing its potential to improve business, Ford lead their early adoption of OpenStack.

AT&T focus on minimum vendor relationships - many of the existing partnership dating over 100 years. “It's difficult to become an AT&T partner, which has its problems. But OpenStack has become a proxy to get innovation into AT&T.”.

Already deployed in seven data centers today with three more deployments by the end of year, its reliance has expanded from a few back-end systems and gateway API's to, more recently, Big Data services and some other larger scale use cases.

After these first ten OpenStack data centers deployments there are plans to jump to 20+ for their broader NFV use cases.

Inspired by what Netflix was able to do in such a big scale and they are taking this implementation philosophy to their TV own offering.

Through OpenStack and, inherently, data center agility, AT&T are driving towards multi-tenancy to improve efficiency with their framework for a multivendor cloud ecosystem being based on Openstack – “the heart of NFV”, said Ford, describing ecosystem as having the attributes of: Plug-ability, Programmability, Innovation and speed, Reliability and security, Telco expertise & features, Distributed and scalable, Cost & Performance optimized hardware.

Now they are looking at how to virtualize switching and routing but not just an implementation of SDN. “This is bigger than that and includes application services.”, said Ford.

When asked what OpenStack challenges they are looking to address he mentioned the need for an Openstack distributed deployment, to communicate across numerous sites. Ford closed with, "Imagine how many locations we can put servers and move content closer to its insurers".

Next up was Joel Johnston, Platform Architect Global Hosting at Sony Entertainment who started his presentation by promoting a PS4 baseball game after which Mark Collier asked the obvious question, “What’s this got to do with Openstack”?

Sony run their user community infrastructure on it for both in-game and out of game hosting and communication. This includes game scoring and rank data, and for the intelligent pairing of on-line gamers based on their player data, history skill level.

Joel, advised that in the past, the ticket to provision new services was a couple hours to a couple of days, but now, with this OpenStack platform, they can provision by themselves delivering faster time to market and much happier developers, improving both new feature developments and time to react to increased gaming activity. “So we're connecting people from one basement to another all over the world?”, enquired Collier.

In response to what’s important to the now, Johnston replied, “In place upgrades!”. They want the ability to upgrade in production deployments with confidence - from Nova Network to Neutron, for example. This is the type of thing he is looking for now – agility within OpenStack in addition to the agility derived from OpenStack.

Following we had Guillaume Aubuchon, CTO of DigitalFilm Tree, who, looking at lot like Zac Galifianakis, requested two ferns be brought out and places behind himself and Collier and proceeded to conduct an interview discussing topics including:

 

• Origami as a form of Software Define Paper.
• The fact that public clouds killed 20 people last year yet, open clouds killed none... Or, maybe it was one. 
• SWAS - Sweater as a Service. IT professional are exposed to hyperthermia everyday in data centers. They want sweaters to be available in all DC's around the world at just an arms reach.

Moving on from the comedy, Guillaume discussed their use OpenStack to rapidly deploy infrastructure for the camera acquisition through to digital production for show's like NCIS and Nip/Tuck.

DigitalFilm Tree set out to try and change the way the process works. Standing up their first OpenStack deployment with just 4 people and in 6 months, it's now used in almost everything they do - even running on-site instances of OpenStack for editorial implementations. They call this their ‘Software Defined Post Production”.

Next was Dave Wright who left Rackspace in 2008 to start SolidFire and is CEO of the orgainzation. “Public cloud is about speed, and time to value.”, he shared. Enterprises are attempting to replicate the speed and agility of public cloud. They have learned that the technology silo approach is dead, that the entire architecture of the data center is changing and it’s all driven by a shift from manual provisioning to automation. Wright then explained that this has lead to a fundamental change in application design. Now they can span hundreds or thousands of VM's machines with service-to-service communication and mobile interfaces.

“The entire DC has become a giant computer. It has programmability, and if the DC is a giant computer, then OpenStack is its operating system.”. said Wright before welcoming Subbu Allanaraju and  John Brogan of eBay to the stage where the two explained that agility drove them to look at cloud, for ops and for developers with a vision, that any team in eBay should be able to request resources from IT and it should be readily available.

OpenStack delivered benefits to eBay both internally and externally. Its how they've broken away from technology silo's internally. “OpenStack is also a career development opportunity for engineers” said John, “Storage guys becoming more knowledgeable about networks, for example. This never would have happened before.”. The 3rd thing eBay got from Openstack is that it's now a gateway to get into their infrastructure. Instead of working around vendor strategy’s they now say to vendors, tell us what your OpenStack story is. The vendors now must play by eBay’s rules.

Mike Werner from RedHat and Steve Stover from Dell were then welcomed to the stage where they discussed their alliance with SolidFire and their tested and ratified software defined infrastructure.

Last to the stage for the morning keynotes was founder of Canonical, Mark Shuttleworth. He took the approach of an ecosystems for all to succeed in. Suggesting that while many vendors in the audience believe that much of their success will come from Openstack, but also, Openstack itself will only be successful from their involvement.

Yesterday (May13th) Canonical announced OIL, their OpenStack Interoperability Lab. With 14 supporting vendors signed up they are constantly building automated clouds and testing different deployment configurations with these new partners for their joint customers. Shuttleworth said they generate a Cloud every 15 minutes using Juju, an OpenSource cloud automation solution.

Shuttleworth closed with a live demo – the only one of the keynotes so far.

With the closing out of the keynotes, we make are way – via the extensive queue at Starbucks back into the expo hall for another day of conversations about making things better. Bring it, I say!