Cisco’s Intercloud: Policy, momentum, and the Internet of Things
Wed 12 Nov 2014
In the second part of our interview with Nick Earle, SVP Cloud & Managed Services Sales at Cisco, he discusses Cisco’s take on policy and considers how emerging technologies will influence the cloud marketplace.
The first technology enabler we’ve got is Intercloud fabric which enables workload portability and security across any cloud. The second enabler is policy replication. Compliance and security policy are set in multiple parts of your IT stack, at the device level, the network level, and the data centre level. Typically what happens is that most IT departments are organised by how the IT industry used to look; there are server groups, storage groups, network groups, application groups, security groups, and one of their roles is to ensure policy is applied to applications. That’s a very slow process, about six weeks for all of those policies to be applied.
Independent of Intercloud, we announced ACI (Application Centric Infrastructure). ACI allows you to set policy at the Application level. We’ve abstracted policy up so again you have a rules-based engine.
First of all we created this technology based on a controller called APIC (Application Policy Infrastructure Controller). I can connect policies through that controller in about 5 minutes, as opposed to 6 weeks. With Intercloud, if the workload moves clouds, the policy gets replicated.
We’ve now got the ability to do two things that have never been possible before. Set policy at the application level, which is where the explosion is, and then replicate that policy automatically across any cloud, private or public, and any cloud deployment model, hosted or hybrid.
If I’m a CIO I now don’t care where the employee stores their workloads as I have full security and full policy control. That’s what Intercloud is – the ability to do that.
We are using a partner ecosystem so that we can have data centres in as many countries as there are partners. We’re not using our balance sheet, so we’re not going to run out of cash either.
The challenge with the IoT is that it’s not actually about the things. Information about things is more valuable than the things. What you need to do is to get the information from all the things, at the edge of the network, but you can’t replicate it. Right now all the hot action is around Hadoop clusters, and enterprise data warehouses, which are all a form of replication, but as the number of things starts exploding getting data from those things, is a huge challenge – people would have to replicate the internet inside their own company which isn’t going to work.
What you want to do is to be able to change the paradigm for Big Data. Instead of grabbing data, replicating and then analysing, you leave the data where it is and you query real-time – that’s called data virtualisation. We believe we have the leading offering in that area.
This model and these technology enablers – this whole thing about recreating the internet in the cloud, is a massive project. We started talking to cloud partners and providers; so far you’ve seen us announce seven from a number of months ago but since then we’ve been quiet. For instance we have Telstra in Australia, who have put all their cloud business on us. We also announced Citrix who run DaaS with us. We announced Dimension Data, who have got about 18 data centres which are going to offer Cisco-powered IaaS.
After announcing the seven we got some feedback asking, is that it? I can say is that in the month of September, there will be a significant increase in the number of partners. More countries, more data centres, more partners. People want to start sharing workloads, or brokering. The name of the game is moving to workload brokering. Our view is that workloads will be brokered across clouds.
Let’s take for example an employee who needs to look at data in Germany, Australia, and the US. Due to data sovereignty laws, they want to keep the data in Australia, they definitely want to keep the data in Germany, and they want to keep the data in the US. So instead of trying to find one service provider who will handle all of that, they can broker the workload using the Intercloud fabric in APIC. The workload gets brokered or federated, and the data gets stored separately, but is still one experience from the employee’s point of view.
We believe that OpenStack will be the predominant model and we’ve done a big deal with Red Hat. For example, we’re managing Telstra’s OpenStack environment for them. We believe that workload will start to be shared across a curated catalogue or marketplace.
Let’s say hundreds of companies, service providers, and clouds were able to connect and move things seamlessly. What will happen is a marketplace will appear, and it will exist on top of the clouds. People will buy and sell services from the marketplace, and things will connect because it becomes like a black hole, it has the most gravity. Companies will connect to it because they can go to it knowing that they still have security and policy. We can guarantee security for Intercloud because we manage the stacks.
This global marketplace for buying and selling is a massive partner play. Partners, instead of having to rely on reselling hardware as they do today, will start to resell managed services. But these could be somebody else’s. I could be selling your services, for example you put them on the marketplace but I sell them and you give me a fee or vice versa.
There’s going to be a cloud marketplace or a big supply chain. Whoever has the most supply and can create the most demand, will create the biggest cloud marketplace and that is where we believe cloud will go.
We have 62,000 channel partners, every SP you can think of runs their core network on Cisco. But we’re not just going to bring SPs in, we will also allow them to leverage their investment in their core network to do what is called NFV (Network Functions Virtualisation). This allows them to create higher value added services and storage.
They can do virtual CPE, so things like Meraki. If you’re Telstra, or a big cloud partner, you will be able to run Meraki, our Network as a Service startup. As a partner, you’ll be able to take Cisco’s catalogue of managed services and run them inside your data centre.
We’re not just distributing the hardware, we’re distributing our entire portfolio of software to our partners and that’s called CCS (Cisco Cloud Services). We plan to seed the marketplace with our portfolio of managed services.