Sramana Mitra: Let me try to get clarity on exactly what I am asking. I understand the application acceleration and the cloud acceleration value proposition, when you are dealing with an enterprise’s private cloud, private data center, and so forth. If it’s a public cloud environment, then let’s say you have the Global 3000 using a Salesforce.com service. Where do you come in to that picture? Isn’t the onus of delivering the application at the right speed on Salesforce.com?
Willie Tejada: Well, let’s use Salesforce.com as a direct example, because we embarked on a relationship that we announced with Riverbed to accelerate that exact environment. In this case, when you consider the network configuration the topology for the enterprises, it is not unusual for someone who is accessing Salesforce.com to have to connect through the private network and through a handful of Internet gateways that are part of his private network environment.
So, consider it is not an unusual configuration for an enterprise that currently has thousands of branch offices to have Internet gateways in fewer than 15 to 10 locations, which means that all that user community has to back call their access. The byproduct of that is a very long link they have to traverse to get to their Salesforce.com instance. This link creates performance and reliability problems as it traverses over the public and private networks to get to the end user. The solution that Riverbed and Akamai are building together is to optimize that link. That use case is that as more people increasingly want to access things like Salesforce.com, they are accessing it from the private network environment. While the responsibility certainly is on someone like Salesforce.com to deliver that user experience as best as possible, typically, they can only reach the doorstep of where that Internet gateway is.
SM: Even in a public cloud use case, your client is still the Global 3000?
WT: That is correct. We put something like Salesforce into two categories, because certainly they have their traditional CRM team, their Salesforce.com users. Then they have things like Force.com or Siteforce. So, there are really two types of services and businesses that Akamai focuses in on. The area of something like customer using Salesforce.com, what we look at is, How do we optimize that? We sell specifically to the enterprise to try and solve that problem for them. Part of that is because their enterprise network architecture is something that also contributes to the performance problem that is part of their accessing Salesforce.com. In the case of something like Force.com, it is typically the enterprises. We classify the use of any one of those platforms as a services as people delivering applications as they’re built organically in the cloud itself, whether you are talking Microsoft, Azure, or Force.com.
The effort that Akamai has put forth in that is to work directly with the cloud provider so that value-added services such as application acceleration or security and protection services can be selected dynamically at the timethey stand up the application. It is similar in many cases to what Amazon provides, the basic utility in relation to something like CloudFront, but Akamai has done a number of these in-network services from dynamic application acceleration to web application firewalls in the cloud. We believe that those services should be available directly from the cloud provider when they stand up the application. So, we are working closely with the platform as a service vendors, and even the infrastructure as a service vendors, as they select these value-added services for the applications that inherently start in the cloud. They’re developed directly in the cloud. Most of our other services are for those applications are still in the customer’s data center or some mix thereof.
SM: When you are working with these Global 3000 customers, what are some trends you are seeing in terms of cloud infrastructure adoption and their needs and open problems?
WT: That’s a good question. I think it really depends. In the Global 3000, in many cases what we have seen that it depends on the vertical or the customer segment in terms of where they are in their adoption or their openness of the use of cloud computing. I use the term cloud computing here primarily [to mean] Internet-based or outside of their own private data center cloud computing.
An example I use is that a common practice or trend we are seeing in our segments and verticals such as healthcare and financial services is that the data will reside behind the firewalls in their on-premise server. These customers are looking very much to put the front end of the application outside and use the scalability of the cloud. Doing so provides them with the ability to deal with a flash increase in terms of their workloads and dealing with their workloads whether they be app, computing, or storage workloads. It creates challenges in providing the same kind of performance and synchronization of having the front end of the application exist in the public network side and the data in their private network side.
More often than not, in financial services, they do not see that data moving directly into the cloud any time soon. Some of it, in some cases, has to do with compliance and regulations. In some financial services institutions, once they have data they have to keep it for for more than five years. So, in some of those cases it makes sense for them to own the data set side of it.
But they are trying to take advantage of the agility and the right cost savings that happen when they can leverage infrastructure as a service and, for that matter, platform as a service in the cloud. On the other hand, when you take a look at companies that are primarily in the media space or in the rich content adoption space, companies like Adobe and Autodesk, some of the ones we have seen from that standpoint, they are looking in relation to their adoption of the cloud and strength their content in the cloud are, and they are looking for the best cost or computer storage, whether it be inside their private data centers or out external in the public data centers. We see most of the financial services and healthcare companies really doing a lot of testing and development in the cloud and looking at how they bridge over security and some of the issues related to performance and connectivity between the public and private cloud environments. We see rapid adoption in segments where they are using the cloud primarily for storage. They are much more comfortable with placing digital assets and things of that sort directly into the cloud. So, the systems of record have made slow migration into the cloud. Maybe many of them still in testing and development, where much of the collaboration that companies would do collaboration or a content of media applications, are doing much faster adoption actually over that.