Last week, Battery brought together six of the industry’s top global enterprise-IT CEOs and founders with a group of noted technology journalists for a dinner in San Francisco. The topics ranged from big data to cloud storage, mobile apps to cloud computing, and “bare-metal networking” to disaster recovery. Not everyone agreed on where enterprise technology and buying trends are going in this market environment. But all were unanimous in their support of the great food, conversation and lovely views of the Bay Bridge.
After the dinner, we sat back and posed some questions to the participating executives about enterprise-IT trends affecting growing companies today. Here are their answers.
Powered by Battery: What is the typical “tipping point” for companies seeking to migrate their applications and data to the cloud from traditional data centers? What finally makes them make the move?
Ziv Kedem, co-founder and CEO, Zerto: Many companies initially consider cloud to gain efficiency — reducing the need to build and maintain additional data centers. Specifically, “hybrid-cloud” use cases enable IT to use on-premise and cloud-based infrastructure seamlessly for cost reduction and disaster recovery. As companies get their feet wet with cloud, they will find that some basic but very important functionality required to use any cloud is missing: workload mobility, cross-hypervisor portability, seamless disaster recovery and data protection. In many cases, workloads on one data center are not interoperable with any other data center environment, so moving these workloads to and between different clouds can be quite challenging.
The tipping point for cloud adoption will come when companies see flexibility from vendors in these areas – particularly the ease of moving, managing and protecting production workloads (applications, data and associated processing power, and information) from on-premise data centers to public and managed clouds. At Zerto, we have identified the key functionality required for production workloads to utilize any cloud, namely, workload mobility, cross-hypervisor portability, disaster recovery and the encapsulation of application dependencies. These are the core building blocks needed for hybrid cloud environments to thrive.
PBB: What aspect of data-center infrastructure is the hardest to scale?
Ariel Maislos, co-founder and CEO, Stratoscale: One of the biggest challenges in scaling data centers is balancing infrastructure growth between what have traditionally been independent silos of technology. In a non-converged environment, adding more compute infrastructure (servers) means having to separately add more storage capacity, often managed by a separate group, as well as networking capacity in order to avoid bottlenecks and performance degradation. So when thinking about scaling-out a data center, one must consider how to simultaneously scale compute, storage, and networking concurrently in a balanced, economically feasible manner. Converging the technology is the answer. When a new server is added to a converged data-center, all of the components of the server–compute, storage and networking—are made available to the entire cluster, removing the need to separately scale individual silos of technology.
PBB: How are companies using Big Data today?
Ben Werther, founder and CEO, Platfora: Today, every CXO is laser-focused on delivering results for his/her organization. And the business leaders of today that are using data to make decisions will outperform their peers. As a result, we’re working with companies that are moving beyond the exploratory phase of Big Data, and are investing in three core areas:
1) Security – CISOs (Chief Information Security Officers) are tasked with ensuring that customer data is safe, secure, and that their company is protected against security threats, both internal and external;
2) Customer 360 – CMOs need to better understand the customer experience and customer touch points, and drive customer acquisition and revenue; and
3) Internet of Things – CXOs are focused on improving existing products and driving new product introductions by analyzing trends and pinpointing issues across millions of devices.
It’s never been more important to keep customer data secure, drive customer satisfaction and increase revenue. Big Data is becoming the CXOs secret weapon in driving business results. And because of the advances in technology, getting started with Big Data has never been easier – everyday business users can now gain valuable insights from their data in minutes, not months. Also, it’s important to note that IT continues to play an important role in this process, and is becoming a strategic partner for the CXO. But what we’re finding is that these decisions are increasingly being led from the line-of-business business leaders who want to achieve these capabilities, and the days where this is purely IT-purchased and implemented technology are history.
PBB: What are the roadblocks now facing companies trying to develop Big Data apps quickly?
Jonathan Gray, CEO and co-founder, Continuuity: At its core, Big Data is about scale out, and scale out means distributed systems on commodity hardware, often in the cloud. These systems are not only a difficult computer-science topic, but they remain very hard to use in practice: It’s challenging to build and deliver applications on top of Hadoop and other scale-out architectures. The biggest problem today is that most developers don’t have the expertise to build these types of applications as they are being asked to understand complex architectures and build full-blown systems, rather than developing application logic.
At the same time, line-of-business use cases are getting insufficient attention from vendors as much of the focus remains on low-level infrastructure and operations, rather than solving problems and delivering value to businesses. Organizations need to be able to apply Hadoop to specific business challenges, but to do this today, they must invest significant time and resources at the infrastructure level. Hadoop is still very much in the early days, but what we ultimately need to unlock the potential of Big Data is for the ecosystem around Hadoop to mature in order to create a more usable platform for data applications. This is where the real opportunity lies and where Continuuity is investing our efforts—allowing developers to create the next generation of great applications, powered by data and operated at scale.
PBB: Why are business-intelligence (BI) adoption rates so low today?
Elad Israeli, founder and CPO, SiSense: On the one hand, there’s a lot of pressure on executives these days to make sure their organizations are data driven. On the other hand, BI adoption rates are relatively low. A lot of BI requires expensive hardware and complicated software. There are also often integration issues between database technologies and visualization tools. Legacy BI solutions are often pricey, difficult to use and require the help of IT or outside consultants to implement. And when it comes to mobile BI, adoption rates have been even worse because, quite frankly, mobile solutions often severely limited. There are some solutions that offer a pretty front-end, but the problem is not with the UI; it’s a back-end issue. What good is a glittery front end if the solution cannot handle a lot of simultaneous queries? At SiSense, we worked for years on the back-end in order to overcome all these problems.
If you want some guidelines about questions to ask to ensure you implement the most useful strategy for your organization’s particular needs, see this advice I shared in VentureBeat recently.”
PBB: I’ve heard that industry-standard switches are becoming more popular. Why do they matter to the broader market, and to growing companies in general?
JR Rivers, co-founder and CEO, Cumulus Networks: More and more technology vendors are moving towards a model that disaggregates hardware from software, removing vendor lock-in and all of the huge markups and inefficiencies that come with it. What we are now seeing is a much more sophisticated and streamlined approach in which customers are using industry- standard switches, then installing the software of their choice on top of them. This new operating model lets businesses scale their data centers at software speed and in the most efficient manner possible, allowing them to realize benefits on both the capex and opex side.
Rather than buying from an incumbent that will lock them in to only using their software and hardware components, businesses can now select the technology that best suits their needs from the vendors they choose. And what we are seeing is a number of networking technology vendors forming partnerships with other like-minded companies in the space, which ensures compatibility between the various components in the networking stack. Case in point would be our recent partnership with Dell. Ultimately, separating hardware from software on networks gives customers choice, making it simpler and more affordable to build high capacity networks much like the mega-scale datacenters similar to the ones Google and Amazon have.