Integration gets back on track with the vertical stack
We don't need no vendor lock-in
What is the role of integrated stacks, in which compute, networking, storage and management are vertically integrated? And how do you avoid vendor lock-in?
Let's get the discussion started with three expert submissions, one from an analyst, one from an IT practitioner and one from a vendor. And tell us what you think in the comments below.
Andrew Buss, service director, Freeform Dynamics
The delivery of IT applications or services requires a number of elements to come together.
While some solutions may need specialised kit to work, we can usually distill the required elements down to a common core of compute, data storage and communications. Mostly, we tend to think of these as physical boxes: servers, disks and networking.
In the past, these were often highly integrated like the IBM mainframe, which was a complete system end to end providing processing, storage, networking and dedicated client terminals and peripherals.
This brought the benefits of reliability, scalability and manageability but also disadvantages such as the high cost of purchase and a certain inability to adapt to new requirements.
With the arrival of minicomputers and then the workgroup server, IT infrastructure opened up. Cheaper, PC-based servers boosted the trend towards scaling out rather than up.
Difficulties with managing a lot of locally attached storage in these servers led to dedicated storage servers and the rapid evolution of the storage area network.
Alone in a crowd
Meanwhile, connectivity settled on Ethernet and TCP/IP, which moved from small, shared hub-based networks to large-scale switched networks with highly configurable management.
The upshot of all this has been a proliferation of devices and solutions at each level of the IT infrastructure stack, providing a vast choice of vendor, performance, features and price. It has been great for flexibility but our research shows a tendency towards fragmentation and isolation between the servers, storage and networking, which hinders the ability to adapt to change and makes consistent and coordinated management that much harder.
This is where the concept of the vertical stack is making a resurgence. Like the mainframe, compute, storage and networking are bought together and closely integrated, enabling a more seamless approach to management and configuration and boosting agility.
The common view is that of a single-vendor solution, such as the HP Cloud Matrix, IBM PureSystems or Dell vStart. But it also applies to packaged solutions encompassing multiple vendors such as the VCE vBlock or the Cisco/NetApp FlexPod.
These vertically integrated stacks are engineered to work together, but should not require exclusivity. The key approach is that if the solution meets requirements the standard hardware can be used, but if there are other needs then it should be possible to plumb in third-party kit.
To avoid lock-in, customers should look for standardised and open interfaces at each layer of the stack, with the ability to swap out kit at any level and replace it with an alternative, or to plug in third-party equipment.
That is not to say that no work is required: some integration will almost certainly be needed. But by insisting on common interfaces and standard integration packs, it is kept to a minimum.
Trevor Pott, Sysadmin and Reg columnist
A vertically integrated solution is one in which a single vendor owns a unified stack of products. These are extremely difficult and costly for vendors to assemble because the modern data centre has many components. So why do they keep trying?
You can really trace it all to IBM. The vendor’s strategy of vertical integration has seen it through highs and lows and it keeps turning a profit. IBM owns the CPU (Power) the operating system (AIX), and the manufacturing to make the boards, chassis and so forth to create individual server nodes.
IBM provide the racks, the power, the cooling, the databases, the middleware, the networking – all of it. Right through to acting as a managed service provider for everything from a small x86 server to its lucrative mainframes.
The entire stack is designed in such a way that the overall margins can cope with a downturn in any one segment. Power going out of style? That's okay, the company will bring in x86.
x86 going out of style? No problem, IBM will have ARM all wrapped up. Services a little short? Turn the knobs on the hardware to compensate. It is economics designed by engineers. Instead of relying on a single product or product type, IBM has Raid 61ed its income.
It is no wonder that everyone wants to be like IBM. That's the vendor side of the equation. What role do vertically integrated stacks play in the enterprise? It depends entirely on who you ask.
Single-vendor stacks are a wonderful concept for the kinds of empty suits that never actually have to implement the stuff or deal with the cold reality of vendor relations.
Human resources people like single-vendor stacks because that way they know exactly what to look for. Relying only on certifications will probably not get you the best people for the job but no HR drone is going to get fired for hiring Cisco-certified meat to look after Cisco-provisioned iron.
Middle managers like single stacks because they represent one company to negotiate with, one training path and one conference set for their employees.
Everything about single-vendor stacks is aimed at reducing costs and effort. Although single-sourcing drives up the cost of buying the gear, for a large business this is a good thing. Capital expenses can be written off whereas humans are fickle and ultimately a liability.
Lock-in rears its head from time to time but this is not a massive concern on the data centre front. There are a number of vendors playing the integration game: IBM, Dell, HP, Cisco, Oracle, Supermicro, VMware/EMC and more.
Each has its own strengths and weaknesses along the stack, but when refresh time comes around the thing preventing you from moving your whole stack from vendor A to vendor B is never the hardware. It is who owns your data.
Data lock-in is why Oracle bought Sun and why losing the appeal over the termination of its Itanium database software version was such a big deal.
Once you have them trapped, turn the knobs and buy some more yachts
The idea was simple: Oracle database plus Oracle Middleware on Oracle-controlled Java running on Oracle-owned Solaris on Oracle hardware supported by Oracle services.
Get enough enterprises locked in and then simply stop making versions for anyone else. Once you have them trapped, turn the knobs as hard as you can and buy some more yachts.
What happened instead? Hadoop and NoSQL. Oh, Hadoop is pretty crap and NoSQL can't really compete with Oracle for the important stuff, but they are learning. This is the risk of vertical integration for vendors and for their customers. Oracle made a play and lost.
In the next few years, those who invested in the Oracle ecosystem (data hostages) will start to see their competition fly past. The competition will be using a more agile, more open offering whose costs are much lower. It will be two decades before that market consolidates and those not locked in will have a huge advantage.
This is exactly the sort of innovation that is the true hedge against lock-in. Enterprises are willing to put up with a little gouging every now and again but there are limits.
Back the world's largest corporations into a corner and they will respond by dumping resources and funding alternatives on a scale that would make many nations weep.
Wayne Meyer, Windows server product manager, Microsoft UK
Referring to an integrated stack means thinking about how we can build out a platform that can be simply managed and put together seamlessly.
The role of IT is complex, not just because of the products but because of the need to align to business process and requirements. It is imperative that as we build solutions to meet those requirements the underlying technology fits together and can grow and adapt along with the business.
An integrated technology platform simplifies the way information is stored and used
The business world is changing rapidly and customers need to address opportunities before they fall by the wayside. Having an integrated technology platform, from user to data centre, from communications to data simplifies the way information is stored, shared and used.
Integrated stack does not equal single vendor. Clearly there is large value in building a single-vendor solution but customers can also gain value by building an integrated stack from strong industry partnerships.
Standards and joint development programmes are in place to give customers a choice of hardware and software.
Let’s take data centre management as an example. System Center 2012 enables customers to manage multiple hypervisors – Hyper-V, ESX and Citrix Xen – three different hypervisors managed through a single toolset. Within that toolset they can also manage storage, networking and operating systems – Windows or Linux.
Integrated stacks can reduce the dependency on vendors and avoid lock-in. As shown above, System Center gives customers the opportunity to choose the hypervisor that best fits their needs, either from a cost or capability perspective.
Lock-in, however, is not just about technology and implementation, it is also about licensing, skills and the ability to expand services seamlessly. The platform I choose today should enable me to build onto that platform, acquire secondary products to provide new capabilities and accommodate future growth simply and cost effectively.
Customers need a technology partner with a clearly stated vision and a core set of technologies that are designed to work together, rather than being pieced together by acquisition, and which provides the best economics. ®