Listen to the podcast. Find it on iTunes/iPod. Read a full transcript or download a copy. Sponsor:HP.
Cloud computing has sparked the imagination of business leaders, who see it as a powerful new way to be innovative and gain first-mover advantages -- with or without traditional IT's consent.
This now means that the center of gravity for IT services is shifting toward the enterprise’s boundaries – moving increasingly outside their firewalls. And so how can companies have it both ways -- exploit cloud's promise but also retain rigor and control that internal IT influences and governance enables?
In a special BriefingsDirect sponsored podcast created from a recent HP Expert Chat discussion on how to best pursue cloud computing models, I recently moderated an HP Expert Chat session with Singapore-based Glenn West, the Data Center and Converged Infrastructure Lead for Asia-Pacific and Japan in HP’s Technology Services Organization. [Disclosure: HP is a sponsor of BriefingsDirect podcasts.]
In our wide-ranging discussion, we deliver actionable recommendations for how IT can support cloud benefits and re-architect their data centers with low risk, and how IT can quickly take on the role of cloud benefits enabler.
Here are some excerpts:
Glenn West: The cloud is an exciting environment, and it's changing things in quite incredible ways. We're going to be focused now on how cloud is enabling data center transformation.
In the data center today, there are quite a few challenges, both from the external world, as well as from internal changes. In the external space, there are regulatory risks, natural disasters, legal challenges, and obviously technologies are changing.
Whether the IT department chooses to change or not, businesses are changing anyway. This is putting pressure on the data center. They must adapt and transform. Internally, greater agility and consolidation are needed. Green initiatives to save money and cost are putting great pressure on change.
So all of these things are causing the data center to converge, and this convergence is pushing the cloud.
What is a data center? In HP we have a very holistic approach. We'll start at the bottom from the facility point of view -- the location, the building, the mechanical and the electrical. Data center densities are growing quite rapidly and electrical costs are changing incredibly fast and rising. So the facility is very important in the operational cost of the data center.
If data centers change, the management and the efficient operation become even more important. Then, controlling, governing, and the organization have key parts. Without having the right organizational structure it's very difficult to manage your clouds.
In reality, cloud computing is an irresistible force. It's moving forward, and things are changing.
Scalable and elastic
So what does cloud mean? Cloud means going to a more service-driven model. It's scalable and it's elastic. Think about the public cloud space. How do you handle it when something is very, very popular? One day, it may have a hundred users and the next day it becomes the next hot thing for that instant in time. Then, the demand goes away.
If we use a traditional model, we can’t afford to have the infrastructure, but this pay-per-use is the foundation of cloud. We start looking at a service concept delivered and consumed over the Internet as needed.
The key word that keeps coming up is service, service orientation, the elasticity and the pay-per-use. Clouds ideally are multi-tenant. That can be within a company or outside a company.
Cloud is for automation, orchestration, automating control and a service catalog. All of a sudden, instead of calling somebody and saying, "I need this done," you have a portal. You say, "I want a SharePoint site," and boom. It’s created.
It's radically different than traditional IT. You move away from managing servers, and you manage services. In a data center, over the next couple of years, focus is going to be on private clouds. There will be public cloud providers for certain things, but the focus is going to be on the private side.
Private will slowly push into a hybrid and then slowly adds additional from the cloud services. The majority initially a private cloud will be infrastructure as a service.
The key drivers of this are agility and speed. When a business unit says they need it tomorrow, they're not joking. The agility that a private cloud provides solves a lot of opportunity in the business. It also solves the pressure of going into a public cloud supplier and getting it outside the IT framework.
Management and processing
The challenges over the next few years are management and the processing. How do we fund and charge back the whole business model concept? Then, building the cloud service interface, the service description. All of this is before the technology. Cloud is more than just the technology. It’s also about people and process.
Only a small portion will fit in the cloud today, but things are rapidly moving. We were talking about the future. Look at the current sprawl that's occurring. If IT doesn’t get in the front, this probably will get worse. But if the cloud is managed properly, then IT sprawl can be reduced, controlled, and slowly moved into a more standardized structure.
As we move into the cloud and talk about private cloud, service function of IT starts coming into reality, and this is referred to as hybrid delivery. Hybrid delivery is when you start looking at the different ways of providing services, whether they are outsourced, cloud private-based, or publicly-based.
You start looking at becoming a service broker, which is the point at which you say that for this particular service, it makes best sense to pay it here. Then you start looking to manage it and be able to fully optimize your services.
Going further out into 2015, 18 percent of all IT delivery will be public cloud, 28 percent will remain as private cloud, and the rest will be in-house or outsourced. You can see the rapid change going forward.
Gardner: What kind of applications do you think we are going to see? When you mention the service enablement, these different cloud models, I think people want to know what sorts of applications will be coming first in terms of applicability to these models?
West: If you're referring to public cloud, the first ones a lot of times are collaboration applications. Those were the first ones that moved into the public cloud space. Things like SharePoint, email, calendaring applications were the early adopter models.
Later we have seen CRM applications move. Slowly but surely, you're seeing more and more application types, especially when you start looking at infrastructure as a service (IaaS). It’s not so much the type of application, but the type of application load.
As you see, the traditional model is all about selling products, fixed costs, fixed assets. Everything is fixed. But when you start looking at a service model, it’s more pay-per-use. It’s flexibility, it’s the choice, but also a bit of uncertainty. In the traditional model you have controls, but when you start looking at the service model, it’s all about adaptability and change.
So there's a big gap here. On one side, we're all about things being fixed, and on the next side, we're moving to being cloud ready, to hybrid services, and hybrid service delivery. So how do we get across this great divide? We really need a bridge. We really need a way to move across this great divide and this big change.
The way we change this is through transformation. It's a journey. Cloud is not something that you can wake up one day and say, "We're going to have it executed instantly." You have to look at it as going through levels of maturity.
This maturity model starts at the bottom. Some organizations are already at the beginning of this journey. They've already started standardizing, or they may have started virtualizing, but it’s a process. You have to get to the point where you're looking at moving up. It’s not just about technology.
Obviously, you have to get to the point where you're consuming cloud services. If you look at the movement to cloud, you can look at it as pulling organizations into it. This is driven by the rapid adoption by the masters in cloud. There’s a great push from the business side. Business is gearing their customers to talk about cloud and cloud-based applications. So there is a pull there.
Also from the data center itself, there is a push. The IT sprawl, the difficulty in management, are all pushing towards cloud quite rapidly.
The question is, where are we now? Right now, a lot of companies are in this environment where they have started virtualizing. They've moved up a bit and they've started doing some optimization. So they're right at the edge of this.
But to move forward you need to look at changing more than just some of the technology. You also need to look at the people, the technology, and the process in order to bring organizational maturity to the point that it’s starting to be service enabled. Then, you're starting to leverage the agility of cloud.
If you are just simply virtualized, then guess what, you're not going to see the benefit that cloud offers. You need to increase in all of these areas.
Gardner: As we look at the continuum, how do organizations continue to cut costs while they're going about this transformation?
West: This journey is quite interesting. To a large degree, the cost optimization is built in. When you start the journey in the standardization process, you start reducing cost there. As you virtualize, you get another level of cost reduction. At each step, when you start going to a shared service model and a service orientation, you start connecting things to business. You start getting the IT concepts dealing with the business cost.
Moving up to the point of elasticity, things are further optimized. This whole process is about optimization, and when you start talking about optimization, you're talking about driving down the costs.
Currently, between the beginning of this journey and the end of this journey, we're reducing cost as we go. Each stage of this is another level of cost reduction.
We mentioned that the cloud isn't just about technology. Obviously, technology is part of it, but it's also about automation and self-service portals. The cloud is about speed. Imagine the old traditional process, you say, "Let me weigh the capital equipment required. Let me get that approved. Let me write the PO."
As part of the process, you also have to get into standardization. You have to get into service lifecycle. A cloud that never throws anything away is not an optimized cloud. Having a complete service lifecycle, from beginning to end, is important.
Usage and chargeback are key elements as well. Anything that's free always has a long queue. In IT, a cloud without a chargeback model will be a cloud that is over-utilized and running out of control. Having a way of allocating and charging back to the consuming parties, be it an internal customer or outside customer is very important.
Elements often forgotten in cloud are people and having a service orientation. If you look at a traditional IT organization, you have a storage manager and a network manager. If you look at cloud, you have service managers. The whole structure changes. It doesn't necessarily reduce roles or increase roles, but the roles are different. It's about relationship management, capacity management, and vendor management. These are different terms than traditional IT.
Opportunities for cloud
High performance computing, web services, database services, collaboration, high volume, frequently requested standardized and repeatable. That pretty well identifies those great opportunities for private cloud.
Once you start moving into public cloud, you need the ability to understand that things will scale a business, meaning that you need to look at the variability of cost. They need to be tied to the level of business.
Things like backup ability, interoperability and standards, and security are additional things that we need to look at as we move into public cloud services and the hybrid model.
Let's talk about the things that are pushable to hybrid cloud models. First, core activities that are essential to the business are not suitable to go to cloud. Those are best in a private cloud. But if you start looking at things that are not unique, immediate, but not a differentiator or are cost-driven, then those are ideal for public cloud.
Basically core activities are very, very good for private cloud and less core activities or that are cost-driven are more ideal for a public or hybrid cloud offering.
Lock-in and neutrality?
Gardner: Glenn, what prevents us from getting locked into one cloud and not being able to move out?
West: There are a significant number of cloud standards at every level, and HP does everything it can to remain part of those standards and to support those standards. The cloud industry is moving fast, and if you look at cloud, it's about openness. If you have a private cloud then you cannot have the ability to burst to public cloud.
Gartner actually did a study, and found that HP is one of the most open players in the industry, when it comes to cloud. A significant number of the public cloud suppliers actually use our equipment. We make a point of being totally open.
The cloud industry as a whole, because of the interoperability requirements, has to be inherently open.
There are a significant number of cloud standards at every level, and HP does everything it can to remain part of those standards and to support those standards.
Gardner: So it's not only important to pick the technologies, but it's very important to pick the partners when you start to get into these strategies?
West: That’s absolutely right. If the viewpoint of the company that you're getting your cloud from is to lock you in, then you're going to get locked in. But if the company is pushing hard to stay open, then you can see it, and there are plenty of materials available to show who is trying to do lock-in and who is trying to do open standards.
What do we need to think about here? Flexibility is obviously important. Interoperability -- and I think Dana nailed that one on the head -- being able to work across multiple standards is important. The cloud is about agility. Having a resource pool and workflow that can move around the resource pools on demand means that you have to have great interoperability.
Data privacy and compliance issues come into play, especially if we move from a private cloud into public cloud or hybrid offerings. Those things are important, especially on the compliance side, where the cloud supports data being anywhere.
Some requirements, depending on the industry, actually restrict the data movement. Skill-sets are important. Recovery and performance management, all of these things can be managed with the right automation and the right tools in cloud as well as the right people.
We've talked about moving forward and now we're getting into the full IT service broker concept. This is where we have the greatest flexibility. One of the things you said very well was about dynamic sourcing. We can look at the workflow and we can push and share these workflows internally and externally to multiple cloud providers and act as a service broker, optimizing as we go.
You should have this even from a corporate point of view. You could be a service provider where you take those services and you broker and manage those services across multiple delivery methodologies.
At this point, you have to get at an organization very good at doing service-level agreement (SLA) management. SLAs, when you are growing cost and managing workflows through this is very important. When we start talking about going across multiple clouds, advanced automation gets to be very important as well.
As we start looking at the future data center, it is very business-driven. You have multiple ways of sourcing your IT services. So you have both, the physical as well as the virtual services and you have the appropriate mix. It’s changing practically on a daily basis, as business needs demand.
Let's talk about these physical side and the changes in the data center. One of the things that looks quite interesting, if we look at resiliency, is that a lot of data centers are looking at moving further up the resiliency levels, and each level of this has significantly increased cost, practically exponential cost increase.
Once you implement cloud within your data center, you can get a lot more flexibility all of a sudden, because instead of building a single Tier 4 data center, using the efficiency of cloud, you could potentially build Tier 2 data center and have greater resiliency and greater agility.
The big change is in the way data center physical infrastructure is done, but the thing that's changing quite rapidly is density. If you look at it in a traditional data center, infrastructure is reasonably low to moderate density.
When you start looking at cloud enabled data center, high density is the norm. Greater efficiency, power, space, and cooling are all typical of cloud-enabled data centers. This true IT resource is where anything can run anywhere, and it becomes quite different.
The density change is radically different. The power per rack and cooling all change. The next thing with power is that even if you start looking at a traditional data centers, things such as structure, cabling, and power have to have flexibility and have to have the ability to change.
Orchestration also becomes important. If you start looking at a cloud-enabled data center, everything needs to scale. All the cost factors should scale with the amount of business.
Standardization and efficiency
The standardization level also changes as well. Standardizing configurations allows rapid redeployment of equipment. Finally, it’s efficiency -- this dynamic power and cooling during the work loads.
These are pretty radical changes from traditional data centers. Data centers are evolving. If you look at traditional data centers, they were quite monolithic -- one large floor, one large building, that’s pretty well it.
Slowly moving up to multi-tiered data centers, followed by flexible data centers that share resource utilization, and everything can change.
Most organizations when you start looking at the different areas, categories, and types of culture, the technology is there. If you looked at a company today, they will have different levels of maturity. This maturity modeling is a scorecard or a grade card that lets you understand where you are compared to the industry. The thing is that if you look at this example, different areas have different levels of maturity.
The problem is for cloud is that we need to look at something a little different. We need to get an even playing field across all of the areas so that the organizational maturity, the culture, the staff, the best practices are even in the level of maturity for cloud to work.
If you bought the best technology, but you didn’t upgrade your governance or the culture, and you didn’t implement the best practices, it won’t work. The best infrastructure without proper service portfolio management, for example, just isn’t going to work. For cloud to work properly, you must actually look at increasing maturity across all areas of your data center, both the people and the process.
Some of the criteria for cloud include technology, consolidation, virtualization, management, governance, the people, process and services, and the service level. Managing the service level can often reduce your cost quite significantly in cloud.
In the process, adopting ITIL and looking at process automation and process management. These organizational structure and roles are quite different in cloud.
Think services. Understand what you have. Decide on what your core and your content are. What is the foundation of your business and what is something that we could start considering moving into public cloud sector?
Get your business units and your businesses on your side. Standardize and look at the automation of processes and explore the infrastructure conversion. Then look at introducing your portal and making sure you have charge back. Start with non-critical or green-field areas. Green field are your new activities. Then, slowly move into a hybrid approach.
Evolve, optimize, benchmark, cycle through -- and optimize further. HP has been doing this for a while. We did a very large transformation ourselves and out of that journey, we've created a huge amount of intellectual property. We have a Transformation Experience Workshop that helps organization understand what changes are needed. We can get people talking and get them moving, creating a vision together.
We have data-center services for looking at optimization, the physical change of data centers. And then we have comprehensive data center transformation services and road map. So get some action going. Let's start at doing the transformation.
A great way to do this is do it a one-day Cloud Transformation Experience Workshop. This is done in panels with key decision makers and it allows you to start building a foundation of how to go through this journey in transformation.
Listen to the podcast. Find it on iTunes/iPod. Read a full transcript or download a copy. Sponsor: HP.
You may also be interested in: