VMware outlines management suites strategy

Virtualising computer infrastructure has many advantages, but easily managing some aspects of increasingly complex environments has not been one of them. Last month, virtualisation technology vendor, VMware announced three management suites designed to address this, at VMworld Europe, in Copenhagen.

Between them, the suites drive better infrastructure efficiency, right some of the management wrongs created by virtualisation itself, and help IT teams with different agendas collaborate better. The ability to 'see into' virtual machines (VMs) has also been added.

Because these management suites are highly complex themselves, VMware's CEO and CTO spoke about significant resources allocated to quality assurance, aiming to relieve customers from having to deal with teething problems themselves.

In his keynote address, VMware CEO Paul Maritz explained his take on the need for better virtualisation management, to get maximum operational efficiency at the infrastructure level.

“We are working very hard to assemble a complete suite of products that will allow you to realise complete efficiency in the private cloud. That starts with virtualisation, the layer that provides the resource pooling and the automatic scheduling of resources in the pool. It adds in the functionality you need for cross-data centre disaster recovery.

“Also, [the product suite] adds new ways of doing management; because when you move to a bigger pool of resources, it means you will have more and more applications eating out of the same pool. You're no longer managing on a silo-by-silo basis. So if you're the custodian of that big pool of underlying infrastructure, you have a different set of needs.

“So what we've meant to implement in the core of our vCenter Operations Management Suite is an example of a new, real-time analytic application. We applied modern analytics statistical analysis to that pile of information coming out of the pool. With that, we build a statistical model of your infrastructure, and then monitor it. We compare the infrastructure in real-time against the model we built. Every now and then, we tell you that it's gone outside the bounds of what's normal, and here are the things we think are outside the bounds of normal.

“It's a fundamentally different approach to management.”

Solving teething problems with complex management solutions to complex virtualisation problems will take some effort; however, Maritz promised a strong focus on doing that.

“We have all the members of the [management] suite in existence now. Our goal is to move these technologies forward as a suite. We want them to go out on the same schedule. We will take the time to make sure they are properly tested, so that you don't have to spend time doing that.

“Businesses increasingly view infrastructure as a means to an end. Infrastructure does not differentiate one business from another. Over time, it just needs to work, and fade away into the woodwork. To do this, we need an integrated approach.”

Maritz was asked whether VMware is really able to maintain high quality in its core offering and integration, given that the company buys numerous technologies and companies to extend its abilities.

“It's hard for any company to strike the right balance between doing a few things very well, while reacting to the realities of the world, which won't stop changing around you,” replied Maritz.

“In general, we're trying to do that. At each layer of our strategy, we are trying to pivot towards a more integrated approach. For instance, our announcements ... are about pulling things into a true [management] suite.

"So if you look at what we're doing with these individual products that we've either acquired or developed over the years, there is a very strong priority now to pull those together into a more integrated offering. [We also want] to make sure that we release them on a common schedule, take the rough edges off between them, so our customers don't have to be in the business of trying to individually test components. That is very much on our minds now, without compromising the quality of our vSphere product, the foundation on which everything is built.”

Continued Maritz: “One-third of the world's server applications are running on vSphere today. That is a big responsibility, and we have to take that into account as we go forward. We do, and it is a huge expense for us. For every engineer we employ developing code, there is another two people just providing quality assurance and checking functions around our product.”

In his keynote address, VMware CTO Steve Herrod outlined more reasons to improve management of virtualised environments: management problems caused by virtualisation and recent changes in the IT environment.

"This is one of the biggest management launches we have ever done. We want to focus on all the areas you need to be ready for the cloud,” said Herrod.

“There is a requirement to do a lot more in the management space, because, candidly, we've caused a lot of problems for management tools over the last several years. It used to be a nice, clean world where I had my entire data centre, that I owned, and I had one application, mapped to one server. But virtualisation came along and severed the tie that existed between the two.”

“We separated the application from the infrastructure, and I can't tell you, starting six or seven years ago, how many tools and processes broke just from that. Not only from knowing that they're not tied, but even when they begin to get consolidated. This core notion of an application-tied infrastructure has been built into a lot of the original ITIL processes, and has really kept us from moving as quickly as we need to do in this new [virtualisation] world.

“The second thing we have done is that we moved to a world with pooled infrastructure. This is a more dynamic world, where the applications are changing the mappings between where they run and how they run, more quickly than ever before.

“Thirdly, the users are more dynamic than ever; there are so many more virtual machines out there, and with cloud computing, we're allowing people to self-serve virtual machines as quickly as possible.”

Given all these factors, new virtualisation management is required, continued Herrod.

“We need management tools that can keep up with massive amounts of change, make sense of it, and provide you with the right information. Add to that, you have more choice than ever before: Most CIOs are not just looking at their own infrastructure, but also see that a public cloud could be used for some applications.

"On the other side, we see software as a service (SaaS) as a great option for many companies. They can choose to build an app themselves and deploy it locally, or to grab one from outside their own infrastructure. This choice is liberating, but is also a real challenge. How can a CIO know which one to move towards? So for all these reasons, and several others, we see it's time for a new approach to management.

“The majority of our engineers are working on automation and management tools at this point. The core of our approach is our belief that the platform should shoulder more of the burden for making sure that you can automate things and enforce guarantees. Smart infrastructure knows about availability, performance and security, and can make sure it is satisfied.

“But there are two other things which we think are interesting, which can make management step up a notch.

“Firstly, to converge what are silos today. You see, teams are supposed to work with each other, but they have different tools, different processes, and when the time comes to communicate, they're throwing job tickets back and forth in a way that's not particularly productive.

"Secondly, it's about dealing with the data deluge in large, fast-moving infrastructures. We need to be able to filter that data and provide just the right amount of information, depending on who's looking at it.

Sometimes infrastructure and operations teams throw job tickets 'over the wall' at each other, instead of collaborating to solve a problem. This is one of the areas targeted by the vCenter Operations Management Suite. The suite integrates performance, capacity and configuration management to automate operations.

"Most companies have at least two teams to manage infrastructure: an infrastructure team and an operations team. The teams have different concerns, and in many cases, they're billed differently. An infrastructure person often worries about physical entities, like buying servers, storage and networking. But an operations team is worried about keeping the devices running, and satisfying the demands from the clients. This team deals with end-users, [job] tickets and service level agreements (SLAs). There is almost a wall that forms between the two teams."

The new infrastructure management suite gets over that wall.

"vCenter operations management 5.0 builds on vCenter ops we launched in April, and adds in several capabilities from other tools that were shipped independently before, such as CapacityIQ and Configuration Manager. It also adds a killer new capability to really understand more about the applications running on top of your infrastructure. This is called Infrastructure Navigator.

Previously, a virtual machine was a 'black box', and it was not possible to see inside it using management tools. That has changed.

Navigator has the ability to understand the applications and services that are running inside a virtual machine without making any changes to the virtual machine, to the operating system, or to the application. Navigator is agentless and automatic, and can be used to filter information. Navigator also helps in understanding the relationships between VMs. As an example, one can see a SharePoint VM, a SQL server database and a second SharePoint VM on one screen. When making changes to the first SharePoint VM, one probably needs to make the same changes to the second VM.

"[vCenter Operations] is about giving a converged view, to both the infrastructure and operations teams, on the key things they care about. It brings them together in a nice way, so they can talk in a common language and a common view about performance, capacity and configuration. What's interesting about this is that it shows you what filtering means."

The first screen one logs into shows a dashboard, allowing the user to drill down deeply to where the problem is, but doesn't compromise the ability to troubleshoot and really understand where the core issue is. Herrod expects a change in managing infrastructure, using this tool; or a new field even.

"We're calling this cloud operations, around this efficient, converged way of managing. With vCenter Operations Management, we are converging the fields of infrastructure and operations, to make sure customers can handle performance, capacity and configuration changes."

The second suite Herrod spoke about enables users to provision and monitor applications in the cloud. It also aims to bring two different sets of people together: application developers and those who put the apps into production. These two teams usually have different priorities.

"A developer is typically worried about development cycles, saying, 'I got to get the latest feature out'. Or he may have deployment challenges, looking for the infrastructure so he can build the feature. Bugs are another worry - the application needs to run correctly, but the performance, when deployed, matters, too.

"The operations team typically cares about the end-user experience, because they get trouble-shooting requests all the time. 'This app is slow', is the most common one. Another concern is change control - being accountable for how all of this works.

"So what we're trying to do here is work on the emerging field of 'dev ops' (development operations), which tries to bring the two teams together in a better way, and get over the brick wall that exists between them.

"So ... we also announce the vFabric Application Management Suite. This is the first time we're going into the application management space in a way that will converge teams, give you the right information, and make sure that you can efficiently monitor and deploy applications.”

vFabric Application Management Suite allows users to do this across private and public clouds.

Advancing peace between infrastructure and operations teams, and between applications and operations teams, helps, but the CIO faces another real challenge, and that is what the third suite addresses, said Herrod.

"CIOs ask 'how do I deal with all this complexity, and how do I make the right choice? CIOs have to deal with risk in complying with regulations. It is also around speed - every business wants to move quicker than they can so far. And every CIO in the world is pressured to save money. So there are lots of pressures, but also a lot of choices, more choices than ever before.

"Do I run this in my own data centre? Shall I use the public cloud? Or go software as a service for this particular application? The final big announcement is on how we can help CIOs to choose and measure different approaches to satisfy the needs of their business.

"We're also launching the VMware IT Business Management Suite. This is the newest suite from us, and also in a very new area that we're seeing quite a bit of traction for. It is aimed at CIOs and line-of-business owners, and is really about visibility and control, all the risk and cost facing them.

In the suite, a CIO can see a simple bar graph depicting how the IT organisation is doing with its SLAs against each of the lines of business that it is servicing. One can also see what the company spends to get that service, and a roll-up of year-to-date IT spend by service, by infrastructure. All the information comes from thousands of data points, which is then filtered, to give senior management the information they need.

Continued Herrod: "IT Business Management is really around letting the CIO have a chance to understand costs and risk in a world where there is a lot more choices than before. What we're trying to do is pull all these teams together, as a step closer towards IT as a service.

"One of our engineers says 'it should work and it should work well'. This is something that is happening in all these products. Almost everything is shipping today, or is in use. It's about the services in the cloud, but it's also about the people.”

The suite is intended to act as a broker of IT services.

Concludes Herrod: "We are currently in one of the biggest transitions happening, and it's a privilege to be part of it. These changes are really forcing us to think of services and people, rather than individual servers and desktops.”