Posted by Crystal Nichols on Thu, Feb 02, 2012 @ 08:56 AM
In IT there is a critical relationship between size and simplicity. That is, the larger or more important a service, the simpler it must to be to meet the reliability and availability needs of the business.
Too much complexity and the application runs the risk of failure – usually at just the wrong time.
To combat this, the infrastructure must be constantly monitored, nurtured and nursed just to meet the minimum needs of the business. This, in short, drives up costs and slows down progress.
We have come a long way over the last decade in IT. Virtualization via hypervisors has helped tremendously.
The abstraction of the application from the underlying hardware has enabled a first leg of improvement in both the cost and agility of IT, but hypervisors do not help bring simplicity, resiliency or agility to the physical infrastructure.
Bringing that same sort of abstraction to the compute and networking layers is the next obvious step in the drive to modernize the data center. This is the idea behind fabric based computing.
Through abstraction, fabrics (aka converged infrastructure) simplify the physical infrastructure thus increasing the flexibility and resiliency of the datacenter. This is why compute fabrics are a top CIO priority as we enter 2012.
For those of you familiar with blade technology or converged infrastructure, you might have already heard of Egenera Inc. – Unitiv’s newest vendor partner.
Egenera was a pioneer in blade systems and converged infrastructure fabrics over ten years ago.
They launched their “BladeFrame” product, which was a combination of the company’s blade server technology married with their powerful fabric “PAN Manager” software solution.
For years, the company performed extremely well selling their own platform, but recently, urged by its customers, Egenera made the strategic choice to port their fabric management software to other blade platforms.
Launched in May of 2011, Egenera’s PAN Manager for HP is now available for Unitiv’s c7000 blade customers.
By running PAN Manager on open systems –such as HP BladeSystem – customers get the best of both worlds – simplicity of management on existing and industry leading blade systems.
The open systems approach was important to us at Unitiv as we have heard your concerns about potential vendor lock-in with alternative fabric offerings - such as Cisco’s UCS.
But, unlike Cisco UCS, Egenera enables stateless compute fabrics without requiring customers to rip & replace their existing hardware investments.
And for us at Unitiv, that was critically important in choosing Egenera as our partner.
Learn more about Egenera or to arrange a demo, contact Unitiv today.
Posted by Crystal Nichols on Mon, Jan 30, 2012 @ 10:03 AM
It’s just been an accepted fact of the life of an IT professional (and of a data center manager in particular) that sometimes you have to go into the data center to fix things. A single phone call at 3 AM means you’re tooling down the road half-awake, hoping to find an open coffee shop on the way to the data center so you can approach the problem with at least some semblance of logic and attention.
Remote tools have been somewhat lackluster, and only able to handle certain types of problems. While you can probably connect remotely to restart a given virtual machine or a server, there are countless data center components you can’t access from home.
In particular, some data center components just seem to get left in the dust when it comes to remote management. These include:
• UPSes
• AC Units
• Low-level network equipment
In some environments, even servers, routers, and switches go unmonitored.
Convenience and cost
Part of the problem isn’t that these devices aren’t capable of being managed remotely; the problem is that the remote management tools are either not installed or not configured.
This makes sense, to some degree. When you’re pricing out equipment, it’s easy to drop that “remote maintenance” line item. After all, remote maintenance doesn’t affect the equipment’s core functionality. And, while remote management can reduce downtime by making it so that you don’t have to drive into the data center, you know that it won’t make a difference during staffed hours.
For other equipment, remote management capabilities exist but simply aren’t turned on. Configuring remote management is an afterthought, one which rarely produces action.
Developing a remote management strategy
Remote management can improve your uptime, and it can even improve staff morale. To make the most of remote management, however, you need to develop a remote management strategy.
This involves several steps:
• Identify remote management candidates. Not every piece of equipment can or should be managed remotely.
• Document remote management functionality and procedures. Make sure your data center staff know what can be accessed remotely, and how.
• Implement the appropriate remote management solutions. When necessary, acquire the additional hardware required.
Posted by Crystal Nichols on Mon, Jan 23, 2012 @ 10:08 AM
More and more, IT professionals are recognizing that tech projects really aren’t that different from other projects. They all involve managing people and processes, identifying and completing tasks, meeting measurable objectives, and doing so in specific timelines and with limited budgets. That said, there are aspects to tech project management that make the project management process unique and, in some ways, more challenging.
Understand the purpose of Tech Projects
Effective tech project management starts with an understanding of why exactly the project is being done in the first place. The particular application being implemented is never an end in itself. It’s easy to lose sight of that fact.
The purpose of tech projects is – and should always be – to support and improve business objectives.
Notice that we’re not just out there to satisfy business needs. By implementing tech projects, we become agents of business change and improvement.
Correctly define success
Give that the objective of a given tech project is never the technology itself, we need to be able to redefine what success means. Success doesn’t mean deployment or completion of the project. Success means true improvement of business.
Now, most IT project managers aren’t business change experts. In fact, we’re more likely to be wired toward the status quo. We’re told that we need to make sure our solutions guarantee continuity and uptime, both of which are more maintenance focused than improvement focused.
Arguing that a project met specs so it was successful – even if the business objectives aren’t complete or improved – is arguing for failure.
Managing implementation
Implementation shouldn’t be something that happens after a project; it should be a project in its own right, or it should be a milestone in the overall project. In the case of implementation, project ownership shifts from the project manager to the operational manager.
By treating implementation like a project, you make sure everyone involved knows what is supposed to happen. It puts the tech where it needs to be – in the hands of the users. Success for implementation means measuring business objectives and making sure they’re complete. Only then can the tech project be deemed a success.
Posted by Crystal Nichols on Fri, Jan 20, 2012 @ 08:23 AM
According to recent studies, more and more companies are relying on colocation services in order to handle the capacity issues of their own data centers. Rather than building brand new facilities (which come with a significant capital price tag as well as plenty of operational costs) they are outsourcing their data center capacity needs.
The data is stark: more than one third of large companies expect that they will have at least one data center run out of capacity within the next year and a half. That means they need to boost capacity.
There are, of course, a number of ways to boost capacity:
• Build a new data center. The capital expense involved in building a new data center is often prohibitive for many organizations, to say nothing of the possibility that the organization will run out of capacity before the new data center is online.
• Consolidate and upgrade. Server virtualization and data center consolidation can buy you some time. However, most organizations have done just about all of the consolidation they can at this point. There is still the potential for upgrading cooling or power capacity, but these are bandages on a bleeding wound.
• Outsorce. This is the option that many companies are choosing today. It makes sense from a capital expense perspective, and outsourced resources can be brought online much quicker than the other options.
Colocation vs. Cloud
There are two primary ways that data centers are outsourcing their capacity today: colocated services and cloud solutions. About a third of companies intend to lease some colocation space, while about one in five intend to move some applications to the cloud.
Colocation gives companies a greater degree of direct control over their data center services. However, it also requires more from a capacity and systems management perspective. Cloud solutions, on the other hand, offer scalability that is often unmatched. They come, of course, at a higher price.
Ultimately, the problems of data explosion and data center sprawl aren’t going away. It’s up to data center managers to identify where the potential exists for outsourcing, and then determine what kind of outsourcing makes the most sense for their business.
Posted by Crystal Nichols on Wed, Jan 18, 2012 @ 07:53 AM
Consumer technology solutions have, for the most part, maintained a healthy distance from enterprise IT operations. Sure, executives in the mid-2000s all needed help getting their Blackberry synced with your Outlook server, and company laptops have been a relatively standard thing for longer than that. By and large, however, most consumer solutions haven’t created too many issues for IT. The release of iCloud and iOS5, however, might prove to be a different story.
Starting with security
One of the biggest concerns enterprises are going to have with the use of iCloud for business purposes is security. Meeting compliance requirement can get a little bit more challenging, given this new technology. That being said, there are a number of important features in iOS 5 that should alleviate some of the concern.
There are three major areas of security to think about with iOS 5 and iCloud:
1. Location-based services
The new Find My Friends app offers some wonderful potential for things like employee monitoring or accountability. It also creates the opportunity for theft. Anyone on the user’s friend list can know where the user is. On top of that, the feature could be used to monitor a user covertly during times when they’re not working, making it a privacy risk.
2. iCloud
iCloud is designed to be a personal, rather than professional, data syncing product. The ability to have access to important files anytime, anywhere, is also tempting for users at work. This means that confidential corporate data may wind up being synced to a user’s iPhone or iPad, and then subsequently syncing to their home computer. Even after that person leaves the company, that data may then be available to the person. This, of course, says nothing about potential phishing scams and other potential security leaks from home.
3. The problem with Siri
Siri uses voice recognition to enter commands, read text messages, and more. If a user asks Siri to read business emails, for example, a nearby person might overhear something confidential. This can be managed through user education, but still remains something of a danger.
User education, along with a variety of security protocols on enterprise-owned devices and systems, should be able to mitigate most of these risks.
Posted by Crystal Nichols on Wed, Jan 11, 2012 @ 08:13 AM
HP has long been a proponent of virtualization solutions, and one recent move by the company bears out that commitment. HPs new VirtualSystem for Microsoft ships with HP servers using Microsoft Hyper-V virtualization software already preconfigured. This release promises to help companies deploy their virtualized server workloads much more quickly.
Best of all, the servers and their workloads can be managed using both Insight Control from HP as well as Microsoft Service Centers.
VirtualSystem for Microsoft provides customers with a direct path to private, as well as public, cloud solutions. It’s built on the same proven architecture that HP uses in its CloudSystem. This architecture combines three key areas – storage, servers, and networking – to help reduce latency, scale capacity, and effectively deploy cloud solutions.
While virtualization of course isn’t ideal for every type of cloud solution, there are many applications that will perform extraordinarily in this manner, and HP’s system is aimed squarely at those applications.
Scaling up, scaling down
When you talk about virtualization and cloud computing solutions, most of what you hear about scalability has to do with increasing capacity. Indeed, virtualization is key in offering more resources to an organization in a faster way.
But, what about those organizations who need to scale up only temporarily? It just doesn’t make sense to make huge investments in infrastructure that will cost more than what the temporary need will justify.
VirtualSystem for Microsoft gives you the kind of scalable flexibility you need for those high-capacity projects, but does so in such a way that coming back down to normal operations isn’t a problem.
Here are some quick details on VirtualSystem for Microsoft:
• Release is scheduled for later this year.
• Pricing will be around $175,500.
• Two models will be available. The VS1 can host around 750 virtual machines, while the VS2 can host around 2,500 virtual machines.
• Various configurations can contain up to 12 HP ProLiant servers, many modules for network and storage connectivity, and up to 84TB of SAN storage via the P4800 LeftHand SAN.
If you need to add to your virtualization capacity or want to implement a virtualization strategy, it’s worth taking a look at HP’s VirtualSystem for Microsoft.
Posted by Crystal Nichols on Mon, Jan 02, 2012 @ 08:03 AM
Discarded electronic equipment is becoming a major concern for our age. Electronic waste creates a new factor in the environmental problem. When electronics are discarded, there are at least two major concerns: the missed opportunity to reuse or recycle, and the release of various toxic materials into the environment.
Fortunately, there are some things that IT can do in order to reduce its e-waste footprint:
1. Start with e-waste policies. Curbing e-waste is something that should be modeled from the top down. CIOs are responsible for lifecycle management, and making sure that e-waste is a part of that management process only makes sense.
2. Consider it during the purchasing process. When a new IT purchase is made, take into consideration how it will be disposed of when the time comes. Identify those products with hazardous materials, and figure out which products have components that can be reused.
3. Reduce consumption. While most IT departments are already stretching things a bit when it comes to longevity, others are more cavalier. Identify ways to make do with the existing infrastructure, as long as it doesn’t impact efficiency.
4. Reuse. Peripherals, power cables, hard drives, and more all have the potential for reuse. Be careful here, of course; for example, older hard drives are likely to use more power than newer ones, and you don’t want to simply reuse a hard drive and then find that it’s actually using more energy.
5. Recycle. Many vendors have a recycling policy when it comes to old equipment. Old equipment is bought back and then redeployed or donated to those in need. In a worst case scenario, the precious metals in your IT equipment should have some value to local recycling firms, so contact them to determine whether or not they can help dispose of old systems in a way that’s environmentally friendly.
6. Participate in the community. Connect with other businesses outside of the supply chain and encourage curbing e-waste. Use your influence to make an impact on other people in your community, including those in your own industry.
Posted by Crystal Nichols on Fri, Dec 30, 2011 @ 07:59 AM
By now, many companies have embraced virtualization as a way to control IT capital costs, increase efficiency, and make better use of IT resources. Still, some haven’t made that leap, and others have only dabbled. If you find yourself feeling behind the times when it comes to virtualization, here are some tips to get you started on the right road:
1. Start small. IT staff can start very small – even on the desktop. For example, setting up a virtual desktop on your local system to test application compatibility issues with VirtualBox or a similar technology can give you a feel for how virtualization can work.
2. Set up a lab. Retired servers make great practice specimens for virtualization. The key is to make sure you pump them full of RAM and give them gigabit network cards.
3. Implement shared storage. Shared storage is necessary for virtualization, but it’s useful outside that context, too. Shared storage will give you a peek into this important aspect of virtualization.
4. Profile your existing infrastructure. If you want to virtualize, you need to start by knowing what exactly you have in production. Identify what can be virtualized and what cannot. There are a number of capacity-planning tools out there that can help you through this process.
5. Carefully choose what to move first. Most organizations will benefit from a step-by-step move to virtualized environments. Start small, and with systems that aren’t business-critical. This gives you plenty of opportunity to make mistakes, before it matters the most.
6. Monitory the transition process. Once you start to virtualize, make sure you’re watching performance on several levels, including storage, physical host, and virtual server. Watch those load leveling functions of your servers to make sure they’re working right. Make sure those resource forecasts you have are accurate, and if not find ways to adjust before they become a problem.
7. Calculate and tout the benefits. There are a number of real-world numbers that you can use to show why virtualization was a good idea. From utilization to TCO calculations, use these numbers to demonstrate the benefit of virtualization and to make the case for moving forward with virtualization strategies.
Posted by Crystal Nichols on Wed, Dec 28, 2011 @ 07:58 AM
Virtualization has made significant inroads in certain markets. For example, you’d be hard-pressed to find an enterprise or data center that, today, is not using virtualized server technologies. Indeed, virtualization has become less of an option and more of a necessity for companies that want to be able to continue to offer high levels of service, more applications, and do it for the same or even less money than they did before.
One area virtualization hasn’t taken off – at least for midsize and small companies – is the area of desktop virtualization. We were told that 2010 would be the year that virtual desktops would reign supreme. When that failed to materialize, experts told us that the release of Windows 7 in 2010 would mean that desktop virtualization would take over in 2011. Here we are in late 2011, and that hasn’t happened either.
Is it any wonder that many are prediction 2012 to be the year of desktop virtualization?
Higher priorities than desktop virtualization
At this stage, many companies are still working through their own server virtualization issues, and dealing more and more with the implications of moving applications to the cloud. Desktop virtualization has not been a priority, even if the technology is where it needs to be.
Is that likely to change in 2012? It’s hard to say. Other trends – such as the move toward computing devices like tablets and smartphones, or the proliferation of personal cloud solutions working their way into business – may take priority yet again.
It’s all about ROI
Some companies have held off on desktop virtualization because they’re not yet convinced that they will get a sufficient return on their investment. While the cost of upgrading PCs and OS in order to virtualize desktops are comparable to what they are for servers, the sheer number of machines and user accounts means that deployment requires significant resources.
Bit by bit
Still, desktop virtualization is moving forward. The percentage of companies using these technologies isn’t dropping; market share is rising. It may well be that desktop virtualization, because there are other priorities and because the ROI isn’t as great as other virtualization technologies, will have to settle for long-term growth rather than taking the market by storm.
Posted by Crystal Nichols on Mon, Dec 26, 2011 @ 07:57 AM
Everyone pretty much agrees that the new data center will be more streamlined, more uniform, more efficient, and less costly to build and maintain than today’s data center. In some ways, of course, this is wishful thinking; that ideal data center with limitless capacity that requires no staff to maintain it and no energy to use just doesn’t exist in the real world, and never will.
Further complicating the matter is the fact that there are plenty of technical and political obstacles for most companies to build the next generation data center.
What the data center would look like
If you were to start over from scratch, today’s data center would look very different from most of the data centers out there today. For example, it would likely feature x86 virtualization servers, converged storage, converged networking, and even cloud services to handle when demand exceeds supply. All of the storage capacity in the data center would be shared. The network would be flat within the data center, eliminating plenty of unnecessary hops.
In short, a brand new data center would utilize the best of what virtualization and convergence trends have given us over the past several years.
Cold reality
So, what is it that prevents companies from creating this data center utopia today? Well, for starters, existing organizations already have significant levels of customization. Organizations implement solutions and them tweak them according to business requirements, turning them into valuable resources that can also become sacred cows.
Calculating TCO is key
Building a case for the next generation data center is made even more difficult by the fact that it’s often hard to express TCO for a given system today. Pulling in all of the components – hardware, software, and more – is sometimes tremendously complex to do in a meaningful way. This means that it’s much more difficult for CIOs and other decision-makers to make the right decisions about how they can make the data center more efficient.
One step at a time
Few organizations can simply wipe the data center clean and start from scratch. The key to getting to the next generation data center for them is to do it one step at a time. Utilize virtualization technologies, implement convergence whenever possible, and avoid adding complexities at all costs.