Scalability is the capability of a system, network, or process to handle a growing amount of work, or its potential to be enlarged in order to accommodate that growth. - Wikipedia
I outlined how communication challenges can lead to problems in IT organizations in my last post. In this follow-up, I want to examine how infrastructure and development teams can work together better.
As far as most people are concerned, a company has “The IT team” and no one seems to understand the different roles within the team. Since IT services are gaining importance for businesses, it is important to bridge any communication gaps. But let’s not rush in. Every company treats this differently, but let’s look at generic definitions…
(Warning: You’ll never look at the toilet in the same way again)
Your Mission, If You Choose to Accept It…..
In a previous life, I worked with a Managed IT Services Provider in the UK. My role involved designing enterprise digital platforms using our state of the art data-centres and the public cloud.
What’s your vMotion Story?
If you have worked in IT for long enough, then you will have a vCenter vMotion story. The very first time you saw vMotion work. And the slow realisation about how it was going to change our jobs forever.
Let’s talk about ESX baby,
Let’s talk about Hyper-V,
Let’s talk about all the good things and the bad things,
That may be……..
OK, looks like I’ll stick to my day job.
In the world of Virtualization, it’s safe to assume that the two biggest players are VMware and Microsoft (Hyper-V). A quick look at the Gartner Magic Quadrant for x86 Server virtualization from July 2015 clearly illustrates this fact.
SOURCE: Gartner Magic Quadrant x86 Server Virtualization Infrastructure 2015 Gartner (July 2015)
If you know anything about the industry, then no one needs to tell you that VMware is the current market leader. With by far the largest market share, a very established product, and a very extensive portfolio of complimentary products, you would imagine it was all smooth sailing.
However, the competition is hotting up. The main challenger at the moment is Microsoft’s Hyper-V. After some growing pains, Hyper-V now has a fairly mature product, and several organizations are looking at moving some workloads to Hyper-V, particularly Small and Medium Businesses.
With the release of vSphere 6.0 earlier in the year and Hyper-V 2012 R2 (released in 2013), let’s take a look at both products.
Let the battle begin.
VMware – The undisputed Champion (for now)
But for how long?
Conventional wisdom was that VMware was the only option for larger virtualization deployments, and Hyper-V was great for small to medium deployments. This is no longer true. With the 2012 version, Hyper-V has a lot to offer large environments; and on the other hand, VMware can be a great option for smaller environments.
Round 1: Thinner is better
VMware’s base hypervisor - ESXi - is as thin as you can get, it’s what is known as a bare-metal hypervisor. With version 6, the footprint has been reduced to 150 MB. In contrast, Hyper-V requires some components of Windows, the bare minimum is Windows Server Core which is about 5 GB. However, most people use a full blown Windows installation, with a massive footprint of 9.1 GB, over 60 times more than ESXi !
ESXi’s light footprint means there is very little overhead and a very small attack surface for malware and other threats. Think about how many patches and reboots a Windows Server installation needs every Patch Tuesday, you can see why having a slimmed dow n hypervisor makes sense.
VMware wins this round convincingly on points.
Round 2: Bigger is better
As previously stated, VMware has for a long time been the first choice for larger virtual estates, however, Hyper-V is muscling in on that territory.
In terms of scalability, VMware now supports up to 64 hosts in a cluster in vSphere 6.0, double the previous limit in vSphere 5.5. A single host can now accommodate up to 1,000 Virtual machines, up to 480 physical CPUs and up to 12 TB of RAM. In comparison, Hyper-V 2012 supports up to 64 nodes per cluster, up to 8,000 Virtual machines per cluster, and individual Hyper-V hosts can have up to 1,024 running Virtual machines, 320 physical CPUs and up to 4 TB of RAM.
Also, with vCenter Server, you have access to several features that are simply more mature than Hyper-V’s offering. In order to get a similar level of capability, you will have to license System Center 2012 R2 with Virtual Machine Manager (SCVMM). You’ll also need to use some PowerShell, and multiple interfaces.
ESXi is still the hypervisor of choice for larger estates, but only just. A close fight, but VMware only just edges it here.
Round 3: The best things in life are free (well, almost free)
The perceived wisdom around Hyper-V is that it is free. When I speak to people about why they are looking at moving workloads to Hyper-V, the primary reason is almost always cost. But is this really true?
On the one hand, a Windows Server 2012 license allows you to run two virtual instances, while a Windows Server 2012 Datacenter license comes with unlimited instances on a single machine. If you’re already going to invest in the Windows Operating System anyway, then yes, it really is free. Well almost free.
In reality, for you to manage multiple Hyper-V hosts efficiently, you will need to run (and license) System Center 2012 R2 with Virtual Machine Manager (SCVMM). By the time you take this into account, the price savings vs. vCenter effectively disappear. VMware’s marketing department are very keen to point this out.
It’s not so clear which is cheaper, it depends on your circumstances. Organizations that are large enough and make extensive use of Microsoft Technology will probably have to run System Center 2012 regardless of which hypervisor they use for virtualization. Smaller organizations can probably get away running standalone Hyper-V. Based on these assumptions, Hyper-V comes on top. But only just.
Round 4: Waiting on that Ready Queue
This ‘little known’ problem is getting more and more prevalent, and is unique to VMware ESXi.
The reason a lot of people are reluctant to virtualize certain applications is the overhead that virtualization adds to any server. It’s an extra layer of management for your IT Administrators to keep an eye on.
And sooner than later, Ready Queue will start to creep its ugly head.
So, what is Ready Queue?
My colleagues have pretty good explanations here and here, so I won’t reinvent the wheel. In one sentence, it’s the time a virtual machine must wait in a ready-to-run state before it can be scheduled on a CPU. This is down to the way the CPU scheduler works in VMware ESXi.
Basically, the larger your Virtual Machines (in terms of CPU cores), the more likely you will have Ready Queue. This problem primarily affects people who erroneously think that a larger Virtual Machine gives them better performance.
While this is true in the world of Physical servers, the reverse is the case on VMware virtual infrastructure, it leads to performance problems that are difficult to pin down, and even harder to resolve.
Not to say that Hyper-V is all smooth sailing, but on this one, easy win for Hyper-V.
All depends on your personal circumstances.
The truth is that in terms of capability, both Hypervisors are pretty much neck and neck. There’s loyalty to spare on both sides of the equation, people have used Microsoft technology for years, and will consider them for any solution, and VMware has also has a fairly established army of supporters and users.
Hyper-V is due to get an upgrade with the release of Windows Server 2016 which is literally round the corner, so let’s see what the future holds.
And finally: Watch this space
There’s a new kid of the block. OpenStack.
What is it? It’s a free (and we love FREE!) and open source cloud computing software platform.
“The OpenStack Mission: to produce the ubiquitous Open Source Cloud Computing platform that will meet the needs of public and private clouds regardless of size, by being simple to implement and massively scalable. OpenStack is open source, openly designed, openly developed by an open community.” - openstack.org
Keep an eye on this one, the only way for OpenStack is up.
“Hello, Helpdesk? Email is down”
“I’ve just sent everyone home. We have lost a whole day of business. I don’t care how you do it, but when we get in tomorrow morning, email had better be working.”
There are some conversations you never forget. The date was July 21st, 2005, and the Chief Operating Officer walked out of the server room, and my manager and I sat back down, after our dressing down. We had been battling a dead Exchange Server for over 3 hours, with no resolution in sight. It took another 3 hours to get it back up and running.
A few weeks ago, a patch was issued for the widely used Xen hypervisor to fix a massive security hole. It didn’t get as much coverage as the VENOM bug earlier in the year, but the vulnerability is similar, it allows a guest Virtual Machine to access the Host Operating System, possibly running code on it. It’s called a “VM escape” or “Guest escape” bug.