Extending the hardware

An interesting conversation yesterday with folks from AMD, on what’s going to be happening with the next generation or two of their virtualisation technologies in Opteron.

There’s a lot to be said about “Trinity”, their secure virtualisation platform, and “Raiden”, a client model for blade servers – but the really interesting story is “Terrazzo”. This is where AMD opens up its HyperTransport pipeline to third-parties – as well as its socket specifications. So on a multicore, multiprocessor motherboard, you could drop in a physics coprocessor for fast gaming, or (and this is where I think things will get very interesting) a dedicated processor for additonal server functionality.

This is where AMD needs to talk to companies like Azul. Dropping one of Azul’s 48 core VM-specific processors onto a HyperTransport bus alongside a set of Opterons could really speed up your Java applications (with direct access to the system memory) – and get rid of all that nasty non-deterministic garbage collection…

AMD is taking the enterprise server game in a very different direction to Intel. Let’s see if the industry takes them up on it…

Technorati Tags: azul, amd, terrazo, processors

 

Hyperjacking hits the headlines

Eweek has an interesting piece on a proof of concept hyperjacking rootkit that’s about to do the rounds of the security shows…

“The idea behind Blue Pill is simple: your operating system swallows
the Blue Pill and it awakes inside the Matrix controlled by the ultra
thin Blue Pill hypervisor. This all happens on-the-fly (i.e. without
restarting the system) and there is no performance penalty and all the
devices,” she explained.

Rutkowska stressed that the Blue Pill technology does not rely on
any bug of the underlying operating system. “I have implemented a
working prototype for Vista x64, but I see no reasons why it should not
be possible to port it to other operating systems, like Linux or BSD
which can be run on x64 platform,” she added.

Interesting times…

 

Writing about Azul

I spent last Friday morning braving the delights of Highway 17 over the Santa Cruz mountains in the rain at Azul Systems' offices next door to Google in Mountain View, learning lots of interesting stuff about their Vega processor and their network attached processing tools, including their "pauseless" Java garbage collection.

You can read about some of my morning at The Register:

Adding storage to a network is straightforward; adding processing power tends to involve a lot more complexity. This is something Azul Systems aims to change. Following the recent announcement of its second generation Vega processor, is today’s news that BT will be using the company's processing appliances to handle both its existing web applications, as well as providing the foundation for a utility computing farm – part of BT’s 21st Century Network.

The Azul platform is more than just a box you connect to your network, which replaces software virtual machines. It’s also a set of tools for managing application performance and handling how you bill the rest of the business for CPU usage. Mainframe administrators will be familiar with these techniques, but they’re still new to the arrays of application servers that now run many of our businesses. Being able to bill for actual CPU and memory usage is a key part of any utility computing platform – whether it’s Sun’s $1 per CPU per hour or an IT department billing the rest of the business for application operations.

They've got quite an impressive server room too, especially when you realise that each of those boxes has 384 cores – so that's the equivalent of 9600 CPUs in this rack alone:

Not bad – and what's more important, not too power hungry.

 

The Hypervisor Wars

I’ve realised I’ve mentioned the idea of the hypervisor wars without explaining what I mean by it.

The underlying virtualisation technologies used in Intel’s VT and AMD’s Pacifica curently only allow a single VM Manager to run. This means that the VMM (the hypervisor) installed has an incredible amount of power – it controls what runs and how it runs. Install yours first, and the machine is yours – especially if you lock your hypervisor into TPM or similar security mechanisms.

So what would the hypervisor wars mean? Firstly an end to the open systems model that’s been at the heart of enterprise IT for the last 25 years.

If Microsoft and VMware fell out, VMware could reduce the priority of Windows partitions. Other hypervisors might have licensing conditions that make it impossible to run non-free OSes as clients.

You could end up with a situation where each OS installation would attempt to insinuate its own hypervisor onto the system partition. Security partition developers may find that they are only able to code for one set of hypervisor APIs – locking end users into a closed platform.

The end state?

Co-opetition breaks down, the industry becomes enclaves built around hypervisor impementations, and the end user finds that they’re unable to benefit from the possibilities of an open hypervisor architecture.

Can we avoid the hypervisor wars? Optimistically I think we can. There are pre-requisites. We need an agreed hypervisor integration architecture, and we need it quickly. Let VMM developers compete on ease of operation and management, not on who controls your PC.

Technorati Tags: hypervisor

 

How long before there’s an Apple Hypervisor?

One thing to note about the new Apple Intel machines is that the Yonah chipset supports VT.

With Apple saying that they’ll let Windows run on their hardware, the question is – will they let a third-party hypervisor run? I suspect not – especially if they are using TPM in secure startup mode. Of course, they’ll first need to enable VT in whatever BIOS they’re using…

So will Apple produce its own hypervisor, or will it badge a third-party tool? My personal suspicion is that Apple doesn’t have the skills to write it’s own hypervisor (there are only a limited number of people with the deep combination of hardware internals and OS knowledge required, and they’re mainly at Microsoft and VMware) that they’ll announce a partnership with VMware at the WWDC. Unless Apple’s been hiring the Xen dev team on the sly…

Apple will quickly need to gain the high ground in managing virtualisation on their platform – as they’ll need to maintain contol of OS X running as a VM. Otherwise, will Apple be the first casualty of the hypervisor wars?

Technorati Tags: apple, hypervisor, intel, yonah