I’m on my way to Orange County for a few meetings this week and I find myself with some time to actually sit and write a couple of postings without any interruptions!  I always find cross-country flights a good time to catch up on my writing.  My only problem is that my two batteries usually don’t  last for the whole flight so I’m going to have to type fast!

In a previous posting I mentioned that one of the technologies that is changing the storage industry is Solid State Drives (SSD’s).  Back in January EMC announced that all of it’s core products will now be offered with SSD’s.  We’re also beginning to see startups like Pilant Technologies sprout up that are focusing solely on this space.  These events really mark the beginning of the end for spinning media.  SSD’s have been talked about for many, many years but the technology is finally at a point to make the technology viable for this applications.

SSD’s today are made from FLASH memory devices.  FLASH technology is what you find in digital camera’s, some MP3 players, and USB memory sticks.  Over the past decade or so FLASH memory has proven itself as a reliable and rugged technology.  Although FLASH devices haven’t hit the cost/MB that spinning media has there are other industry trends that are driving the adoption of SSD’s.

FLASH based SSD’s provide four key attributes that are attractive to storage vendors these days.  First and foremost the densities are approaching parity, although spinning media still has a slight edge.  Second is power.  SSD’s offer a significant savings in power consumption.  These days anything "green" seems to be able to sell and SSD’s have a strong lead over spinning media on this front.  Another important metric is performance.  SSD’s are inherently faster than their spinning media counterparts.  One reason is that with SSD’s there is no rotational latency, the proper devices are addressed directly.  The last major factor is reliability.  SSD’s have much longer Mean Time Between Failures (MTBF’s) than spinning drives do.  Mechanical parts are prone to higher failure rates than their solid state counterparts.

What we are seeing today is that Storage Array vendors are swapping their spinning drives for SSD’s.  They have fabricated SSD’s to fit the same footprint as the traditional spinning disk drive.  I guess one benefit of this approach is that customers now have a choice and they can decide whether to use spinning media in one tier of storage while using Solid State Drives in other tiers to improve performance and reliability.

As long as the Array vendors don’t unduly influence the purchasing decisions of customers I think you’ll see a large number of customers opt for the SSD technology.  Why wouldn’t they?  And once that really begins to take hold I bet you’ll see some innovative "redesign" of the Storage Array.

I wonder if this transition will be the death to the traditional storage protocols like SCSI.  Wouldn’t RDMA technologies such as iWarp make more sense?  I think the SSD revolution will have a much greater impact on the storage industry than any other previous transition.  I’m looking forward to be part of the evolution of the storage array over the next decade….it’s going to be an exciting ride!

VMware announced today it is acquiring B-hive Networks.  I think this is a good acquisition for both companies and extends VMware’s presence in the data center.  Obviously VMware is hearing from its customers that they want to manage their virtual data centers as a service.  Web, application, and database performance is critical to that goal.

B-Hive’s Conductor product monitors network traffic and uses that information to map applications to the supporting infrastructure and reports on measured response times of the monitored transactions.  This gives a good view of how well your service is performing.  However Conductor does not have visibility into the SAN and storage layers and can’t help you monitor, troubleshoot, optimize, and plan those pieces of the infrastructure.

I would think the next step for VMware would be to make the information from the B-Hive product accessible via Virtual Center.  Products such as BalancePoint will benefit from that web and application centric analysis and complimenting that with the analysis and modeling that we do will give the customer a complete end-to-end view of their IT service.

Well it’s trade show season again.  May and June are typically busy months for industry trade shows and this year is no exception.  Next week we’ll be demonstrating our award winning product, BalancePoint, at two of these shows, Citrix Synergy and EMC World.

I’ll be at the Citrix Synergy Show in Houston.  Our booth number is 423.  If you are going to be there please stop by and see our product in action.  Tom Joyce will be representing us at EMC World in Las Vegas.  Our booth number there is 108.  If you are there please stop in and say hello.

Recently I was asked to give a talk  to one of our partners about the Future of the Data Center.  Now there’s a broad topic!  The industry has been talking about the "Future Data Center" for decades and there have been many definitions of what it will look like.  I see the Future Data Center as a "lights out", service delivery operation.  This Data Center will provide compute and storage services and do so autonomically based on workloads and service level agreements.

Until recently though this vision was impossible to implement.  The Data Center was a fairly static place where modifying resources to support business applications took hours if not days to accomplish.  Providing true service levels was impossible, IT managers didn’t even know how to define a service level let alone implement one. 

The adoption of virtualization technology changes that though.  Compute, network, and storage resources all become very dynamic and modifications to these resources can happen in real time or near real time instead of the hours and days it would take in the past.

Two other key pieces of technology have evolved that will make the transformation of the Data Center complete.  One key technology is autonomic elements.  In order to provide true "lights out" operation elements within the Data Center need the capability to "self heal" and automatically provision based on changing requirements.  Recently Xiotech announced their ISE line of storage arrays.  This storage array has the capability to predict when drives are becoming problematic and re-calibrates the drives before issues arise.  It also has the capability to self heal using it’s "Spare-In-Place" technology.  This is the first autonomic storage array I’ve come across and I’m sure you’ll see more vendors going down this path in the future.  Of course this technology will become easier to implement as we adopt SSD’s (Solid State Disk) and obsolete the spinning rust we deal with today.  I’ll talk more about that in a future post.

The last piece of technology that is needed in order to realize this Future Data is what my friends, Ellen & Richie Lary, calls the DCRM or Data Center Resource Manager.  They describe the DCRM as

performs several important policy-based functions, utilizing the various system management utilities available in the data center as its eyes and hands.  The DCRM maintains knowledge of what resources are available to allocate to applications, and schedules an application for execution based on priority and the availability of appropriate resources.

Beyond providing policy-based functions I believe the DCRM needs to also offer a rich set of Analytical and Modeling capability in order to truly manage the Future Data Center.  A policy engine alone can not provide the dynamic, system level knowledge that the DCRM requires to make it’s decisions about which resources to deploy and when to deploy them.  Another aspect of the DCRM is that it needs to not only respond to current conditions but also must have predictive capabilities.  In this new paradigm IT service providers have to have an insight into the future needs of the business.

Akorri’s BalancePoint is a good example of a DCRM.  BalancePoint provides the necessary visibility across all layers of virtualization and with a rich set of analytics and modeling is able to identify trouble spots and make recommendations about future service capabilities. 

All three technologies, virtualization, autonomic elements, and DCRM can now be combined to offer a true Future Data Center.  My prediction is that you’ll see the first "service based" data centers rolled out in the very near future and then the real excitement begins.

One of the powerful aspects of Akorri’s BalancePoint product is that it works across the technology domains within the data center. You’ll hear us refer to our technology as being “cross-domain“. This means we do not just monitor a particular technology within the data center, such as servers, but we monitor all “domains” including applications, servers, switches, and storage devices.

So why is being “cross-domain” important? The answer is pretty straight forward. Without looking at all the domains you can not get an accurate understanding of how your infrastructure is working as a complete system.

Let me show an example of this. One of the most popular i/o monitoring tools in the UNIX/Linux world is iostat. Server administrators use iostat to figure out how well utilized their storage devices are. The problem with this is that iostat was written for simple disk models, such as a Direct Attached Storage (DAS) device. Once we migrated to more complex storage models and introduced SANs, then iostat doesn’t produce accurate results anymore. If you used iostat to fix a problem or plan for future needs then you are probably in for a big surprise.

Figure 1 below shows a graph that was generated using the data from iostat. It compares queue depth vs busy time of the storage device. Notice all the blue points clustered on the far right. Analyzing this graph you’d assume that the drive is very busy and highly utilized. You might even make the assumption that this is the root cause of some of your performance problems and you may go out and buy more storage in order to offload the existing device. You’ll probably find out that it actually doesn’t solve your problem.

Figure 1

Figure 2 below shows another graph of the same test just from a different perspective. This time we are monitoring the traffic using a Fibre Channel Analyzer at the storage device itself. As you can see a very different picture. This time you see the storage device is about 80% busy, still highly utilized but probably ok.

Figure 2

So as you can see if you used your server centric tools to understand how your server was interacting with your storage subsystems you’d probably be making some very bad assumptions about it’s operation.

A lot of products these days are beginning to appear in the marketplace that claim they can help you manage your data center. Products like VKernal’s Capacity Bottleneck Analyzer have the problem of being domain centric. Capacity Bottleneck Analyzer can not possibly tell you with any accuracy how well performing or utilized your storage system is due to it’s server centric view.

In order to fully understand the operation of your complete system you need to have a true cross-domain view. My advice before you purchase any tool to help you manage your data center is to ask the question of the vendor “Does your product support a cross-domain view of my infrastructure.”

Gartner recently released it’s Cool Vendor Report in Server Virtualization Management and Akorri was named one of the Cool Vendors in this space.  I’m excited about our company being recognized as a "cool vendor" in this emerging space. 

It’s clear to me that Gartner understands that the Virtualized Data Center requires a new paradigm when it comes to managing this dynamic infrastructure.  A quote from their Key Findings points this out. 

The ease with which virtual machines (VMs) can be created and/or removed as well as dynamically relocated is calling into question the capabilities of many traditional management tools that were not originally designed for this environment.  

In the report Cameron Haight states that one of the challenges he sees for Akorri is

Akorri’s most direct competitor was a recent introduction by Onaro, which was recently acquired by NetApp, and the combination of these two companies may provide the means to potentially cut into Akorri’s technology lead. 

I don’t really understand this since Onaro and Akorri provide solutions in very different spaces.  See my previous post on this topic. 

Shortly before the acquisition, Onaro started talking about providing a solution called VM Insight.  From looking at their website it seems to me that what VM Insight does is that it pulls statistics from VMware Virtual Center and redisplays those stats in their GUI.  It seems that is a popular thing to do these days.  Another vendor, VKernel, announced that they were going to provide a product that does the same thing.  It’s not clear to me what value that really brings to customers since all that information is obviously available in Virtual Center.  In VKernels case they can’t even provide a cross domain view.  The only visibility they have is what Virtual Center provides to them. 

Akorri’s BalancePoint on the other hand collects information from the Database, Server, Switch, and Storage layers and uses that information in it’s Analytics and Queuing Model Engines to give customers a detailed analysis of what’s going on with their systems.  Other than seeing the complete system it’s hard, if not impossible,  to provide information that will allow customers to effectively manage their virtualized data center.

Last month Akorri announced it’s release of BalancePoint 2.0. This is an important release for us and it’s been a long time in the making. In the next few blog postings I’ll talk about some of the new features and functionality that BalancePoint 2.0 brings to the market.

Today though I want to talk about one of the major enhancements we made in BalancePoint V2.0 that is not obvious to our customers. That enhancement is the implementation of a “services based architecture” (SBA).

There is a lot of discussion in the industry today about service based architectures but for Akorri the main reason to implement a SBA is the flexibility and agility it provides in offering new features and applications. Having well defined infrastructure services such as data access, reporting, security, etc will make the delivery of new features and applications easier and quicker to develop and deliver to the marketplace. Our SBA implementation will allow us to offer new applications without have to “rip up and replace” any of the existing infrastructure. Scalability, overall performance, and adaptability are other benefits we’ll see with this architecture.

In an upcoming post I’ll talk about another important feature of BalancePoint 2.0 which is our delivery of BalancePoint as a virtual appliance. There are a lot of exciting things happening in the virtual appliance world and I’ll discuss some of the benefits we get from offering a virtual appliance as well as some of the challenges we faced.

We found out today that we won the Silver Award for Product of the Year in the Performance Management Products from SearchDataCenter.com. As you can imagine we’re pretty proud of our accomplishments in 2007 and we are looking forward to a busy 2008.

It seems everybody is getting into the server virtualization game these days.  Back in October Sun announced it was entering the server virtualization space with it’s $2B investment into Sun xVM.  In November Oracle announced it’s getting into the space with Oracle VM. 

It’s interesting to see two strong partners of VMware decide that this technology is too important to leave in VMware’s hands.  I think what these and other vendors have figured out is that the hypervisor is the new Data Center OS and if you want to control the Data Center you need to own it’s OS.

Another interesting fact is that both Sun and Oracle have decided to build their solutions around the open source hypervisor, Xen.  Although I haven’t seen a lot of Xen deployed into my customers Data Centers I predict we’ll see a lot more of it in 2008. 

It’s going to be interesting to sit back and watch as VMware, Citrix, Microsoft, Sun, Oracle, and the host of others do battle for the customers dollar.  The good news is customers have a lot of choices now, the bad news is customers have a lot of choices.

In a recent blog Tony Asaro of ESG writes about his view of the future of IT Infrastructure. I think Tony rightly sees the future data center as one where resources such as CPU, Memory, and Storage are all resources that can be “pooled” to perform a particular task. As Tony points out

Perhaps it is time to consider gluing all of this stuff back together again. Why not run user applications and data management within the same infrastructure? You can have dozens and even hundreds of CPUs in a single rack. You can support 100s of TBs and potentially a PB of capacity in a single rack. Servers are made up of CPU, memory and disk capacity. Storage is made up of CPU, memory and disk capacity. Why aren’t we consolidating these resources?

These pooled resources are then defined by what type of application they are running. With the addition of virtualization technology these resources can be dynamic and shared. Resources can be allocated and reallocated on an “as-needed” basis. Today’s virtual infrastructure really takes the modern data center back to the past. With pooled resources, multi-threaded processors, and high speed IO channels the data center of today looks and works a lot like the data centers of 3 decades ago when mainframes ruled.

One big difference between the mainframes of past versus today’s virtual infrastructure is that in the past the mainframe was manufactured by one vendor and it was very well instrumented. Today’s virtual infrastructure is heterogeneous with components coming from many vendors and is not very well instrumented, therefore making it difficult to manage.

Tony recommends we call this new data center “Infrastructure 2.0″. I think maybe we should call it “Mainframe 2.0″.