VSPEX – EMC’s entrance into Reference Architecture

Hello everyone and welcome back to Journey to cloud blog, where we look at the various cloud options available in the market from leading vendors and their advantages and dis-advantages. I know that in my last blog, I did talk about exploring UCS in detail, but that would be more product specific and it would be good, if we explore the options that we have in the market first before getting down into the product details.

In today’s blog, we will try explore EMC’s latest offering (VSPEX). Though the product has been out there in the market for a few months now, it is still to gain momentum compared to the other reference architecture products like FlexPod or Hitachi. So, with that lets try to explore VSPEX.

VSPEX as the product is called (really don’t know why – EMC has a passion for naming its products which ends with an X, VMAX, VNX, Symmetrix) is EMC’s latest offering to cloud based on the reference architecture model.  Now as discussed in my last blog, reference architecture is a set of best practices by which different components that make up the cloud are connected together to get predictable performance.  Apart from that, reference architectures deal more with the support offered in administering the product and also on various levels of orchestration that can be implemented on the product.

VPEX being a reference architecture product is essentially made up of EMC storage. But unlike it big brother the VBlock, it does not have Cisco and VMware tied to it as a bundle nor it contains the monster called VMAX.  The VSPEX offering comes with VNX storage’s various flavors and along with that, various flavors of network, servers and virtualization solutions can be coupled.

The core component of VSPEX is VNX- its EMC’s second flagship storage product after VMAX. VNX is basically a combination of two of EMC’s previous storage offerings, the Clariion and the Celerra. Clariion was SAN storage offering from EMC that served the mid size customers effectively and Celerra was the NAS solution from EMC which used either the Clariion or Symmetrix storage for backend and provided file level access to the end users. But Clariion could never provide file level access by itself and Celerra could never block level access. I think EMC wanted to compete with Netapp at all levels since one Netapp box can provide both file and block level access and they came up with VNX. Now, only the name is different. VNX offers pretty much everything the Clariion and Celerra had to offer in terms of replication, disaster recovery and high availability.

The other components that make up the VSPEX are servers and networks along with Virtualization. Now, this is where EMC has taken a giant step by allowing other vendors/competitors products to integrate itself with VNX. For example, EMC and HP are direct competitors in terms of storage, but the VSPEX model is fully compatible with HP servers. Other than HP, EMC has extended its collaboration to DELL. Now, what about UCS, the main component in Vblock.  EMC is still keeping the relationship intact with Cisco on the server front, so expect to see UCS in some Vspex configurations.

Similar to servers, EMC has also extended its network collaboration to Brocade and other vendors while still continuing its association with Cisco with Nexus switches. But one of the biggest strategic move is the decision to work with Citrix and Hype V along with VMware. Citrix and Hyper V are direct competitors to VMware and VMware as everyone would know it is actually a EMC company. But EMC have gone the smart way for trying to keep their core business separate which is storage and including the other two Virtualization companies was a good move. Now again, this is to compete with Netapp’s Flexpod who have also moved from a VMware centric solution to a Virtualization vendor free solution. This is move is both good and challenging for EMC as with any reference architecture solutions, once you sell the product, you have to support it and supporing HP or Dell along with Citrix can be quite difficult. But from a customer point of view, I don’t have to rely solely on VMware or UCS if my existing Data center is a HP workshop with Hyper V.

Other than these core products, there are quite a few additional products from EMC that add value to VSPEX like Data Domain, which provides backup and recovery solution, RPA which provides multi snapshot recovery.  Overall, VSPEX is a good solution from EMC to compete with Netapp on the Flexpod market, but how good the solution is can only be determined in the days to come.

Take care,

For DCV

K

Cloud Architecture at a glance

Journey to Cloud

DCV Engineering

Hello Everyone, DCV Engineering would like to wish all of you A Big Welcome to a Journey to the Cloud.

DCV Engineering was formed in 2011 as a Cloud and DC services company based in APJC.  In addition to cloud services we offer other solutions such as server, storage and network based services. If you want to know more visit our website at www.DCVEngineering.com.

I know for sure that my boss is not going to be happy with me for putting such a small introduction, but this is a blog site and I am free to do whatever I want :)  Additionally, the purpose of this blog is not to talk about DCV but to talk about Cloud and Cloud solutions. Instead of me telling you what DCV can offer or do, I want the people who read this blog to decide based on what we put into this blog.

So, enough with the introductions and lets get going with what we are going to discuss in this blog. Now I know that lot of people, when they start a blog on cloud, they start directly talking about the different cloud offerings, IAAS, PAAS, XAAS and other offerings. Now, I am going to be very frank here. I am going to be very biased and am going to talk a lot about Cisco and the cloud offerings from Cisco.

Now, why Cisco? Is it because I am a solutions architect and also a trainer who does lots and lots of trainings for Cisco? Or may be due to the fact that I am a Ex-EMC employee who spent the last 7 years of his life doing nothing but working on EMC and Netapp products (I know, EMC employee, Netapp products, can be confusing, but hey, that’s the way I used to work at EMC, working on everything the customer had) and most of the cloud offerings from EMC and NetApp are tied with Cisco. Or is it because Cisco is the market leader in network based solutions and they have a very good server market in a short span of time.

So, to start of the blog, we are going to have a small discussion on the different type of cloud infrastructures available. Now, this is not the IAAS or any offering, but rather, how Cloud data centers are setup. I feel the need to discuss this as I find most people (apart from cloud architects or cloud pre-sales people) don’t have an idea on the different infrastructures available. Once we have discussed that, we will discuss UCS, Nexus, Vmware, Cloud Orchestration solutions and Storage over the course of the next few weeks.

Cloud Infrastructures:

Many people I know want to move into Virtualization and Cloud, but they are not able to decide what is the best way to move forward. For some them, server Virtualization is cloud. Also, people need to know what kind the Cloud infrastructure the vendor is offering them so that they can make sound business decisions.

There are three different type of infrastructures available in the market today:

1.  Stack Architecture – As the name sounds, stack architecture is just that, a stack of components that come from one vendor. A good example is something like open stack or a cloud solution from HP or IBM. Most, if not all of the hardware are from one vendor. Now there are advantages and disadvantages.

Advantage – everything is from one vendor, which means that one throat to choke.

Disadvantage – everything is from one vendor, which means one throat to choke :) The vendor might not have the equivalent of some of the industry innovating technologies from other vendors. And since it’s a stacked architecture, we cannot incorporate other vendors inside the cloud offering. Not so good from customer point of view.

   2.  Mixed vendor: Now in mixed Vendor, we have two type of architecture available. Validated Design vs. Reference Architecture

      a.  The best example I can think of for a converged infrastructure – Validated design is something that 8 out of 10 people have right now – an apple IPhone.  Now, I know most of you people think that an IPhone is stack architecture, but no, it’s not stack architecture. Samsung and other companies make most of the IPhone components for Apple. Apple is just rebranding it and selling it in the market. Similarly we have VBlock which is again a converged infrastructure made by VCE, which in turn uses products from EMC, Cisco and VMware. Again, we have advantages and disadvantages.

Advantage – Guaranteed performance, how many times have you seen an apple IPhone crash? How many applications or apps, than an IPhone can run, without compromising on performance?  Same thing with Vblock, its guaranteed for performance. 

Disadvantage – You cannot run flash in an IPhone, you cannot install applications on IPhone without jail breaking it, and if you do that, there is virtually no support for your phone anymore. In short, your IPhone was designed with a purpose in mind, and we cannot alter it to suit our needs without compromising on support and performance. Sometimes this can be itchy for some people.

           b.  Reference Architecture – The best of both worlds, kind of using a Samsung phone. It still uses components used by leading industry vendors, but its fine tuned so suit the customer’s needs.

Advantage - It’s a set of best practices by which you should connect the different components to achieve requested performance, but unlike a converged infrastructure, the components are not tied to each other. It means that, a customer for some time, can compromise on performance, while running a higher load on storage, and he may buy the additional server and switch later, when he has the money and time. Another advantage when compared to a CI is that, the customer need not buy all the components that make up the cloud initially. He may start with a simple virtualization solution with shared storage and then move towards a RA cloud, when he feels he is ready.

Disadvantage – There is no single point of support in a RA solution. Try calling Samsung and tell him that an app is not working properly; the reply is going to be, please call the app owner. But, its not that bad, considering that, most of the RA partners are interlinked, and they will eventually, find a solution for you, but during that time, you may just have to wait.

Now, I hope that this gave us a good picture on the different type of infrastructures available to us in the market. In my next blog, we will probably try to look at Cisco UCS and its features. Till then, take care.

For DCV

K