 Hello, hello everyone My name is Venku Mujankov, and I'm solutions architect at store pool in my daily work. I spend a lot of time listening to our customers and prospects and One of the things that they tend to go on about is how the traditional storage area network model even when build with a really good storage race just doesn't work well for them at its scale and Now we will see eight of the most common flows of the traditional storage When it's deployed at large scale clouds Enterprise environment are usually fairly static They are with less frequent changes cloud environments on the other hand are much more dynamic During the life cycle of a typical storage system that is three to five years. They can scale significantly Very often the cloud. We will need to start with a very small capacity relatively small capacity that soon will be upgraded significantly and It is difficult to plan for these several years ahead The traditional storage arrays When it comes to scalability are limited They can scale up by adding shelves and drives to increase the capacity But the system is limited by the performance of the controllers and the bandwidth of the buses and the storage interfaces Bandwidth Internet is the primary protocol in data centers and modern internet Networks are optimized for this use case on the other hand fiber channel Sorry Sun architecture is built on fiber channel network and this requires that you need to build parallel storage network to your existing Internet network and large suns Can be very expensive and difficult to build and to operate modern cloud data centers operate in an always on operation model without maintenance windows applications work 24 by 7 and building an infrastructure for always on Operation is not just about fault tolerance. It's about everything you need to provide reliability high availability and data integrity it includes search things like in service software upgrades in service hardware upgrades Online reconfiguration these are all things that you just can't get with the most ordinary Sun models storage controllers often becomes performance bottleneck in the traditional arrays and if controller becomes performance bottleneck for your service then You have your choices are limited to basically three pleasant options. First is to upgrade the storage controller That is usually a disruptive operation It is it could be very expensive and not always possible. The second option is to upgrade the storage array That is often referred to as a forklift upgrade that is basically replace the Storage appliance with a new model more powerful model Again, this is a very complex task. We can see this later it can be expensive and it's almost always is service disruptive operation and The most common choice is just to find a way to live with your limited performance. This means to optimize your applications to partition your cloud Into several systems do load balancing and a lot of stuff to deal with a limited performance You you have so this makes your cloud Very complex and difficult to operate large-scale or Multirack availability is something that is difficult to manage with a traditional race For example, taking several racks down to address a power issue Would require a major project with a synchronous application with a with a sound solution and cloud providers can't afford to Have their customers down Even when the outage is planned After several years of operation every hardware has to be replaced replacing compute nodes is a trivial task because compute nodes are interchangeable replacing the storage system on the other hand can be a cumbersome operation that involves time-consuming data migration from one system to another and If the application use integrated features of the storage system like snapshots and remote Replication it gets even more difficult in clouds typically storage systems are tightly integrated with the cloud management's platform and replacing the storage system with a new model often from a different series or from a different vendor with a Different set of features and most importantly with a different integration interface We'll induce a new separate project to integrate the new system with your cloud management What was initially a simple task of hardware refresh? Cycle now becomes a complex project that involves building an integration data migration and possibly services reconfiguration Traditional storage arrays Generally have a slower adoption of innovation compared to industry standard servers whether it's a hundred gig ethernet or a larger NVMe drives or a faster drives or Communication protocol like let's say NVMe TCP All these are available in the specialized arrays years after these technologies are available in the general purpose servers The result of this is that the storage arrays are with smaller capacity less storage density They have a they consume more power are more expensive to operate and Maintain and upgrade They cannot provide the features that you need and the end result is that they make your cloud services not competitive Traditional arrays are also Difficult to manage at scale by running multiple separate storage appliances You always end up with running many multiple silos in your cloud to take care of and instead of focusing on Improving your services cloud and storage administrators are busy more and more time dealing with issues that this individual silos creates like capacity management load balancing dealing with hotspots optimizing the traffic Inefficient utilization of available resources So there are products that are designed to address these problems, but these are partial solutions and It's better not to have these problems in the first place All these flows Comes from the fact that the Sun architecture was designed for enterprise environment It fits very well there, but the clouds are very different from enterprise environment they have a different requirement and They need a new different type of storage storage that is designed from the from the ground up for for this specific use case and not retrofitted from another use case So thank you for your attention and I'll be happy to hear your thoughts about About this and about operating Sun in cloud environment Do you have What are the most challenging part that you see with with Sun in this environment and Yeah, do you do you see some of this in in your daily daily work? Thank you