Oracle Exadata Reviews

It provides performance for our highly critical applications.

They seem to be doing real well with keeping up on the performance and everything, especially what I have seen with the new chipset that is being offered. I can't really think of anything that they are not doing that they should be doing.

It can be difficult to patch and maintain because there are so many databases running on it.

We have been using Exadata for about five years.

We have had no down time with it, but it can be difficult to patch and maintain because there are so many databases running on it.

It is an absolutely scalable solution.

The support is very good.

I'm not involved in operations that much, but from what I understand, they get back to us; they always find the right person.

Oracle comes in and does the initial setup.

I recommend it for highly critical, performance-sensitive applications. You have to do a TCO on it because it is a fairly expensive solution. There has to be a cost-benefit analysis done on that.

Oracle Exadata provides very good performance and reliability. It's a highly reliable system and provides maximum performance.

The business outcomes are much faster; end users are very happy so they can focus on their day-to-day work rather than waiting for the computer, the response from the system. In many customer cases, reports that used to take 10-12 hours get completed in 1-2 hours, a massive reduction. A business cycle that used to take about 3-5 days to turn around can be turned around in 1 day, so that's where we optimize our business utilization; much more efficiency.

A major concern from customers is that it's very costly, but if they think about the total cost of ownership over 3-5 years, I think the total cost attests to the business readiness and it doesn't come out to be very costly.

Exadata is pretty stable. There are built-in high-availability features. For example, every component – including the power supply, the fan, the CPU, and the memory – is redundant; you're buying a highly available system.

It absolutely scales to our customers’ needs. You can start with a two-node configuration, and you can go as high as eight racks, I believe, connected together with eight nodes, so you can scale up to 64. That number might have changed recently. You can scale it very well.

Technical support is fantastic. Oracle supports Exadata customers as the top tier, and you get support when you need it. Oracle has separate Exadata support. Once you buy Exadata, you go to that dedicated team and the turnaround is very, very good.

Initial setup is complicated and you need to engage Oracle to do that. Obviously, nobody else can do that. It's complicated, but it's a part of the process.

It's a very good product. I think if you have a business use and can justify the cost, go and buy it. You'll never regret it.

The system is engineered to run both databases and enterprise applications unlike other engineered systems, which are either tailored only for databases or only for applications. The ability to consolidate all of the databases is a big plus.

The SuperCluster provides best of both worlds – Oracle Exadata functionality along with virtualization at the firmware and the OS kernel layers.

On the Exadata side the features most valuable are – Hybrid Columnar Compression (HCC) for both Data Warehouse and OLTP workloads, Storage Indexes for Smart Scan, ability to use Flash Cache for DB storage, and more.

With SuperCluster and Exadata, all of the servers and storage are integrated within the same rack. This reduces the configuration and setup time, increases performance, and makes maintenance and patching easy. We are able to consolidate all of the databases and application stack on a single SuperCluster with Exadata.

In the area of Solaris zone-level virtualization, it would be good to have memory capping as a tool for memory management. Currently for 11g databases running on Exadata with smart scan against hybrid columnar compressed (HCC), tables could result in errors. It would be nice to have a patch rather than the current solution of upgrading the databases to 12c.

Currently, we have a Half-Rack SuperCluster with 2 SPARC T5-8 Compute Nodes and 4 Exadata Storage Servers. Current version of Exadata is 12.1.2.1.3. The compute nodes are running Solaris 11.2 with Oracle11g 11.2.0.4 databases. We've been using it for close to three years.

We had few issues while deploying DB zones. This virtualization has to be carried out differently when compared to DB LDOM virtualization.

There have been no major problems so far with stability.

There have been no major problems so far with scalability.

Overall it has been good so far.

Overall it has been good so far. In the case of engineered systems like SuperCluster/Exadata, the patching has to go through the support team and there is definitely room for improvements in this area.

Previously, we had multiple servers both with and without physical partitioning. The storage for all of the servers had to be zoned to a SAN. The servers and the storage previously were from different vendors and we had to integrate those. Other product offerings were evaluated, but with all of those we had to explicitly integrate the compute, storage, and networking components. In addition, we could not get the benefits of database optimizations with Exadata and have to pay the penalty for virtualization overhead and network traffic between compute and storage layers.

There is a certain degree of complexity with respect to the initial design of the Exadata storage cell disks and grid disks to meet the customer’s application needs. This especially true when migrating from an existing setup. Care has to be taken with regard to the initial domain configurations since this will determine the LDOM and zone-level virtualization. The Exadata disks have to be exposed to both the DB LDOMS and DB Zones.

The initial setup was from the vendor team (for any engineered systems from Oracle). But later we had to continue the setup to cater to our application and business needs. Prior to the implementation, all of the IP allocations are to be completed for the three layers – Client/Public network, InfiniBand private network, and the Management network.

At this time, I do not have the actual numbers but would rate the ROI is pretty good.

In the long run one can consolidate the various DB related licenses. The number of cores required to run the DBs and applications is much lower on a SuperCluster with Exadata, thereby leading to fewer licenses and reduced cost. Since the entire stack is owned by one vendor (in this case Oracle) the core factor for licensing is 0.5 leading to fewer licenses for the software components.

The SuperCluster in addition to Exadata storage also comes with ZFS storage cluster. Since the compute nodes, storage and networking components are fully integrated with InfiniBand I/O fabric this provides very high performance between various components. Also, it has built-in hardware encryption to provide data security.

The following features of Exadata are the most valuable to me:

• Cell offloading
• Smart Scan
• Flash Cache
• Infiniband Fabric
• IO Resource Manager
• Hybrid Columnar Compression
• Storage Indexes

The Exadata database machine is a perfect solution to cater to the needs of large data warehouses. It aptly addresses the problems of database consolidation and marks an end to the argument of running mixed workloads together. With Exadata, we were not only able to consolidate our large data warehouses in a single machine, but we also stabilized the peak workloads. The throughput issues simply vanished with the features like Flash Cache and Infiniband Network Fabric.

On the operational fronts, Exadata streamlines the tasks and responsibilities in a distinctive manner across the levels of an enterprise IT structure. Solution architects act as machine administrators and make sure that the machine is being used in optimal way. At the same time, database administrators, being the familiar heads, don't find much change in roles. Overall, Exadata justifies the total cost of ownership (TCO) by balancing the capital and operational expenses - thereby achieving a higher ROI.

Starting from Exadata X5, virtualization is supported. The feature, though, for some reason, is not much talked about. If Oracle starts promoting OVM and VMWare on Exadata through white papers and case studies, I'm sure it will enable new IT setups in practice. For instance, virtualization support can enable ETL nodes to be resident with Exadata.

I have been using this solution since 2012 (four years).

I did not encounter any issues with deployment.

I did not encounter any stability issues.

I did not encounter any scalability issues either.

I rate customer service 8/10.

I rate it 8/10. My experience with Oracle Support has been moderate. Many times, I find support engineers working on an SR seem to be in a hurry to change the status to 'customer working'. In the past, I've asked them to understand the issue over a call (rather than adding updates) and then suggest the solution.

A vendor team with good expertise implemented the solution.

Find the appropriate workload for the machine. Looking at the data sheet, Exadata is a beast meant to solve bigger problems. You will not be able to gauge its benefits with nuggets of data.

Oracle Engineered Systems are the best solution for Oracle and non-Oracle applications: hardware and software are combined together to ensure high availability and maximum performance. Oracle Exadata contains all the advantages of standard Oracle Database, but on top of that there are exclusive features such as IORM (IO Resource Manager) and HCC (Hybrid Columnar Compression). IORM (in combination with Database Resource Manager (DBRM)) permits you to isolate I/O requests to avoid I/O saturation. HCC provides different levels of compression to reduce space allocation and to decrease I/O.

Oracle Exadata and Oracle Exalogic are standard-based. Customers can easily migrate their existing applications based on Oracle Database and Oracle WebLogic: they can take advantages of Engineered Systems with minimal effort during the migration process. Customers can benefit from high performance without modifying Java or SQL code and they can decide to use specific features after migrating to Exadata/Exalogic. Also, Engineered Systems can be used to consolidate complex environments: for example, our customer was able to decommission about 20 blades, using a single 1/8 Exalogic x4-2. Finally, Exadata and Exalogic simplifies patching/upgrading process, unifying firmware upgrades, ILOM upgrades, OS patching, Database/WebLogic upgrade/patching in one "single-vendor" activity.

We installed the first Exadata in 2011.

The most critical issue was due to a bug with SDP (Sockets Direct Protocol): after connecting Exalogic and Exadata to the same Infiniband fabric and after enabling SDP listeners on Exadata compute nodes, we faced unexpected and random crashes of database hosts. Fortunately, we discovered that SDP caused the reboot, so we disabled it. We permanently resolved the issue with the latest firmware of Exadata, and after that we can take advantage in terms of performance, of SDP protocol.

We used to perform upgrades of Exadata from quarter to half in order to increase CPU/storage capacity. When Exadata hardware is too old (for example, if you want to upgrade from Exadata X3-2 quarter to half, adding x6-2 hardware), you should evaluate the possibility of installing a new Exadata and using Oracle Data Guard to move Oracle Databases.

6/10

There was no solution in place previously.

Using Exadata and Exalogic, you can get the best performance optimizing the number of CPUs, saving costs on software licensing. Also, if you consolidate your environment using Engineered Systems, you can save money on hardware (storage/compute capacity).

No other options were looked at.

As best practice, before starting the installation you should have a look at the Oracle Documentation. Oracle provides a "Getting Started Guide" to get up-to-date checklists to verify requirements (for example IP addresses, name syntax, etc.) and to be sure your infrastructure is ready for the installation. Also, we recommend to install Oracle Enterprise Manager as a monitoring solution: it is fully integrated with Oracle Exadata and Oracle Exalogic, you can monitor hardware (compute nodes, switches, ILKOM etc.) and software (operating system, Database, WebLogic, business applications etc.).

The place where I've used it are mainly on the data warehouses with larger implementations, and having everything in one enclosure is really a lot nicer. Having the disk, all the memory, and all the networking, makes not just the implementation easier, but as you're running it, a lot simpler to administer. Because we're asking it to handle everything, it's a simple solution to administer and support. It also has plugins that makes it easier to monitor.

Definitely having the ability to have one machine do everything has been the biggest improvement for us organizationally. From the administration side, it helps. If one or two admins can handle it, they don't have to go all over the place to, such as to the storage team, the network team, or anywhere else because they can do a lot of things internally. Definitely, we've seen cost, time, and HR savings with several implementations.

Sometimes pricing can be a bit of an issue, especially if customers don't know exactly what they're getting. Maybe they've gone through the whole process and they didn't get the right size or the right amount. I've worked with several customers who had implementations that were too small and then had to upgrade, as well with others who were overpowered with their implementations.

The deployments have been issue-free.

We haven't really had any stability issues. The experiences we have had were when there have been hardware failures, but a Oracle took care of that pretty quickly. Our customer had databases and memory modules die and everything kept running. With the failures we've seen, it's stayed very consistent on the ones we've done.

If it's done with the right specs in mind, it seems to scale pretty well. We haven't really had any scalability issues.

Exadata is pretty much set up when it ships. Afterwards, you just have to copy data over and it's ready to go. So the setup is fairly easy and straightforward.

Definitely do a POC on it. Work closely with Oracle Professional Services to help you with the right configurations.

Some of the features which are specific to Exadata, like histories, offloading and accurate processing. One of the customers we work with had mostly thousands of batch jobs, which used to take, 30 to 40 hours, with the Exadata, and re-architecturing some of those solutions on the Exadata, moving them. It turned out to be running in one hour, or less than one hour. It's a huge, huge performance gain. And in the end, the company realized the return on the investment they have made into Exadata.

Some of the key benefits with our customers moving to the Exadata, they can get it consolidated, all the databases into one platform, one environment. And other advantages, you get a lot of the new features out of the box of the Exadata itself. We don't have to shelve around a lot of different parts of the machines, the networking component, the storage components, you know, all together, one into one machine, into the Exadata. It's a huge, huge advantage for the customers, right? They don't have to look into the history, they don't have to look into the database machines, you know, networking part of it all income passes in one machine. And Oracle has engineered it specifically for the purpose of data. Optimized the database solutions. So once you move it, you get the benefits of your query optimization, your history optimization, your machine data with computer optimization. Everything is blended into one machine.

The key features with the Exadata is offloading the query processing, some of them, at the storage level. That's where Oracle has to make it smarter. I think it's already smarter, the Exadata storage. I can't think of a specific feature that can put in, but that's where I think there's room for more improvement.

Adaptability is a little bit challenging for the customer because of the licensing and the pricing. That's where I think they can make a big difference.

It's a product which Oracle has built with scalability in mind. You can have Exadata hooked up with another Exadata. You have a really high bandwidth network, they call it Infiniband. So you can extend it horizontally as much as you want. So there's a huge, huge opportunity to upscale it and Exadata itself comes in two different flavors, like quarter rack, half rack, full rack. So you can choose and pick, based on your need of the scalability and the future need of how your workload and other things are going to be in future.

You've got to have some basic understanding of networking, storage, some kind of background in the database. That's where it makes it very easy for you to step in and do the initial set up of configurations. Of course, there are certain sorts of specific tools and utility from Oracle which you have to be aware, you have to talk to Oracle very proactively. Make sure you know all the requirements, you need to do the initial setups. You are aligned with the people in Oracle and you plan it. You know, planning is the key, and if you have those basic understanding of storage, database, the networking, those things. It really makes it easy for you to step it into the initial setup planning.

I think it comes to pricing. I know a lot of people don't have the better understanding of the licensing and other things. That's where it effectively comes, in the initial logistics of dealing and getting it.

I think it's a great solution for all the existing Oracle clients and if they really want to move ahead with the next generation of the Oracle Database, that's the solution to go for. Consolidate everything, move ahead with this. You get a better performance, you get better management, you get less people to manage the things, the room of errors are really less because you have specialized people, less people, and it's very valuable.

Rating: so I'll say that in terms of the rating it will fall in 8 to 9. In terms of the ease of maintenance, you have all the tools and everything Oracle provides. So you don't have to poke around and see what's where, if somebody should come, so where you have to go. If we're really good, you know the Exadata, you know what you're looking for, you talk to Oracle, get those tools, do it. In terms of the performance, you don't have to waste your night, day and night. It's a lot things come built in with the Exadata itself in terms of the price.

It's pretty simple in my opinion. It's called out the box, and it's called support. Sometimes with the engineered systems based on the support agreement that you have with the vendor, sometimes you hear about problems with your system from them rather than your applications experiencing some issues, and you're having to log a support ticket with Oracle. I think that's huge.

The thing that attracts them to the two products but mainly the Exadata is the storage technology is revolutionary whether it's the only solution of that flavor on the market or not, I'm not familiar with. It's the storage solution, it's the speed, and it's the high availability. If somebody wants 5 or 6 9s worth of availability, a very good place for them to be is with one of Oracles engineered system solutions.

As is always the case with Oracle, when some late-breaking fancy technology gets under their radar, self-invented or perhaps something that they're picking up that they'd like to compete with from another vendor, they're all over it. I can't specifically think of anything myself. Outside I guess of speed and maybe the other two things I could think of are speed and speed, but I'm not suggesting that speed is an enhancement because there's anything wrong with the speed of the system now, but of course we always like to do things in four nanoseconds rather than seven.

I think I would probably give it a capital S. People here peripherally hear about informational stuff that may be going on in the cluster, but as far as catastrophic events that may happen with the cluster, a lot of them in some way, shape or form are not far from self-healing. I think that's a huge advantage of an engineered system.

If you buy a quarter rack, you feel like you need more computing power, you upgrade to a half rack. You go to a 3/4 or a full rack. It's basically I guess we used to call it plug and play except a lot of us found it in the olden days, it was plug and pray. I think they've got that one licked to the max.

Sometimes it's them calling us on behalf of one of our clients, but I would find that once one figures out the best way to work with their support organization in general, but their engineered systems support organization in particular, I can't think anything outside of a world class organization. I would admit that I don't have a lot, if any, experience with the competition, but I don't feel like I need any experience with the competition because of the way they do support those two products.

A lot of the times it ends up being complaints from the user community, and difficulty with strategic activities that the system performs. If you're running a commercial off the shelf application against these engineered system, sometimes it's a little more complex to work with a third party vendor, to speed things up. Mainly it's complaints from the user community. You and I both know that if something takes four seconds all the time and takes seven seconds some of the time, all the people are going to remember is the seven seconds.

I haven't done it on a hands-on basis, but I believe that what I mentioned at the beginning about out of the box is just that. Oracle from my recollection first tried to venture into some form of engineered system in the '90s with HP. It was a red box that you basically took out of the carton and plugged in. I think that's pretty much close to what once you brought it onto the network, exactly what you're going to do with their systems, so I thinks it's plug and play.

Choosing a vendor: obviously cost, high availability, and the strength of the rack technology. It's a corner of the Oracle technology that basically not exactly gets reinvented, but the new features and the new ways of going about to do things and the capabilities of failover and all that is a huge concentration as the product matures, not that it is immature. I think that's a primary reason that people might consider looking at this. There's absolutely no doubt, it's not far from an upgrade basically being done on the fly. There's so many things that can be done online. The plague of course of the '90s and the early-2000s was downtime.

Rating: I'd call it about a 12 out of 10. Of course you have to be able to afford the box. You have to be able to afford the configuration that you're going to be getting into. It's cost-effective for a lot of organizations, but I would rate the technology very high. Some of the specialized mechanisms that Oracle has put in place, especially with respect to Exadata and the late-breaking version of Exadata in particular.

As I said before, I'm not intimately or even somewhat peripherally familiar with the competition, but these guys know what they're doing, and my experience has been in the past that if Oracle ever plays catch-up in a technology spectrum with the competition, watch out. We all know that it's now cheap. It's affordable for a lot of organizations. If cost is going to be something, it's going to ultimately drive an organization's buy or no buy decision. The benefit after the money is spent and an ongoing outflow of cash to the vendor, if it makes business sense for somebody, I don't think they can be in a better place.

Make sure it's the right solution. Make sure that you do indeed need their real application cluster solution, which we affectionately call rack. Make sure it's right for you. It sounds corny, and it's sounds like a cliché, but it applied when I got started in IT in the 1800s, and it still applies. If it's cost-effective, go for it.