WAN optimization continues to evolve; integration with data deduplication products on the horizon

The adoption of cloud storage and virtualization are driving wide-area network optimization implementation and it is becoming clear that WAN optimization more flexible than we've thought about it in the past. Jeff Boles, senior analyst with The Taneja Group, discusses WAN optimization in this Q&A.

The adoption of cloud storage and virtualization are driving wide-area network (WAN) optimization implementation and it is becoming clear that WAN optimization is more flexible than we've thought about it in the past. Jeff Boles, senior analyst with The Taneja Group, discusses WAN optimization in this Q&A. His answers are also available as an MP3 below.

Listen to the WAN optimization FAQ

Table of contents:

>> Have data deduplication products and WAN optimization products been better integrated this year?
>> How does virtualization play into the WAN optimization picture?
>> What are the significant developments in the WAN optimization market in 2009?
>> Are WAN optimization solutions expensive?
>> What are the alternatives to WAN optimization products?

In our conversation on WAN acceleration and WAN optimization in October of last year, you mentioned that end users could expect that WAN optimization vendors would diversify their offerings and integrate data deduplication. Where are vendors at in this area? Has deduplication become more integrated into WAN optimization products?

While that was my prediction then and remains my prediction for the future, it's been a long time coming. At this point in time the question remains: Why hasn't there been better integration with deduplication? Deduplication is a fundamental component of most of the WAN optimization solutions out there. WAN optimization products reduce redundant bit patterns, pack more data into a limited amount of bandwidth and decrease latency. And, that's similar to the technology used to deduplicate disk storage today. So, it would be a natural progression to see some of those vendors converging.

However, the architecture of WAN optimization technology is complicated, and it's also complicated to then convert this in-band data to deduplicated on-disk storage. There has been a lot of energy spent in the industry to make this happen.

Although there's plenty of opportunity there, WAN optimization is equal parts latency and capacity optimization. Many of the current approaches for data deduplication can have an I/O impact on storage. Consequently, they are better suited for secondary storage, and you have to consider how you are going to implement capacity optimization for data movement through WAN optimization.

Data deduplication is only going to become a more important technology for the enterprise. And we will continue to see evolution on the capacity optimization side as well as on the data reduction side.

How does virtualization play into the WAN optimization picture?

Virtualization has created some interesting new opportunities. First, consolidation means that more organizations are trying to move their services back to a centralized location. This makes accessing services over the WAN more important than ever. So, you may be moving some of your branch offices back into your data center. As you move into this consolidation mindset, getting increasingly green and using higher density platforms in the data center, more data will be moving back to the data center and you need more efficient access over the WAN. There's clearly an opportunity for WAN optimization there.

Second, virtualization has made the workforce more mobile. In some cases, this might involve data movement over a WAN. Organizations may be moving virtual image files to and from branch offices or some other location. Or, they might be moving services to and from the data center and the branch office for management purposes. Again, WAN optimization has a lot of value there. WAN optimization can also allow you to implement interesting failover/failback scenarios between those branch offices and the central data center.

Several of the WAN optimization solutions available now actually allow you to run your WAN optimization appliance as a virtual machine.

Can you outline any other significant developments in the WAN optimization market this year?

The enterprise often is not optimized for the cloud services today. Let's think about multiple hops. There are a broad number of businesses out there today that are using WAN office connections back into a data center with centralized access to the Internet. Some may be making multiple WAN hops -- from remote to branch to a central office, for example. And for the cloud, it's more than likely that won't be good enough for you because each one of those hops has some latency associated with it. That may make a cloud service totally unusable for the people in the remote or branch offices.

WAN optimization has a valuable play there. That's one of the things that we see emerging today, and has emerged over the past year. More people in the remote office scenario are using cloud services. So, optimizing the WAN has become more important than ever, and cloud is actually driving WAN optimization in some cases.

Riverbed just recently added themselves to the list of companies offering a virtual appliance. These appliances give organizations the ability to optimize their data movement effectively and address some data movement challenges to and from the cloud. We are just starting to see that use case evolve, and I think we'll see that become a fairly significant use case in the future.

Also, the market has matured in how it thinks about WAN optimization broadly across the all aspects of IT. WAN optimization technology is being used to improve disaster recovery, for example. Service providers who offer disaster recovery can use products, such as Riverbed's Steelhead, to optimize data synchronization, replication for disaster recovery, and extend those services to multiple locations. A provider could use an appliance, and a customer might have a single office appliance and laptop agents to optimize data replication and protection for remote workers as well.

WAN optimization is much more flexible than we've thought about it in the past. It can be used anywhere that data movement in important.

Are WAN optimization solutions expensive? Are there products available at lower price points for smaller businesses?

Most of the vendors out there have a range of appliances, but they certainly come at a price. However, I'm convinced from personal experience and from conversations with users all over the market that you get what you pay for in the WAN optimization market. There are differences between the appliances out there, and the vendors seem to be very aware of the differences between their products and how they compare to each other. Consequently, the products that come at a higher price offer some unique capabilities. So, be aware of that when you are looking at prices.

I think today, most vendors offer something to fit most needs somewhere in their product range that will be an appropriate fit and won't be excessively expensive. If you are at the size where WAN latency or WAN bandwidth is a concern, there is going to be a solution for you out there. The market is very mature at this point.

What are the alternatives to WAN optimization products? Is there anything users can do to increase performance without buying any new products?

Even with an extensive background network architectures and infrastructure management, I have little to recommend. The WAN optimization vendors are successful because their gear is valuable to end users with WAN connections. In some cases, you could try to homegrow some half-baked solution by integrating compression but really these devices offer a combination of compression, data deduplication and caching to address both latency and bandwidth utilization simultaneously. That's a pretty complex orchestration.


Dig Deeper on Disaster recovery networking

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.