2. Introducing First All-In-One Testing
Solution For Storage Professionals
Integrated All-in-One The Insight You Need To
Network Storage Solution Control:
Costs
Risk
Block
Performance
4. Step
1
Set-Up: Production Workload Modeling
PRE-BUILT TEST SUITES
PRODUCTION STATS
PACKET CAPTURES
Detail / realism
Flexibility
Parameterization
5.
6. Step
2
Test: High Performance Loading
FC
iSCSI
SMB
NFS
SwiftTest Appliance Systems Under Test
OBJECT
7. Step
3
Results: Better Insight = Better Decisions
Analytics Insight
Block vs File
Boot storms
Capacity planning
Sunny Day vs
Rainy Day
Failure prevention
8. When Do Customers User SwiftTest?
Before you buy
Before you Storage
Infrastructure
change
Before you deploy
Test at critical points
9. “We were tired of
‘testing in production’.
With SwiftTest, we save a lot
of money by controlling our
cost-vs-performance-vs-
capacity equation. And I
sleep at night knowing our
storage will perform in
production.”
Storage Engineer
Go Daddy
11. SwiftTest Product Overview
SwiftTest 3000
(8) 1GigE ports
SwiftTest 5000
(2) 10GigE ports
SwiftTest 6000
(2) 8G FC ports
▸ Client/Server Emulation ▸ Network Emulation
SMB (CIFS) 1.0, 2.X, 3.0 over TCP and NetBT IPv4/IPv6 support emulating
Client, SMB 1.0, 2.x Server millions of concurrent users
NFS: v2, v3, v4 and v4.1 Client, NFSv3 Server Multiple MAC addresses
iSCSI Initiator VLANs supported
FC Initiator Configurable TCP
HTTP Client and Server Parameters
Authentication: NTLM, Kerberos, and CHAP
12. The Future of Testing:
The SwiftTest Intelligent Testing Portal
WORKLOADS
PROTOCOLS TESTS
STORAGE API’s
NETWORK COMPUTE
Editor's Notes
It is an integrated, multi-protocol storage testing solution. Simultaneously, you can test file, block, and object storage devices - CIFS/SMB, NFS, iSCSI, FC, HTTP.As I mentioned before, it is very high performance – and scalable – so by stacking boxes, you can test @ 40G, @ 100G – all within 1 GUI. If you happen to already have automated testing, we integrate into that as well.It is the most full featured storage testing platform availabl. As one of our customers at EMC said, basically, “We thought of everything”. So what exactly is “SwiftTest”?The solution is an appliance that enables you to emulate real world data traffic, completely paramaterizable (file size, block size, read/write ratio, etc.), then scale it up to extreme proportions – millions of clients, 1000’s of servers, hundreds of VM’s – and then load test your networked storage. It is not a monitoring tool; instead, it is a system that is intended to stress test your architecture to find it’s breaking points. It does not belong in production. Some people ask me if it is like IOmeter- the SwiftTest solution is much more than this. We do not just send a bunch of 0 & 1’s. The appliance does write/read verification and works at the protocol level, giving valuable insight to how traffic is handled by command.
Enterprise customers typically use SwiftTest at 3 key points: before they buy, before they deploy, and before they make a change. That includes changes in infrastructure and applications.For instance, when Ellie Mae decided to migrate to a private cloud, deploying hundreds of remote clients booting to VM’s, they brought SwiftTest in during their planning and implementation to help fine tune the overall architecture. We helped them work through a variety of virtualization issues such as boot storms and rogue VM’s not loading at all. By using SwiftTest, they cut down their deployment time and achieved a successful rollout. In the case of GE, they initially found us while they were on the wrong side of the “don’t test in production” camp. After spending too much time troubleshooting a new application rollout, they brought in SwiftTest and we helped them to emulate their exact traffic pattern and then scale out a test to find the exact bottlenecks that were causing their new app to fail. By testing with SwiftTest, they were able to track down the issues and finally deploy their Big Data application.For customers like Wells Fargo and GoDaddy, they were looking to deploy new infrastructure. Using SwiftTest, they were able to emulate their real world data and test potential vendors to compare ROI. To this day, GoDaddy, uses their SwiftTest appliance 3 days a week testing things like new firmware, hard drives, or any other configuration change to their architecture, ensuring they have the most resilient Data Center possible.
Enterprise customers typically use SwiftTest at 3 key points: before they buy, before they deploy, and before they make a change. That includes changes in infrastructure and applications.For instance, when Ellie Mae decided to migrate to a private cloud, deploying hundreds of remote clients booting to VM’s, they brought SwiftTest in during their planning and implementation to help fine tune the overall architecture. We helped them work through a variety of virtualization issues such as boot storms and rogue VM’s not loading at all. By using SwiftTest, they cut down their deployment time and achieved a successful rollout. In the case of GE, they initially found us while they were on the wrong side of the “don’t test in production” camp. After spending too much time troubleshooting a new application rollout, they brought in SwiftTest and we helped them to emulate their exact traffic pattern and then scale out a test to find the exact bottlenecks that were causing their new app to fail. By testing with SwiftTest, they were able to track down the issues and finally deploy their Big Data application.For customers like Wells Fargo and GoDaddy, they were looking to deploy new infrastructure. Using SwiftTest, they were able to emulate their real world data and test potential vendors to compare ROI. To this day, GoDaddy, uses their SwiftTest appliance 3 days a week testing things like new firmware, hard drives, or any other configuration change to their architecture, ensuring they have the most resilient Data Center possible.
Finally, I would like to share the future vision of SwiftTest. We are constantly adding new protocols like FC or FCoE and features of course, but the true end game is to be the integral portal to your change management process. For storage, we are looking to expand heavily into object storage with CDMI, S3, CEPH. Networking will be expanded beyond what we do for WAN optimizers and Load Balancers. For compute, we are growing our virtualized testing capabilities and moving into Hadoop/Big data. API’s such as JSON & REST will be added. We are also looking to integrate 3rd party tools and even considering some monitoring functionality. We also plan to have a VM version, hopefully by mid next next year.Beyond that, we are building a library of tests and workloads – truly developing methodologies on testing for specific things like VDI deployments and database apps like SQL.That is the 10,000 foot view of what SwiftTest does. Before I turn it over to _____ to walk us through the TDE, can I answer any questions?